Ok, this could have gone in the film grain or American Werewolf threads, but it's relevant to the Gladiator issue and so many other releases. John Landis is giving an interview about the restoration of An American Werewolf in London, and makes a few points that many of you will find interesting. The first is about the limited understanding of digital processes that he probably shares with most directors of his generation, and, the amount of fighting it takes to get us computer geeks to restrain ourselves in the use of digital 'improvements':

g6lWNSDsTZg

I have no problem imagining how insulted some techs feel when you don't want to throw all the toys in the toybox at it, and that's the balancing act of any restoration, I'm sure.

That should be good news for AWiL ;)

I like that he's enthusiastic about what he made and how he wants it displayed. If only every director and DP could take the time to look over the masters for home media. Too bad Ridley is busy right now. And James Cameron could have popped in to look at T2, but he was busy with the blue Thundercats.

So was he saying that the technicians like to DNR and boost contrast?:confused: Those sound like bad habits.

He's the guy that had them 'dirty up' "Animal House" after the restoration made the print too 'clean' for him for the look of that movie.

"No..no I want the grain back!"
The proof in the pudding I might say...;) never ever leave a computer geek alone in the mastering room...:D

I like that he's enthusiastic about what he made and how he wants it displayed. If only every director and DP could take the time to look over the masters for home media. Too bad Ridley is busy right now. And James Cameron could have popped in to look at T2, but he was busy with the blue Thundercats.

So was he saying that the technicians like to DNR and boost contrast?:confused: Those sound like bad habits.

They probaly also like overly compressed music....:(

Does this mean we can expect a degrained contrast boosted An American Werewolf in London ?

Fingers and toes crossed for this one.

"The technicians made [Animal House] beautiful, it looked like a Doris Day movie. So they showed to me and I said, 'No! Where's the grain?' 'We took it out.' 'No, no, I want the grain'".

:D

And that cuts to the core of the issue. If you think that an image with the grain left in is degraded and without the grain it's improved by definition you are unfit to master movies to HD since you don't know the first thing about movies (and HD mastering as well). People with that attitude should never ever be allowed to work in this field without strict supervision. They are a menace to the film heritage. It does not help that everything from scanners to AVC encoders now has these tools integrated and available at the touch of a button. An invitation to disaster in naive hands.

One of the most interesting points he makes here is about the bottle labels in the pub. He says you can now read them, whilst in the cinema they'd be 'out of focus' (though I appreciate he's speaking in short hand).

To me, this suggests that directors - at least this director - presume that when a film is shown in the cinema it quite probably won't be in a way which uses the full resolution of the film. So whilst we've often talked about haw raw film is higher definition than HD, he's actually saying that in practice you'll see more with HD than with film.

Steve W

To me, this suggests that directors - at least this director - presume that when a film is shown in the cinema it quite probably won't be in a way which uses the full resolution of the film. So whilst we've often talked about haw raw film is higher definition than HD, he's actually saying that in practice you'll see more with HD than with film.


Yes testing has shown that a standard theatrical print has suffered from degradation and end up having somewere between 450-750 lines of resolution.

Thats what many directors had in their calculation when they decided to shoot their movies digital. That its good enough for the theatrical print.

But then we have the issue about colors.

One of the most interesting points he makes here is about the bottle labels in the pub. He says you can now read them, whilst in the cinema they'd be 'out of focus' (though I appreciate he's speaking in short hand).


At the theater you're watching elements that are several generations away from the original (ocn -> interpositive -> internegative -> release print). Whereas with BD you may be going as far upstream as the original camera negative.

This interview shows why the 4k remaster of Blues Brothers looks so good

At the theater you're watching elements that are several generations away from the original (ocn -> interpositive -> internegative -> release print). Whereas with BD you may be going as far upstream as the original camera negative.
I doubt very many releases get back to the camera negative. The "lost scenes" from Casablanca are about the only examples I can think of, and I'm think those were interpositives. I'm thinking even the "Seinfeld" stuff was re-edited from dupes. Either way, when you go back that far, you don't have a sound track and you have to go through all the color timing processes again. However, your point remains the same.

Not to mention, in the theater, it's being blown up really, really big, so it ends up being like comparing a 4x6 photo print with an 8x10 of the same source. The 4x6 will usually appear sharper.

In addition, the release print tends to get muddied up more and more as the run progresses.

Finally, there are quite a few theaters that don't project films at full brightness in order to preserve the protector bulbs longer in order to save money.

The irony is, lowering the brightness of those projector lamps does not in any way extend the life of the bulb. The manufacturers have proven this with testing, yet theaters do it to this day.

The irony is, lowering the brightness of those projector lamps does not in any way extend the life of the bulb. The manufacturers have proven this with testing, yet theaters do it to this day.

Exactly, and if you ever see a film with a properly lit and focused projector it's like a revelation; it's almost unbelievable how good it can look.

Exactly, and if you ever see a film with a properly lit and focused projector it's like a revelation; it's almost unbelievable how good it can look.

I dont doubt that this may be true, but

What is eco-mode good for then?

I dont doubt that this may be true, but

What is eco-mode good for then?

He's talking about a film projector in a theater. Eco-mode is for home digital projectors.

I dont doubt that this may be true, but

What is eco-mode good for then?

Lower fan noise. Can be important if your room means you have to sit within close proximity of the projector.

never ever leave a computer geek alone in the mastering room...:D

said computer geek could be just called incompetent.

Does this mean we can expect a degrained contrast boosted An American Werewolf in London ?

Fingers and toes crossed for this one.

Sounds like the opposite to me, assuming of course Landis looked at it afterwords and once again scolded the moron that thought the image would be a degradation without contrast boost and DNR.

John Landis should produce ALL BD masters since he loves the grain and the original look :)

He sure bitchslapped those techs for taking away the grain on Animal House and making it to colorful and bright :)

Does this mean we can expect a degrained contrast boosted An American Werewolf in London ?


If you listen to what he says in the interview ha says he made them darken and grain the picure :)

So he once again bitchslapped some sence in the mastering guys who probably made the first version too bright and clean ;)

One of the most interesting points he makes here is about the bottle labels in the pub. He says you can now read them, whilst in the cinema they'd be 'out of focus' (though I appreciate he's speaking in short hand).

To me, this suggests that directors - at least this director - presume that when a film is shown in the cinema it quite probably won't be in a way which uses the full resolution of the film. So whilst we've often talked about haw raw film is higher definition than HD, he's actually saying that in practice you'll see more with HD than with film.

Steve W


This is what I have been witnessing since looking at HD, especially with BD. As much as I like the grain, not excessive distracting grain that crawls like on some BD's, he is making the case for BD being better than the theater, yet we keep hearing how we want it to be close to film as possible and I agree, but not what we are seeing in the theater when what we are seeing at home is "better".

At the theater you're watching elements that are several generations away from the original (ocn -> interpositive -> internegative -> release print). Whereas with BD you may be going as far upstream as the original camera negative.

See my previous post. While I understand your point that misses the point Landis is making which is we are seeing more at home then in the theater and that is what we want or at least what I want, as close to film like, but also if I can see it better than in the "state of the art" commercial theaters then that is a plus.

If you listen to what he says in the interview ha says he made them darken and grain the picure :)

So he once again bitchslapped some sence in the mastering guys who probably made the first version too bright and clean ;)

No speakers connected to the computer i'm afraid.

I was just reading the post that came with the link and i guess assumed the opposite of what is happening. Thats good news and i think i'll buy it without waiting for any reviews. ( That makes a change )

Thats good news and i think i'll buy it without waiting for any reviews. ( That makes a change )

+1

This is one of my favorite movies of all time, and this time it finally may get the treatment it deserves :)
I wish John Carpenter was present when Uni made the master for the BD, then they might have kept the grain as well ;)

And that cuts to the core of the issue. If you think that an image with the grain left in is degraded and without the grain it's improved by definition you are unfit to master movies to HD since you don't know the first thing about movies (and HD mastering as well). People with that attitude should never ever be allowed to work in this field without strict supervision. They are a menace to the film heritage. It does not help that everything from scanners to AVC encoders now has these tools integrated and available at the touch of a button. An invitation to disaster in naive hands.


post of the year!!!

couldn't agree more.

I hope darkening and graining the image does not mean he is after a french connection look

This interview shows why the 4k remaster of Blues Brothers looks so good

You mean the dvd?

Nope this http://www.cineworld.co.uk/films/2585
I saw it all all the Universal remasters projected digitally at 4k

I know I'm in the minority here on AV enthusiast forums... but I hate directors like this. Give me the clean/bright look of digital cinema over film look any day of the week. I hate film grain! It's a flaw of film technology and yet we embrace it???

I hope directors like this one retire soon and leave the movie industry for good. :mad:

I know I'm in the minority here on AV enthusiast forums... but I hate directors like this. Give me the clean/bright look of digital cinema over film look any day of the week. I hate film grain! It's a flaw of film technology and yet we embrace it???

I hope directors like this one retire soon and leave the movie industry for good.

Is it too much to hope you are being sarcastic? :p

I know I'm in the minority here on AV enthusiast forums... but I hate directors like this. Give me the clean/bright look of digital cinema over film look any day of the week. I hate film grain! It's a flaw of film technology and yet we embrace it???

I hope directors like this one retire soon and leave the movie industry for good. :mad:

In the meantime at least you can enjoy movies like Waiting and Gladiator:rolleyes:

And that cuts to the core of the issue. If you think that an image with the grain left in is degraded and without the grain it's improved by definition you are unfit to master movies to HD since you don't know the first thing about movies (and HD mastering as well).


The real question is-- why are they the ones making the decision (to "clean" or degrain, etc)? Shouldn't someone be supervising them? I'm sure if they were properly informed (or truly realized that they weren't making it better), this wouldn't be happening. I don't blame the encoders, I blame the management.

I know I'm in the minority here on AV enthusiast forums... but I hate directors like this. Give me the clean/bright look of digital cinema over film look any day of the week. I hate film grain! It's a flaw of film technology and yet we embrace it???

I hope directors like this one retire soon and leave the movie industry for good. :mad:If all movies were shot digitally, I could agree with that. However, digital-cinematography did not yet exist when Landis made most of his movies, and the whole point of home cinema media should be to try and replicate its original look from the theater.

I know I'm in the minority here on AV enthusiast forums... but I hate directors like this. Give me the clean/bright look of digital cinema over film look any day of the week. I hate film grain! It's a flaw of film technology and yet we embrace it???

I hope directors like this one retire soon and leave the movie industry for good. :mad:You make an interesting point, but I'm not sure it's the one intended. I can't tell if the members here like grain or not. I'd just as soon not have it. And yes, it's an artifact that doesn't have to be there if they had settled on a proper standard way back when (which would have given us grossly better motion as well.) I mean larger negative and faster frame rate here.

So, the reason I hope people here object to grain removal is it's not possible to do without screwing up everything else. If you enjoy grainless digital 35mm film sourced media then you haven't educated yourself on what it should look like - you haven't discovered all the detail you're missing. If you don't care, then you're just saying a version of the old saw "I don't know art, but I know what I like." That's a euphemism for "I like crap and I don't care what anyone thinks about it."

There are people here who think 35mm 24fps film is some kind of beautiful treasure and I think they're nuts. It's a lowest acceptable quality standard that should have been gone long ago. But it's still film, and film is magic. You either get that or you don't (and obviously many industry people, including directors, don't.)

I hope darkening and graining the image does not mean he is after a french connection look

This movie has always been dark and grainy on all formats to date.
The idiots who made the BD transfer probably brightened the picture and whiped all the grain away, and that´s why he made it darker and grainier again.

I wish Blue-Underground didn´t brighten and de-grain their DVD of "Zombie". It went from (in the original version) dark and murky to bright, sunny and colorful. NOT the original intent of Fulci! ;)

Incredible how a little bit of alteration from the original look can destroy the entire movie experience ;)

You make an interesting point, but I'm not sure it's the one intended. I can't tell if the members here like grain or not. I'd just as soon not have it. And yes, it's an artifact that doesn't have to be there if they had settled on a proper standard way back when (which would have given us grossly better motion as well.) I mean larger negative and faster frame rate here.

So, the reason I hope people here object to grain removal is it's not possible to do without screwing up everything else.
Absolutely totally wrong, wrong, wrong!

Have you ever heard of something called "composition"? Film is an artistic expression, not a peek out a freshly cleaned window. How the final image appears is an expression of what the director is trying to achieve, not how the street looks outside your home. Grain, color pallette, and absolutely everything else has nothing necessarily to do with what you personally see on a sunny (or otherwise) day, but instead has everything to do with what the director is trying to conjure for the viewer. Grain is part of the pallette the director has available, and its arbitrary removal is not an improvement, it's a destruction, just like the application of DNR and edge enhancement.

That video confirms what I've suspected and feared ever since the discussion about Patton: That hidef production facilities are utterly INFECTED with film hating, grain hating cretins who are dying to KILL the original look so they can make it look like bright, shiny HD video. Ugh. It's a wonder we get good transfers at all with so much of that sort of thinking around.

Absolutely totally wrong, wrong, wrong!

Have you ever heard of something called "composition"? Film is an artistic expression, not a peek out a freshly cleaned window. How the final image appears is an expression of what the director is trying to achieve, not how the street looks outside your home. Grain, color pallette, and absolutely everything else has nothing necessarily to do with what you personally see on a sunny (or otherwise) day, but instead has everything to do with what the director is trying to conjure for the viewer. Grain is part of the pallette the director has available, and its arbitrary removal is not an improvement, it's a destruction, just like the application of DNR and edge enhancement.Well sure, I agree with you. But I doubt many directors, if economics weren't a factor and supply was plentiful and distribution and display weren't factors etc. etc., would choose 35mm/24fps over say 65/60. They work with what they've got, but what they've got is a joke is all I'm saying. It's an old argument that I've never understood that you seem to be making - that low res film is some kind of artistic high. It's not to me, it's junk, and the only reason we're stuck with it is the cheapo mentality that settled on it to begin with so long ago. Now it's entrenched and there's nothing anyone can or will do about it. But, as I said, and you didn't quote, in my original post, it still reigns supreme over anything digital has come up.

That video confirms what I've suspected and feared ever since the discussion about Patton: That hidef production facilities are utterly INFECTED with film hating, grain hating cretins who are dying to KILL the original look so they can make it look like bright, shiny HD video. Ugh. It's a wonder we get good transfers at all with so much of that sort of thinking around.

Is this because the engineers sit too close to the monitors where everything look more pronounced prompting them to take desperate measures ?:confused:

It's a flaw of film technology and yet we embrace it???
distortion was a flaw of early guitar amps yet everyone seems to enjoy a good guitar solo ;)

Well sure, I agree with you. But I doubt many directors, if economics weren't a factor and supply was plentiful and distribution and display weren't factors etc. etc., would choose 35mm/24fps over say 65/60. They work with what they've got, but what they've got is a joke is all I'm saying. It's an old argument that I've never understood that you seem to be making - that low res film is some kind of artistic high. It's not to me, it's junk, and the only reason we're stuck with it is the cheapo mentality that settled on it to begin with so long ago. Now it's entrenched and there's nothing anyone can or will do about it. But, as I said, and you didn't quote, in my original post, it still reigns supreme over anything digital has come up.
There are plenty of directors/DPs in love with technical imperfection.
I mean, why in gods name would anyone shoot a movie with anamorphic lenses from a practical perspective? You can get perfect quality anamorphic prints shooting with much smaller, cheaper, better-performing, and faster spherical lenses, and use less film doing so, yet you still see a few movies shot anamorphically once in a while.

Some of the posts in the thread confuse me a little.

I'm hoping they did not first degrain the image which results in some loss of detail and then Landis simply ordered them to add simulated grain. Some of the posts seem to imply that.

I would rather they simply acquired the best possible print to make a master with and left the grain intact and thus the detail intact. In fact can someone just tell these young kids working in the industry today to leave the processing tools alone. Just because they have them doesn't mean they have to use them.

I don't want the images contrast boosted and i don't want the grain removed. Just give us whats on the master. If the master isn't good enough then get a great print of the movie which every studio should have and then make a new master and then transfer to Blu Ray. Yah yah i know it costs money to do that but then it costs money to throw processing at something when it isn't needed. It's cheaper to just cut out the processing. I'll take some dirt and specks on a print over scratch removal and DNR and EE tools and contrast boosting and all the other annoying things they can and do to some of my favorite films.

The words i dread hearing from a studio regarding old catalog titles are "This film is being remastered" That usually just means they are using the existing master and applying processing to it. ( Especially if the studio happens to be Lionsgate, Paramount or Universal )

Ok thats my early morning rant over. Now i gotta have a bath.

Some of the posts in the thread confuse me a little.

I'm hoping they did not first degrain the image which results in some loss of detail and then Landis simply ordered them to add simulated grain.

He clearly ment that they backed up a bit to the point before the de-graining ;)

I'll take some dirt and specks on a print over scratch removal and DNR and EE tools and contrast boosting and all the other annoying things they can and do to some of my favorite films.


I agree!

I actually think the scratches and dirt brings life to the movie by not giving that sterile digital "perfect" look. It IS 35 mm and not digital video ;)
I don´t mean the prints should be in the shape of an old grindhouse movie, but a few lines appering a few times during the 90 minutes of a movie, I have no problems at all about. :)

I would rather they simply acquired the best possible print to make a master with and left the grain intact and thus the detail intact.

So, devil's advocate here for a second.

Let's say the best print they could find was several generations down the line, and grainier than intended.

The director thinks this version has so much grain that it's distracting, and chooses to use a little DNR to calm the grain, and can do this without losing any (or without losing very much) detail.

Is this an artistically valid process?

Steve W

So, devil's advocate here for a second.

Let's say the best print they could find was several generations down the line, and grainier than intended.

The director thinks this version has so much grain that it's distracting, and chooses to use a little DNR to calm the grain, and can do this without losing any (or without losing very much) detail.

Is this an artistically valid process?


Careful noise reduction in a remastering process can be needed in some cases. But this is not what happens on most titles that gets noise reduction. How often does they only have access to a several generations down the line print?

In most cases they do have an acceptable print, but use DNR to make it more "highdef".

And overusage of DNR usually makes the image look further away from the orginal then a to grainy print would. So in the end dont overdo a movie and destroy the organic filmlook in the process.

can we send technicans back to film school again when they assume that grain degreades the image? -.-

But I doubt many directors, if economics weren't a factor and supply was plentiful and distribution and display weren't factors etc. etc., would choose 35mm/24fps over say 65/60. They work with what they've got, but what they've got is a joke is all I'm saying.

Interesting post. Shooting and projecting widely distributed mass-market movies at ultra-high resolutions and frame rates – if it isn’t economically feasible today, it certainly will be in the next decade. (Running a 2160p theatrical digital projector at 60fps shouldn’t present any problems, tomorrow’s hard drives will make the extra data storage cheap enough to be a non-issue and the cameras (digital ones) already exist.)

The real question is the aesthestic one of whether filmmakers and viewers will want the hyper-realistic effect created by super high res, high frame rate video. Perhaps movies aren’t supposed to approximate the motion and detail of real life that closely, just as a comic strip isn’t intended or expected to look as lifelike as a still photograph. Indeed making a cartoon that realistic might destroy its whole point.

It’s unclear whether the economics of film production alone prevented 65mm/60fps movies from catching on. I think the next ten years will tell us one way or the other.

I could not tell exactly what he meant when he said he made them put the grain in and darken the image. Was there an unaltered master from the movie that had the movie "as is" ?

Or did the techs simply take what they had and run it back through adding what they thought would be the "right amount of grain" and change in contrast. big difference.


I would prefer less film grain except where it goes through the processing that makes a result like Patton or Gladiator in BD---or when grain was done intentionally by the director to represent something (such as in Christmas Vacation, when watching the old family movies in the attic).

I want the skin textures, the details as is, as fresh from the original negative as possible. But grain being there just cause it shows up in the transfer, that is not something I want. Reasons are as follows:


In making non-digital duplicates from the original, the duplicates will always have a little less detail---I used to have to deal with this problem as a photographer in making duplicate originals (before the days of digital).

Digital reproductions of negatives are even more of a problem even when the negative is perfect. The process seems to introduce a certain amount of artifical grain or artifact ("digital noise") from some type of sharpening effect of the digital process.

In my work, rather than use a digital noise reduction that can produce the patsy look people complain of, I would blow up the image to very high detail and remove select pixel noise iin photoshop by use of the clone stamp and other such devices---but one has to be careful even with those tools. EVen using 22 mb sensors that are dedicated to reproducing 35 mm photo negatives, the noise just pops up. I tend to do several scans of a negative, and when looking at the digital images in the series, one can see the random changes in the noise that identifies it as such, rather than "film grain".

Frankly, sometimes i think what we may preceive as film grain is not film grain, but simply artificial digital noise. Unfortunately, in trying to get rid of the digital noise, one can do more damage than good. I have dealt with this problem for over ten years in digital photography.

Edit note: The last thing I want is a techie who goes to film school and thinks he is now the "artist" and needs to do some art. All you should want is some intensely anal individual computer tech-nerd who is trying to duplicate the original into as an exact reproduction as humanly possible, without trying to ascertain the director's original intent.

Indeed the problem may be too much "artist" work going into the reproductions, rather than too little.:eek:

can we send technicans back to film school again when they assume that grain degreades the image? -.-That's assuming they ever went to film school in the first place. I don't get the impression these "I'm dying to change the original look" keyboard jockeys know anything about film or have any appreciation for it.

Well sure, I agree with you. But I doubt many directors, if economics weren't a factor and supply was plentiful and distribution and display weren't factors etc. etc., would choose 35mm/24fps over say 65/60. They work with what they've got, but what they've got is a joke is all I'm saying.

They tested higher framerates 40 years ago. And they didnt like the esthetics of it.

When something look more real, it also look more fake.

There is a reason why most of the worlds best directors and DPs dont want to use higher framerates. And its not economics.

Interesting post. Shooting and projecting widely distributed mass-market movies at ultra-high resolutions and frame rates – if it isn’t economically feasible today, it certainly will be in the next decade. (Running a 2160p theatrical digital projector at 60fps shouldn’t present any problems, tomorrow’s hard drives will make the extra data storage cheap enough to be a non-issue and the cameras (digital ones) already exist.)

The real question is the aesthestic one of whether filmmakers and viewers will want the hyper-realistic effect created by super high res, high frame rate video. Perhaps movies aren’t supposed to approximate the motion and detail of real life that closely, just as a comic strip isn’t intended or expected to look as lifelike as a still photograph. Indeed making a cartoon that realistic might destroy its whole point.

It’s unclear whether the economics of film production alone prevented 65mm/60fps movies from catching on. I think the next ten years will tell us one way or the other.

Great point and I was thinking about this as well after seeing District 9 in the theatre. I noticed it was captured with Red One cameras. What is going to happen and be the thoughts for movies then.

The irony is, lowering the brightness of those projector lamps does not in any way extend the life of the bulb. The manufacturers have proven this with testing, yet theaters do it to this day.

I can attest to that. Back in the early 90's, I worked as a projectionist for a major theater chain (I can't say which of course, but its initials are MANN...;)). We ran the bulbs at 1/3 brightness supposedly to "extend the life". Given the level of knowledge repeatedly displayed by management, it does not surprise me that this did no good.

I doubt very many releases get back to the camera negative. The "lost scenes" from Casablanca are about the only examples I can think of, and I'm think those were interpositives.
Just acquired a Blu-ray player (Sony PS3) and also bought the 2-disc A&E/BBC 1995 series "Pride and Prejudice" (P&P)on Blu-ray. In an extra featurette the colorist outlines and illustrates how they used the S16mm negatives for the BD release. The OP for the P&P review thread provided a link comparing print vs negative (http://www.avsforum.com/avs-vb/showpost.php?p=16161815&postcount=23) excerpts from the BR featurette. According to the colorist, what made it possible was a technology for eliminating the 'glitches' in the transfer process when negative splices are scanned. Even though it's S16mm, the color and detail seem remarkable on this release.

Understand, from a colorist's post several years back, that TV productions also use negatives more often. AVSer Glimmie, who works in this transfer area, pointed out that negatives for major features are too valuable for routine transfers. (Thought I'd read that John Lowry used negatives for the super-computer image massaging he performed during restoration of the James Bond series.) -- John

Just acquired a Blu-ray player (Sony PS3) and also bought the 2-disc A&E/BBC 1995 series "Pride and Prejudice" (P&P)on Blu-ray. In an extra featurette the colorist outlines and illustrates how they used the S16mm negatives for the BD release. The OP for the P&P review thread provided a link comparing print vs negative (http://www.avsforum.com/avs-vb/showpost.php?p=16161815&postcount=23) excerpts from the BR featurette. According to the colorist, what made it possible was a technology for eliminating the 'glitches' in the transfer process when negative splices are scanned. Even though it's S16mm, the color and detail seem remarkable on this release.

Understand, from a colorist's post several years back, that TV productions also use negatives more often. AVSer Glimmie, who works in this transfer area, pointed out that negatives for major features are too valuable for routine transfers. (Thought I'd read that John Lowry used negatives for the super-computer image massaging he performed during restoration of the James Bond series.) -- John

Excellent!!!

Should be required study for folks here, before being allowed to talk artist intent, film grain and so forth.

Somewhat matches my experiences in the world of photography

So, devil's advocate here for a second.
Let's say the best print they could find was several generations down the line, and grainier than intended.
The director thinks this version has so much grain that it's distracting, and chooses to use a little DNR to calm the grain, and can do this without losing any (or without losing very much) detail.
Is this an artistically valid process?
Steve W

Yes, but we are talking film restoration here, not the general case of mastering a film for HD.

Frankly, sometimes i think what we may preceive as film grain is not film grain, but simply artificial digital noise. Unfortunately, in trying to get rid of the digital noise, one can do more damage than good. I have dealt with this problem for over ten years in digital photography.

I would certainly hope that a $1,000,000+ Spirit 4k or whatever they use to scan motion picture film these days would be limited by the graininess of the film stock, not the noise performance of the scanner...

Well sure, I agree with you. But I doubt many directors, if economics weren't a factor and supply was plentiful and distribution and display weren't factors etc. etc., would choose 35mm/24fps over say 65/60. They work with what they've got, but what they've got is a joke is all I'm saying. It's an old argument that I've never understood that you seem to be making - that low res film is some kind of artistic high. It's not to me, it's junk, and the only reason we're stuck with it is the cheapo mentality that settled on it to begin with so long ago. Now it's entrenched and there's nothing anyone can or will do about it. But, as I said, and you didn't quote, in my original post, it still reigns supreme over anything digital has come up.

Right, and if Leonardo Da Vinci had access to a computer, he'd certainly have made the Mona Lisa in Photoshop. Because everybody knows that paint is a total cheapo joke, is all I'm saying. :rolleyes:

I would certainly hope that a $1,000,000+ Spirit 4k or whatever they use to scan motion picture film these days would be limited by the graininess of the film stock, not the noise performance of the scanner...

Modern film scanners and telecines have low noise. Problems can arise if the scan is not optimised shot by shot and later there is radical digital color correction on problematic shots. It can amplify (low level) noise to ugly proportions.

Right, and if Leonardo Da Vinci had access to a computer, he'd certainly have made the Mona Lisa in Photoshop. Because everybody knows that paint is a total cheapo joke, is all I'm saying. :rolleyes:

I also hope the guy would have had access to a tablet.;)

Yes, but we are talking film restoration here, not the general case of mastering a film for HD.

Okay, and...

If you don't care, then you're just saying a version of the old saw "I don't know art, but I know what I like." That's a euphemism for "I like crap and I don't care what anyone thinks about it."

When I was attending undergraduate studies, I had to take a class called "Art Appreciation", where the woefully misguided professor (IMO) spent an entire semester trying to make me see "art" in a canvas which consisted of random splotches and paint stains. Just because some certified wacko (again, IMO) named Jackson Pollack called it "art" when he spilled some paint and generally made a mess, does not make it so. Whereas some would elevate Pollack's "work" to something requiring talent, I would place it squarely in the category of "crap" (using your vernacular).

Place me firmly in the category that sees grain as the video equivalent of audio analog hiss: a side artifact of the recording medium. I am all for it's removal, assuming that the process does not take any fine detail along with it... and that's the catch-22: No one thus far has designed an algorithm that can accurately discern between what is noise and what is fine detail. So, we are left with the lesser of two evils: leave the grain and preserve the fine detail, or scrub it clean and take much of the detail along with it. I'll take what's behind door #1.

That being said, I firmly believe that the grain should remain intact for those who want it that way, and variable DNR be built into the display or player (just like EE should be left up entirely to the end user... and can be completely disabled if disired). That way, everyone is happy (actually, I'm sure someone will still find cause to complain).

Honestly, sometimes I get the feeling that 90% of the posters here would rather watch a pristine 1080p, no DNR or EE, artifact-free print of "Freddy got Fingered" instead of the current pressing of Gladiator. It seems more important to some that they can marvel at the detail and grain structure on the screen (imagining that they have a 35mm projector behind them) than to enjoy the story itself...

I've gotten several responses so I guess I owe the thread a reply.

My original post was to the subject of grain in 35mm film. I thought I made that clear when I offered larger negative film as a positive alternative. Anyone who wants to argue against larger negative film producing a more pleasing image when blown up to a large size needs to do some research into still film photography - you'd get laughed out of the room.

It was my mistake to lump negative size and frame rate together. If someone wants to argue that compromising in the direction of motion artifacts with a slow frame rate is the best idea for a standard, I won't argue with them. But all directors and DPs wouldn't agree of course. Directors are human, and humans have similar traits no matter what their job. One of those traits is doing what they do because that's how "it's done." My favorite for instance is depth of field - almost never seen in modern film, but occasionally it is. I wish the situation was the opposite. I refuse to believe that the overwhelming use of medium and long lenses is an artistic choice - it's just how it's done. For a beautiful (to my eye) example of the use of DOF, see the warehouse scenes in Ronin. Simply gorgeous ... and scoped to boot, which a lot of young punk "modern" directors would probably say is impossible.

Having this discussion at all is kind of sad, because of its futility. I think film is dead. Even if the physical medium survives, digitizing in post is killing the beauty as far as I'm concerned. I wonder how often on modern film sets there's a conversation between director and DP that ends with, "nah, don't worry, we'll fix it in DI." I recently say a Red One BD (can't remember which one) and wasn't impressed.

I don't want to be wowed buy the clarity of digital, I want to be soothed by the beauty of film. I just think it could be even more beautiful if beauty and not the dollar was the priority.

On the other hand, maybe I'm wrong. I remember the IMAX scenes in TDK and boy did they suck compared to the 35mm stuff.

When I was attending undergraduate studies, I had to take a class called "Art Appreciation", where the woefully misguided professor (IMO) spent an entire semester trying to make me see "art" in a canvas which consisted of random splotches and paint stains. Just because some certified wacko (again, IMO) named Jackson Pollack called it "art" when he spilled some paint and generally made a mess, does not make it so. Whereas some would elevate Pollack's "work" to something requiring talent, I would place it squarely in the category of "crap" (using your vernacular).

Place me firmly in the category that sees grain as the video equivalent of audio analog hiss: a side artifact of the recording medium. I am all for it's removal, assuming that the process does not take any fine detail along with it... and that's the catch-22: No one thus far has designed an algorithm that can accurately discern between what is noise and what is fine detail. So, we are left with the lesser of two evils: leave the grain and preserve the fine detail, or scrub it clean and take much of the detail along with it. I'll take what's behind door #1.

That being said, I firmly believe that the grain should remain intact for those who want it that way, and variable DNR be built into the display or player (just like EE should be left up entirely to the end user... and can be completely disabled if disired). That way, everyone is happy (actually, I'm sure someone will still find cause to complain).

Honestly, sometimes I get the feeling that 90% of the posters here would rather watch a pristine 1080p, no DNR or EE, artifact-free print of "Freddy got Fingered" instead of the current pressing of Gladiator. It seems more important to some that they can marvel at the detail and grain structure on the screen (imagining that they have a 35mm projector behind them) than to enjoy the story itself...

You are wrong.
That isn't an opinion - it's a fact. Most issues have a right and wrong side.

You are wrong.
That isn't an opinion - it's a fact. Most issues have a right and wrong side.

Um... exactly what part of my post is "wrong"? Are you saying that Pollack was not a wacko...? ;)

Based on these shots, it appears Landis kept his word.

http://www.blu-raydefinition.com/reviews/an-american-werewolf-in-london-blu-ray-review.html

Based on these shots, it appears Landis kept his word.

http://www.blu-raydefinition.com/reviews/an-american-werewolf-in-london-blu-ray-review.htmlAnd the online reviewer loathes the grainy results. What a surprise.

I've never been fan of artificial grain, film grain is fine but I don't see the point of removeing it and then re-adding it digitally. Mouse over comparisons with the HD-DVD will be interesting.

I've never been fan of artificial grain, film grain is fine but I don't see the point of removeing it and then re-adding it digitally.Do we know that's what was done here, or did Landis simply tell them not to remove the grain that was inherent in the source?

The caps look great, cannot wait till tuesday :)

http://www.blu-raydefinition.com/wordpress/wp-content/uploads/2009/09/BDDefinitionamericanwerewolfinlondon-c1080.jpg
http://www.blu-raydefinition.com/wordpress/wp-content/uploads/2009/09/BDDefinitionamericanwerewolfinlondon-f1080.jpg
http://www.blu-raydefinition.com/wordpress/wp-content/uploads/2009/09/BDDefinitionamericanwerewolfinlondon-j1080.jpg

hmm...

Do we know that's what was done here, or did Landis simply tell them not to remove the grain that was inherent in the source?

Landis was talking about Animal House, and he told them not to remove the grain that was in it.

The caps look great, cannot wait till tuesday :)

Hey, it looks like a movie! Count me in.:D

I'm feeling good about this release. It seems to have been spared the "let's change it to look like shiny new HD video" treatment, and apparently Landis had a lot to do with that. I wish more directors would demand that the grain be left alone. It will be annoying, however, to see the usual online reviewers whine and give it a very low picture score because the grain wasn't cleaned up.

Right, and if Leonardo Da Vinci had access to a computer, he'd certainly have made the Mona Lisa in Photoshop.

Photoshop is for photos. Nah, he would use a wacom tablet and Painter.:D

Looks pretty poor to me. Is this supposed to be a new transfer? Doesn't look like it... way too soft and blurry.

I would certainly hope that a $1,000,000+ Spirit 4k or whatever they use to scan motion picture film these days would be limited by the graininess of the film stock, not the noise performance of the scanner...

Hope all u wish.

problem is with the sensor. The top sensors are backed up by the digital processor which certainly uses some type of digital noise reduction system. IT MUST HAVE this screen or DNR.

Pixels react to photons and the problem is that the more pixels squeezed into an area, the more noise potential exists from pixel interactions with the photons. It is a matter of physics and protons.

indeed the technology exists to make a 300 megapixel sensor at the size of a 35mm sensor but the images would require massive computer interpolations and digital noise reduction to produce something of a useable image---at the level of the technology used in satellite photos.

At the same time, the more pixels permit a higher resolution but at a cost of more potential noise.

And those shotos posted above by roover 2002, are clearly showing DIGITAL NOISE, as much if not far more, than film grain.

Seen too much of it from top of the line Canons and Nikon cameras, shots that were taken with the highest quality of lenses of various subjects (outdoor and indoor with excellent lighting) and at various ISO in real life--ie not attempting to reproduce an image from a negative

Here is a very basic article discussing the nature of the problem towards the bottom of the article---
http://www.cambridgeincolour.com/tutorials/digital-camera-pixel.htm


Edit--I forgot to mention that the top of the line 7,200 dollar 22 megapxel Canon 1ds markIII (and older mrkII) has an internal lens right in front of the sensor that softens the focus ever so slightly to cut down on noise issues-------

Modern film scanners and telecines have low noise. Problems can arise if the scan is not optimised shot by shot and later there is radical digital color correction on problematic shots. It can amplify (low level) noise to ugly proportions.

And any of the higher resolution sensors do produce noise ---and the sensor processor is designed to hard wire out or reduce down to low noise.


http://www.cambridgeincolour.com/tutorials/digital-camera-pixel.htm


Landis was talking about Animal House, and he told them not to remove the grain that was in it.

The quote says he told them to put it back, whatever that means.

And the online reviewer loathes the grainy results. What a surprise.

Well that´s because he doesn´t know jack **** about what low budget early 80:s films actually looks like. Reviewers with such zero knowledge should be banned from reviewing anything else than shiny Pixar movies ;)

Looks pretty poor to me. Is this supposed to be a new transfer? Doesn't look like it... way too soft and blurry.

Once again! This will NEVER be as sharp and shiny as a big budget movie. This look is TYPICAL of the kind of filmstock, budget and vintage.

I've never been fan of artificial grain, film grain is fine but I don't see the point of removeing it and then re-adding it digitally.

Why would they add digital grain when they could just take a step back towards the source BEFORE they added DNR :rolleyes:

The caps look great, cannot wait till tuesday :)

+1!! :)

Finally we have the original look of the film.
And I´m happy as hell since I feared a DNR nightmare turning it into VHS quality ;)

The quote says he told them to put it back, whatever that means.

It ment he told them to turn off the DNR function. Hence leaving all the original grain intact.

Once again! This will NEVER be as sharp and shiny as a big budget movie. This look is TYPICAL of the kind of filmstock, budget and vintage.
Well, feel free to draw that conclusion, but I love a natural filmic looking blu-ray as much as anyone and that look is pure digital to my eyes. Grain simply doesn't look like that.

Why would they add digital grain when they could just take a step back towards the source BEFORE they added DNR :rolleyes:

I agree..:)

Is it the same transfer as on the HD-DVD? The above captures look similar to these. (http://www.caps-a-holic.com/hd_vergleiche/index.php?vergleich=american_werewolf)

Place me firmly in the category that sees grain as the video equivalent of audio analog hiss: a side artifact of the recording medium.

It is, in a way. But that's not all. The grain implements the detail. It also simply looks different than the corresponding totally clean picture (as the audio with some tape hiss sounds different from a totally clean version) and as such it always has the potential to have a different effect on the viewer which can be important concerning the intended effect by the film makers. As such it's not just an artifact to get rid of by default but a property of the original picture which has to be evaluated on a case by case basis as to the suitability/necessity of tampering with it or not. Neither the assumption that it's a defect and must go or it's totally required and must stay 100% all the time is per default correct. What is always correct is that the historically accurate version of the picture has specific amounts of grain and a historically accurate digital version of the film must have the same amounts of grain. Whether any deviations from this are legitimate is for the film makers to say.

I am all for it's removal, assuming that the process does not take any fine detail along with it... and that's the catch-22: No one thus far has designed an algorithm that can accurately discern between what is noise and what is fine detail. So, we are left with the lesser of two evils: leave the grain and preserve the fine detail, or scrub it clean and take much of the detail along with it. I'll take what's behind door #1.

There are now digital tools that reduce/remove grain and leave detail pretty much intact. The amount of side effects is very low on the best tools too. That the best tools are very often not used is also no secret looking at what gets released and how it looks.

That being said, I firmly believe that the grain should remain intact for those who want it that way, and variable DNR be built into the display or player (just like EE should be left up entirely to the end user... and can be completely disabled if disired). That way, everyone is happy (actually, I'm sure someone will still find cause to complain).

Fully agree. That would be a win win solution.

Edit--I forgot to mention that the top of the line 7,200 dollar 22 megapxel Canon 1ds markIII (and older mrkII) has an internal lens right in front of the sensor that softens the focus ever so slightly to cut down on noise issues-------
Comparing scanners and still cameras is not so straightforward. The best still cameras today have hardly any noise issues with most kinds of shots as long as you don't push the pixel number too high. 2 Mio. HD pixels is actually very low for a still camera. You can have 12 Mio and use ISO 1600 and have quite clean pictures. Still cameras also have exposure times dictated by subject movement, available light, desired DOF. A scanner must focus one plane, can expose as long as needed (unless it needs to be real time or faster) and has standard illumination inside. There are no technical reasons why a state of the art scanner should have noise issues that are dominating film grain when producing a 1080p picture.

Well, feel free to draw that conclusion, but I love a natural filmic looking blu-ray as much as anyone and that look is pure digital to my eyes. Grain simply doesn't look like that.

That IS what grain looks like, yes it´s heavy, but that truly looks like real natural grain. If you want to see what digital grain looks like, watch 300.

At the very least the image has been sharpened so the grain isn't exactly natural.

Comparing scanners and still cameras is not so straightforward. The best still cameras today have hardly any noise issues with most kinds of shots as long as you don't push the pixel number too high. 2 Mio. HD pixels is actually very low for a still camera. You can have 12 Mio and use ISO 1600 and have quite clean pictures. Still cameras also have exposure times dictated by subject movement, available light, desired DOF. A scanner must focus one plane, can expose as long as needed (unless it needs to be real time or faster) and has standard illumination inside. There are no technical reasons why a state of the art scanner should have noise issues that are dominating film grain when producing a 1080p picture.

Unfortunately, all sophisticated sensors and processors have some form of digital noise reduction hard wired into them.

How much is hard to say, as those things are a closely held trade secret........but when canon feels compelled to a put a softening lens right over the sensor to cut down on noise issues (also related to moire issues, but still the same thing---and probably represents what some are seeing in the first portion of Braveheart, where they complain about the jaggies), that should tell you something, if reading the cambridge article does not

(Do not get me wrong--I still beleive 100 iso film, properly processed and then prints developed thorough pure chemical processing with highest quality paper, still produces the best images, and i think the same for film. I am a 'analogue" film lover--remebr the 70mm film of Lawerence of Arabia!!!!yes oh yes!!!)

BTW--many folks will say, look negatives do not produce any better prints than pure digital and have the prints to prove it. I do not disagree, because what they forget, is when you have NOW have negatives reproduced, the machines scan the negative into digital form, and then it is printed from the digital scan image, using the same dot process as a pure digital image, rather than the pure process of light and chemicals. The latter still produces the best images on paper, by far. There is a smoothness, yet even more detailed image that is produced, than produced by any of the curent printers using digital in some form or fashion..

But it also rapidly becoming a lost art, heckfire, trying finding the highest quality paper to produce black and white images, it just is not being made....like it was a mere five years ago

The quote says he told them to put it back, whatever that means.

Thats a worrying statement to me.

Did he look at the finished article and then say put the grain back in and so they simply added artifical grain using the computer ?

That quote he made is open to interpretation.

If they did remove the grain and soften the image too much then just added the grain back in using a computer then i imagine screenshots would show a less than detailed image but with additional fake although authentic grain.

You sound like an expert on photography and photographic processes and we don't get enough of them at this forum so what do you think he meant by his statement ?

Looking at the DVD Beaver pics, it looks very close to the HD-DVD (watched it this afternoon). In fact I would have assumed it came from exactly the same master and it's just a new encode.

That quote he made is open to interpretation.


Why not just take the mothballs off the National Lampoon's Animal House HD DVD and find out?

Thats a worrying statement to me.

Did he look at the finished article and then say put the grain back in and so they simply added artifical grain using the computer ?

That quote he made is open to interpretation.

If they did remove the grain and soften the image too much then just added the grain back in using a computer then i imagine screenshots would show a less than detailed image but with additional fake although authentic grain.

You sound like an expert on photography and photographic processes and we don't get enough of them at this forum so what do you think he meant by his statement ?

I would hope that they had an untouched master (well as close to "untouched" as anything that is digital, because it is all being processed in form or another inside a computer) and just went back to that.

Nothing worse than adding artificial grain into a photo---beleive it or not, one can do exactly that with Photoshop, but whenever I have tried it, my response has been to turn a possible excellent photo into the comment in Predator, you are one UGLY mo.........and while i liked much of the camera work work in 300, it still struck me as less than necessary to have some of the processing

Prints made for theater distribution are like slides or or paper prints of a negative. In the negative, the colors and light are different, but in the "print", the colors are what you see on the screen. So the grain one might think they are seeing is from the material used to make the print or composing the print and not necessarily what was on the original negative.

http://www.pictureshowman.com/articles_restprev_negsprints.cfm
is a good discussion.

Also go back to the earlier post about this excellent discussion about grain, different colors in a print vs. the negative of a film:
http://vimeo.com/3869589


Bottom line, there is all sorts of noise reduction and EE when making a dvd, esp with blu ray, some good and some very bad.

Personally I think whoever is making it should not be some art expert who thinks he needs to make it better, capture the artist intent or whatever but some anal tech/nerd type who is simply trying to produce an image as close to the original negative as possible.

Bear in mind that sometimes all one has is the print of the movie and that can vary greatly in quality.

Here is another excellent article about comparing film and digital photos and why all the comparions are usually very inaccurate.
http://kenrockwell.com/tech/filmdig.htm

it is written by someone who makes a living doing movies in Hollywood

HVisone, as interesting as your comments are, still photography and motion picture photography are completely different disciplines with completely different aesthetics. What may be considered a "flaw" in one can be desireable in the other. You can't directly compare them, or try to impose the rules of still photography to a motion picture.

Unfortunately, all sophisticated sensors and processors have some form of digital noise reduction hard wired into them.
How much is hard to say, as those things are a closely held trade secret........but when canon feels compelled to a put a softening lens right over the sensor to cut down on noise issues (also related to moire issues, but still the same thing---and probably represents what some are seeing in the first portion of Braveheart, where they complain about the jaggies), that should tell you something, if reading the cambridge article does not

This element is an optical low pass filter and not related to sensor noise but aliasing which is something else. Almost all digital still cameras have it. There is no need for noise reduction with still cameras where pixel size is big enough for good SNR and no high ISO is required. The megapixel monsters with small sensors on the other hand...

Thats a worrying statement to me.

Did he look at the finished article and then say put the grain back in and so they simply added artifical grain using the computer ?

That quote he made is open to interpretation.

If they did remove the grain and soften the image too much then just added the grain back in using a computer then i imagine screenshots would show a less than detailed image but with additional fake although authentic grain.

You sound like an expert on photography and photographic processes and we don't get enough of them at this forum so what do you think he meant by his statement ?

Put it back would say to me that it was there and taken off, so it was taken back a step to where Landis was happy

HVisone, as interesting as your comments are, still photography and motion picture photography are completely different disciplines with completely different aesthetics. What may be considered a "flaw" in one can be desireable in the other. You can't directly compare them, or try to impose the rules of still photography to a motion picture.

Principles are almost identical when trying to reproduce any film-movie or still- into a digital format. I was not discussing aesthetic differences as the primary subject but sort of meandered.

Natural film grain bothers me not in the least, but noise created by the digital process does-in stills or movie.

Seems to be a lot of confusion here as to the question of digital noise reduction. DNR is always present and a reduction in a DNR filter that permits "grain" to appear is not always the "grain" to be found in the original, but is just simply digital noise appearing because of less filtering.

Artifical noise in the 300 in the theater film print, added a certain effect that some loved, but it did not make much difference to me, and I could have done without it. In the blu ray, the noise is even more apparent to me as noise rather than the intended film grain, that was intended apparently to be real film grain, but just looked fake to me.

This element is an optical low pass filter and not related to sensor noise but aliasing which is something else. Almost all digital still cameras have it. There is no need for noise reduction with still cameras where pixel size is big enough for good SNR and no high ISO is required. The megapixel monsters with small sensors on the other hand...

aliasing, jaggies, moire, and digital noise are all digital problems created by the interaction between the light photons and the sensor---it is not something created by the later signal processing.....the low pass filter or lens reduces all of these, but at the loss of sharpness of image, as does the digital processing inside the computer chip or sensor processor attached to the sensor.

if u do not want those jaggies at the beggining of braveheart, then increase processing but at the risk of decreasing image quality.

aliasing, jaggies, moire, and digital noise are all digital problems created by the interaction between the light photons and the sensor---it is not something created by the later signal processing.....the low pass filter or lens reduces all of these, but at the loss of sharpness of image, as does the digital processing inside the computer chip or sensor processor attached to the sensor..
Sorry, but the optical low pass filter is not about noise in the sensor, it's about avoiding (obvious) aliasing. Please read up on sampling theorem and Nyquist. Even a very strong low pass filter will not help if there is not enough light and the pixels are too small to gather enough photons or there are other sensor internal issues with reading out data without noise pollution. The sensors have built in hardware measures to reduce noise and there can be external processing chips as well to do noise reduction but this is not what we talk about when we see DNR applied to film transfers. These transfers when made on state of the art equipment have no electronic noise that would even come close to the level of grain in the film itself. These sensors are practically noisefree compared to motion picture film.

Sorry, but the optical low pass filter is not about noise in the sensor, it's about avoiding (obvious) aliasing. Please read up on sampling theorem and Nyquist. Even a very strong low pass filter will not help if there is not enough light and the pixels are too small to gather enough photons or there are other sensor internal issues with reading out data without noise pollution. The sensors have built in hardware measures to reduce noise and there can be external processing chips as well to do noise reduction but this is not what we talk about when we see DNR applied to film transfers. These transfers when made on state of the art equipment have no electronic noise that would even come close to the level of grain in the film itself. These sensors are practically noisefree compared to motion picture film.

dont need to read up, having designed (along with a cousin) early processors and later contributed to the canon processors' use of certain processes used to filter out noise and other artifacts, when Canon purchased certain rights. It was that process of trying to beat out film in terms of PQ, that reignited my interest in non-digital still photography.

The reason those sensors are practically "noise free" is due to the use of DNR, which goes hand in hand with elimination of the other artifacts emitted from the sensor. Grain is seldom as obvious on film negatives as it is seen on some blu rays (although grain becomes far more obvious on some prints of film, due to the process of developing the theater print for mass distribution).

BTW--the presence of the minute imperfections known as grain, along with the minute variations of color that are presented by the molecular structure of the chemical compositions that give a film negative its incredible colors over digital, are read or picked up by the pixel sensors and do contribute to the noise factor.

These variations are on a far smaller scale (basically molecular) compared to that of digital given the number of pixels present in the sensor. While some of my prior associates would hate me when i say it, until some very new, yet to be developed technology comes along (at least any that I have heard of, as who knows what is going on in secret labs)------digital falls far short of pure film for that reason alone, although there are other factors as well.

yeah my former associates all want to say digital pq is better than film, else why spend $7k for a DS1 mark III...........

To this day, I regard myself very lucky to have seen a special screening of Lawrence in the mid 80's of the 70 mm film, reproduced on massive film rolls with a projector bigger than a volkswagen, and have never seen anything like it since.

Pure digital sources, such as a movie like Cars, lack these minute variations and texture, hence do not have the noise issues such as found in film prints reproduced on digital dvds


(and given some of the digital processing used in reproducing the theater prints, (or in making the original movie) more noise rather than grain from the negative should be expected----)

In April of 2008 I went to a theatrical showing of "...Werewolf..." at a local NYC theater with John Landis in attendance. He told us back then essentially the same thing: he made sure the people working on high-def masters of his early movies (specifically "Animal House" and "...Werewolf...") made them less-than-stellar because he didn't want these flicks looking sharp and pretty since it would take away from the 'grime' and dirty look he wanted them to have. Well sir, mission accomplished! Here's a couple of pics I snapped of my 47" Olevia HDTV showing the HD-DVD and DVD versions of "...Werewolf..." (combo disc). Based on the Blu-ray pics posted on this thread it looks like, even watered down, the high-def remaster is a dramatic-enough improvement over SD to warrant interest from fans:

http://i169.photobucket.com/albums/u238/dad1153/DSC00167.jpg
http://i169.photobucket.com/albums/u238/dad1153/DSC00169.jpg

Natural film grain bothers me not in the least, but noise created by the digital process does-in stills or movie.
Sorry but the notion that digital noise would be more visible than film grain is nonsensical. I suppose that scanning at too low a resolution would cause some degree of aliasing, but scanning isn't a particular challenge as far as noise goes, especially if price and scanning time is no concern: you can oversample spatially, you can oversample temporally, etc.

Sorry but the notion that digital noise would be more visible than film grain is nonsensical. I suppose that scanning at too low a resolution would cause some degree of aliasing, but scanning isn't a particular challenge as far as noise goes, especially if price and scanning time is no concern: you can oversample spatially, you can oversample temporally, etc.

Yes, now you know about my specialty, sampling along with some other techniques, and how it is used to enhance a digital image and remove noise.

here are some easy references to look at:

http://www.scribd.com/doc/16404857/Digital-Image-Processing-07-Image-Restoration-Noise-Removal


http://campratty.com/4tooltours/kz/pdf/DigitalCameraNoiseRemovalFilter2006.pdf


the coments about the random nature is actually not accurate, as it is influenced by the colors and bright/dark areas on an image as noted else where in the articles

If you were to see the direct input from complex sensors, you would be stunned at the lack of clarity and the amount of artifacts present, and would begin to understand the complex relationship between process and pixels to bring to you the image that you see.

So sorry to be raining on the "visible film grain (sic) in my blu ray dvd as being more accurate parade" when you are seeing noise and some small amount of grain......

if u do not want those jaggies at the beggining of braveheart, then increase processing but at the risk of decreasing image quality.
In the case of BRAVEHEART, I believe you're completely off the mark. The BRAVEHEART issue being discussed is one which affects the film in its entirety up until the 22-minute mark, at which point it disappears completely. It has nothing to do with there being too much noise, too much detail or anything like that. It has the look of video that has been captured at a lower resolution and then scaled up to 1920x1080 using some form of pixel resize (or "nearest neighbour", as Photoshop calls it), although only the vertical plane appears to be affected. Horizontally, the image appears to be full resolution. In fact, it's actually highly reminiscent of the issue that affected Warner's so-called 1080i releases back in the early days of Blu-ray and HD DVD, although the problem with those releases was that the HORIZONTAL rather than the vertical resolution of the source material was substandard. The remaining 155 minutes of the film essentially look "perfect", with more detail than you can shake a stick at and no evidence of crude scaling.

An example of the issue at its most pronounced (look at the jaggedness on the mountainside):

http://www.landofwhimsy.com/hdcaptures/braveheart3.jpg

Compared to a later shot that is completely unaffected:

http://www.landofwhimsy.com/hdcaptures/braveheart12.jpg

In the case of BRAVEHEART, I believe you're completely off the mark. The BRAVEHEART issue being discussed is one which affects the film in its entirety up until the 22-minute mark, at which point it disappears completely. It has nothing to do with there being too much noise, too much detail or anything like that. It has the look of video that has been captured at a lower resolution and then scaled up to 1920x1080 using some form of pixel resize (or "nearest neighbour", as Photoshop calls it), although only the vertical plane appears to be affected. Horizontally, the image appears to be full resolution. In fact, it's actually highly reminiscent of the issue that affected Warner's so-called 1080i releases back in the early days of Blu-ray and HD DVD, although the problem with those releases was that the HORIZONTAL rather than the vertical resolution of the source material was substandard. The remaining 155 minutes of the film essentially look "perfect", with more detail than you can shake a stick at and no evidence of crude scaling.

An example of the issue at its most pronounced (look at the jaggedness on the mountainside):

http://www.landofwhimsy.com/hdcaptures/braveheart3.jpg

Compared to a later shot that is completely unaffected:

http://www.landofwhimsy.com/hdcaptures/braveheart12.jpg

Yes after a quick look you may or may not be exactly correct and i be exactly wrong.

The problem is that some images when digitally processed with an already captured image, and are "upscaled" to a higher resolution this is a potential artifact that appears as a result. So you may be absolutely correct!!!


OTOH, the same can occur with a fault with the processor that is dealing with problems originating from the sensor, such that the noise reduction hardwired into the process is not working correctly, producing the same result, especially where it may involve sampling processes (which is how many of the better processors or photoshop software increases "resolution" or "detail" to avoid the halo effect and other problems.)

If the halo effect was clearly present, then i would agree completely with you, but its absence does not mean you are wrong.

As to which it is, i can not say based on information i have.

And not having seen Braveheart, does not help. However, i am very puzzled as to why the jaggies suddenly disappear, given the quality everyone otherwise has given to this Blu-ray. Must be some sort of quality control issue, because if the problem was later "cured", why else would it be present in the early scenes and not later?

But I have seen Gladiator, a movie i love dearly, (sorry for you more intellectual types, but i have my quilty pleasures) courtesy of a blockbuster early screw up release......and i am outraged and very disappointed by it-my expression of outrage lead to certain issues in the other thread!!!

I have a region2 superbit copy, and it is better and easier to watch than the blu ray.

Makes me wonder what has been all the problems with Dances with Wolves, if there are no good copies left or negatives that they can use, and the choice is between lousy quality per gladiator (trying to assume that is the problem rather than incompentcy and greed), or being forced to use too much DNR and so forth being necessary for a clean up, so that you get something like Patton. rather than braveheart

clearly having scheduled it for release and its sudden cancellation, suggests to me that somebody must have looked at what was getting ready to go out the door, and said NO,

they should have done the same with Gladiator

and maybe with Braveheart, they should have said, opps small problem with those 22 minutues, short delay while we fix it and then totally set the standard for perfection

(you being in Scotland, I would guess you have access to the region2 super bit??)

Yes after a quick look you may or may not be exactly correct and i be exactly wrong.

The problem is that some images when digitally processed with an already captured image, and are "upscaled" to a higher resolution this is a potential artifact that appears as a result. So you may be absolutely correct!!!


OTOH, the same can occur with a fault with the processor that is dealing with problems originating from the sensor, such that the noise reduction hardwired into the process is not working correctly, producing the same result, especially where it may involve sampling processes (which is how many of the better processors or photoshop software increases "resolution" or "detail" to avoid the halo effect and other problems.)

If the halo effect was clearly present, then i would agree completely with you, but its absence does not mean you are wrong.

As to which it is, i can not say based on information i have.

And not having seen Braveheart, does not help. However, i am very puzzled as to why the jaggies suddenly disappear, given the quality everyone otherwise has given to this Blu-ray. Must be some sort of quality control issue, because if the problem was later "cured", why else would it be present in the early scenes and not later?

But I have seen Gladiator, a movie i love dearly, (sorry for you more intellectual types, but i have my quilty pleasures) courtesy of a blockbuster early screw up release......and i am outraged and very disappointed by it-my expression of outrage lead to certain issues in the other thread!!!

I have a region2 superbit copy, and it is better and easier to watch than the blu ray.

Makes me wonder what has been all the problems with Dances with Wolves, if there are no good copies left or negatives that they can use, and the choice is between lousy quality per gladiator (trying to assume that is the problem rather than incompentcy and greed), or being forced to use too much DNR and so forth being necessary for a clean up, so that you get something like Patton. rather than braveheart

clearly having scheduled it for release and its sudden cancellation, suggests to me that somebody must have looked at what was getting ready to go out the door, and said NO,

they should have done the same with Gladiator

and maybe with Braveheart, they should have said, opps small problem with those 22 minutues, short delay while we fix it and then totally set the standard for perfection

(you being in Scotland, I would guess you have access to the region2 super bit??)
I'd certainly be very interested to hear what you make of BRAVEHEART if and when you see the BD. The issue is just weird - bang on the 22-minute mark, not a second before or after, the whole thing just clears up permanently. It happens during a chapter break, which made me wonder if it was an encoding issue. However, in the documentary on the second disc, any clips shown of the film before the 22-minute mark are also affected, which makes me think it's more likely that the problem is "baked" into the master.

I could probably get my hands on the Superbit copy of GLADIATOR, and I'm actually tempted to, as it seems to be the best-looking standard definition copy out there (plus it has a 1.5 Mbps DTS track), but it's been OOP for years (Sony's deal to distribute Universal titles on DVD in Europe expired in 2002, I believe), and I don't fancy forking out too much cash for it.

I'd certainly be very interested to hear what you make of BRAVEHEART if and when you see the BD. The issue is just weird - bang on the 22-minute mark, not a second before or after, the whole thing just clears up permanently. It happens during a chapter break, which made me wonder if it was an encoding issue. However, in the documentary on the second disc, any clips shown of the film before the 22-minute mark are also affected, which makes me think it's more likely that the problem is "baked" into the master.

it.

Then assumig that the original film of the 22 minutes was in the same shape as the film used for the later portion, they should have redone the first 22 minutes, used the same settings as the later portion and spliced in the revised portion. Might have been a delay, might have been an extra expense, but overall, it seems a shame to not do it.....

bang on the 22-minute mark, not a second before or after, the whole thing just clears up permanently.

I haven't read the Gladiator thread so this may have already been mentioned: a quick googling suggests that standard film reels are around 22 minutes in length. If the film was split into reel-sized chunks at some point during the transfer and the first chunk was not processed correctly, that could explain why the image changes at such a specific time.

I haven't read the Gladiator thread so this may have already been mentioned: a quick googling suggests that standard film reels are around 22 minutes in length. If the film was split into reel-sized chunks at some point during the transfer and the first chunk was not processed correctly, that could explain why the image changes at such a specific time.
That certainly sounds like a feasible explanation. Perhaps it means that whoever was scanning the materials did the first reel incorrectly, noticed the mistake and corrected it for the remaining reels, but couldn't be bothered to go back and redo the first reel?

PS. The issue was discussed in the BRAVEHEART thread, and I think that theory was suggested by someone, but there was some confusion as to the length of the average reel of film.

Do we know if this is the same transfer as the HD-DVD ?

Do we know if this is the same transfer as the HD-DVD ?
I'm going to dig out my HD DVD right now and see if I can match any of the captures.

EDIT: Here we go.

Blu-ray:

http://www.blu-raydefinition.com/wordpress/wp-content/uploads/2009/09/BDDefinitionamericanwerewolfinlondon-c1080.jpg

HD DVD:

http://www.landofwhimsy.com/funbag/werewolfhd.jpg

Barring a vertical shift of a few pixels, they appear to be pretty much identical. It's definitely the same fairly mediocre-looking master. So much for a new transfer! :rolleyes: (We WERE promised one, weren't we?) Glad I didn't bother to "upgrade".

Oh thats odd, I still think it looks good for its age/budget

HD-DVD port so it has the same EE as well.

dont need to read up, having designed (along with a cousin) early processors and later contributed to the canon processors' use of certain processes used to filter out noise and other artifacts, when Canon purchased certain rights. It was that process of trying to beat out film in terms of PQ, that reignited my interest in non-digital still photography.

Then why do you insist optical low pass filters are for noise reduction??

The reason those sensors are practically "noise free" is due to the use of DNR, which goes hand in hand with elimination of the other artifacts emitted from the sensor.

The sensors are built to avoid/reduce noise, yes, but this is not the same as the DNR we talk about in video transfers. I have a state of the art still camera and the extra DNR (which is the kind comparable to intraframe filtering one can do on video) can be turned off completely. It is simply not needed at lower ISOs. It will only degrade picture quality. The images are (far) cleaner than 35mm film in the ISO 200 to maybe 1000 range. Essentially noisefree. After that slight noise starts to creep in. ISO 1600 is still quite clean. ISO 6400 is noticeably noisy. Similar technology is used in the Red One camera. When there is enough light and proper exposure there is no noise problem. They use no DNR of the video kind in camera. Film grain on grainy 35mm footage is way above the noise level of these sensors when they operate in their normal range. We are not comparing large sheet film here with sensors but 35mm.

Grain is seldom as obvious on film negatives as it is seen on some blu rays (although grain becomes far more obvious on some prints of film, due to the process of developing the theater print for mass distribution).

What you see on Blu Rays needs to be analysed case by case. What was scanned? How was it scanned? What processing was applied to it? How was it compressed? Each step can distort grain and make it more obvious than it should be. But there is not a general problem with modern sensors that they are too noisy to deal with 35mm film. The problems are elsewhere, if there are any. Modern DI data is not full of sensor noise that distorts grain a lot. The noise from grain is (much) larger than the sensor noise when the sensor operates within optimal range as it should in high quality scans. Grain can be very large and obvious indeed if the film has higher ISO in low light conditions.

BTW--the presence of the minute imperfections known as grain, along with the minute variations of color that are presented by the molecular structure of the chemical compositions that give a film negative its incredible colors over digital, are read or picked up by the pixel sensors and do contribute to the noise factor.

Quantisation noise is a well known issue in theory. But it's not a problem with todays sensors compared to the noise level inherent in 35mm footage.

These variations are on a far smaller scale (basically molecular) compared to that of digital given the number of pixels present in the sensor. While some of my prior associates would hate me when i say it, until some very new, yet to be developed technology comes along (at least any that I have heard of, as who knows what is going on in secret labs)------digital falls far short of pure film for that reason alone, although there are other factors as well.

Variations on a molecular level are irrelevant as they are invisible to the eye when watching projected film, and not related to perceivable real world detail.

To this day, I regard myself very lucky to have seen a special screening of Lawrence in the mid 80's of the 70 mm film, reproduced on massive film rolls with a projector bigger than a volkswagen, and have never seen anything like it since.

I have seen LOA in 70mm too. Together with many other 70mm films. Once there are 4K projectors with better black levels LOA from camera negative will look splendid in 4K and you won't miss anything compared to watching a print from duplicate negative (all you got now). It will look sharper, more detailed, have more depth and the colors will be quite similar (not always 100% identical, film is subtractive, digital cinema is additive),

First you are attempting to define noise as only being those specks that sometimes appear randonly, when in fact, the only way to look at it, is that noise is a range of issues, including moire, et al, that are all interelated. Without question, the noise (even as you attempt to define it) can be generated by the details of an image, the dark areas, the redish areas, the very bright areas and the borders in those areas.

Indeed, one problem is dealing with color blurring of a sharp edge between two areas having distinctively different color and light--hence the source of the "jagginess" problem in Braveheart, where the result of the computer interpolation to deal with color blurring at the edge, at some point in this process (what point-I do not know) produced an unpleasant image of "jages" instead of correcting the blurring of color (and because this problem originates in the distance--it probably should be blurred, because blurred background and foreground is what produces a depth of field in a picture and the program is identifying the problem as one of a "color blur" that needs correction, then improperly processes the image)

You also refer to the sensor, when in fact, the sensor is connected to a processer and it is the processor that permits the sensor to work. The sensor is just a surface of pixels that collect the light. When the pixels emitted the electical charge as a result of contact with a photon, the individual electrical signal goes to the processor.

The processor perfoms a function called demosaicking which is used to reconstruct a full color image from the incomplete color samples outputed from the sensor which uses a color filter array to reduce artifacts. Notwithstanding all the false advertizing to the contrary that the RAW image information is "unprocessed", even a "raw" image from a camera has undergone massive processing (including DNR) before being placed on a memory card. The difference is that the raw image undergoes further processing in the computer using software where the image can be further manipulated as to certain variables (and this manipulation is extremely necessary to be able to view the image).

For example, CCC pixels have only changed to become smaller and more numerous, but it is the changes in its accompanying processor, that makes the newer "sensors" to be "better".


Then why do you insist optical low pass filters are for noise reduction??
Because noise and other artifacts are reduced as a result of its use-making the job of the processor easier

The sensors are built to avoid/reduce noise, yes, but this is not the same as the DNR we talk about in video transfers. I have a state of the art still camera and the extra DNR (which is the kind comparable to intraframe filtering one can do on video) can be turned off completely. It is simply not needed at lower ISOs. It will only degrade picture quality. The images are (far) cleaner than 35mm film in the ISO 200 to maybe 1000 range. Essentially noisefree. After that slight noise starts to creep in. ISO 1600 is still quite clean. ISO 6400 is noticeably noisy. Similar technology is used in the Red One camera. When there is enough light and proper exposure there is no noise problem. They use no DNR of the video kind in camera. Film grain on grainy 35mm footage is way above the noise level of these sensors when they operate in their normal range. We are not comparing large sheet film here with sensors but 35mm.


This is so funny---what DNR you are turning off, is merely "extra DNR". If all DNR (including EE or sharpness) inside the processor was turned off, you would not recognize the image, except with extreme difficulty

What you see on Blu Rays needs to be analysed case by case. What was scanned? How was it scanned? What processing was applied to it? How was it compressed? Each step can distort grain and make it more obvious than it should be.
Now you are merely saying what I have been saying.."more obvious than it should be"....


But there is not a general problem with modern sensors that they are too noisy to deal with 35mm film. The problems are elsewhere, if there are any. Modern DI data is not full of sensor noise that distorts grain a lot. The noise from grain is (much) larger than the sensor noise when the sensor operates within optimal range as it should in high quality scans. Grain can be very large and obvious indeed if the film has higher ISO in low light conditions.

Again: Noise and other artifacts are reduced by the processor, not the sensor pixels

Quantisation noise is a well known issue in theory. But it's not a problem with todays sensors compared to the noise level inherent in 35mm footage.

Variations on a molecular level are irrelevant as they are invisible to the eye when watching projected film, and not related to perceivable real world detail.


Digital reproduces images as pixelations--tiny dots--the more tiny and numerous, the more smooth are the edges of the image and color changes. And this is perceivable on a concious or subconcious level. But it is still dots that can be easily blown up on a computer so you can see the individual pixels.

The ultimate "smoothness" would be an image composed of "dots" or "pixels" that are the size of molecules.......errr, duh.....traditional film......

It is interesting where one takes a print from a computer printer, using 6,000 dots per inch and the same image generated on a traditional basis, never letting digital be involved in the process, made from a pure chemical process, and in blind sampling, the majority of people were picking the chemical print over the the digital. i am aware of a certain large printer company that spent a massive amount of money to do such a study to prove their superiority, but when the results were becoming known, decided to drop it and call it a "trade secret"....


I have seen LOA in 70mm too. Together with many other 70mm films. Once there are 4K projectors with better black levels LOA from camera negative will look splendid in 4K and you won't miss anything compared to watching a print from duplicate negative (all you got now). It will look sharper, more detailed, have more depth and the colors will be quite similar (not always 100% identical, film is subtractive, digital cinema is additive),
If you knew what i was talking about, you would understand what you are probably watching is a 35mm print made from a 70mm negative (or from a 70mm print). I was watching a 70mm print made from the 70mm negatives--this requires a massive, expensive and very rare projector that will only play 70mm prints.
i have seen LOA projected by a 35mm projector. Compare to the 70 mm prints, the latter is a like a slap in the face as it is so stunning.

First you are attempting to define noise as only being those specks that sometimes appear randonly, when in fact, the only way to look at it, is that noise is a range of issues, including moire, et al, that are all interelated.

Noise in a sensor has several sources but noise is not just any image anomaly. That is not a valid definition of noise. Aliasing is not noise. Show me the definition of aliasing as noise in some standard reference source, please. I would like to see it.

Without question, the noise (even as you attempt to define it) can be generated by the details of an image, the dark areas, the redish areas, the very bright areas and the borders in those areas.
Indeed, one problem is dealing with color blurring of a sharp edge between two areas having distinctively different color and light--hence the source of the "jagginess" problem in Braveheart, where the result of the computer interpolation to deal with color blurring at the edge, at some point in this process (what point-I do not know) produced an unpleasant image of "jages" instead of correcting the blurring of color (and because this problem originates in the distance--it probably should be blurred, because blurred background and foreground is what produces a depth of field in a picture and the program is identifying the problem as one of a "color blur" that needs correction, then improperly processes the image)

(With DeNiro voice): Are you talking to me? If you are, please don't make a hopeless mess out of different image processing concepts and mix it all together into one big noise pot. That is not helpful. Jaggies are not noise and noise are not jaggies. The connection is lose at best.

You also refer to the sensor, when in fact, the sensor is connected to a processer and it is the processor that permits the sensor to work. The sensor is just a surface of pixels that collect the light. When the pixels emitted the electical charge as a result of contact with a photon, the individual electrical signal goes to the processor.
The processor perfoms a function called demosaicking which is used to reconstruct a full color image from the incomplete color samples outputed from the sensor which uses a color filter array to reduce artifacts. Notwithstanding all the false advertizing to the contrary that the RAW image information is "unprocessed", even a "raw" image from a camera has undergone massive processing (including DNR) before being placed on a memory card. The difference is that the raw image undergoes further processing in the computer using software where the image can be further manipulated as to certain variables (and this manipulation is extremely necessary to be able to view the image).
For example, CCC pixels have only changed to become smaller and more numerous, but it is the changes in its accompanying processor, that makes the newer "sensors" to be "better".

That's all very interesting and it's no news that there is a lot of processing going on before the images are saved, but in my still camera there is no DNR going on like for video. Sorry. Whatever they do it's very local to a pixel and does not have the ugly side effects of the usual video DNR filters. And one of the main reasons this camera is so good is not new super duper DNR on the sensor but simply larger pixels, the opposite of more and smaller pixels. And it's not a CCD, by the way. It's CMOS.

This is so funny---what DNR you are turning off, is merely "extra DNR". If all DNR (including EE or sharpness) inside the processor was turned off, you would not recognize the image, except with extreme difficulty

I never claimed that dead pixels are not removed or no debayering is performed, and no other processing is going on to compensate for sensor non linearities etc. There is no video style DNR going on in my camera. Or in the Red One. If you can't believe it you better discuss it with the people who designed them. If they tell you what exactly they do I'm all ears to listen. Good luck.

Digital reproduces images as pixelations--tiny dots--the more tiny and numerous, the more smooth are the edges of the image and color changes. And this is perceivable on a concious or subconcious level. But it is still dots that can be easily blown up on a computer so you can see the individual pixels.
The ultimate "smoothness" would be an image composed of "dots" or "pixels" that are the size of molecules.......errr, duh.....traditional film......

There is no need for ultimate smoothness, only as much as reaches the eye and can be seen for specific applications. Film is a high resolution binary system. Effective resolution is limited just like with digital systems which have in general pixels that are not binary. How many pixels you need for a given film image to describe the real world content on it depends on many things. It's far from infinite. For 35mm motion picture film 4K is in almost all cases enough.

If you knew what i was talking about, you would understand what you are probably watching is a 35mm print made from a 70mm negative (or from a 70mm print). I was watching a 70mm print made from the 70mm negatives--this requires a massive, expensive and very rare projector that will only play 70mm prints.
i have seen LOA projected by a 35mm projector. Compare to the 70 mm prints, the latter is a like a slap in the face as it is so stunning.
You are obviously not very familiar with me. Or you would know that I'm not the person who runs around talking about having seen LOA in 70mm when I saw a 35mm print. Uh.
I saw LOA on 70mm from a 70mm print from an internegative (because all 70mm prints of the restored version come from internegative). I have seen Baraka in 70mm, and many other films also, some of them blow ups and some from real 70mm originals. I have seen Vertigo from 70mm print and Vista Vision original. I have seen Showscan which is 70mm at higher frame rate. Countless IMAX films from IMAX masters and 35mm DRM versions. I have seen 4K projection of 35mm and 70mm originals. I have seen Cinerama (3 times 35mm). I think I have a good idea what film in various incarnations looks like. And digital origination in 4K as well (Dalsa, Red ONE). I have seen 8K Ultra Resolution projected. Anyway, LOA on a good 4K projector scanned from the original negative will look great. Sharper than any new 70mm prints now in circulation.

Hope all u wish.

problem is with the sensor. The top sensors are backed up by the digital processor which certainly uses some type of digital noise reduction system. IT MUST HAVE this screen or DNR.

Pixels react to photons and the problem is that the more pixels squeezed into an area, the more noise potential exists from pixel interactions with the photons. It is a matter of physics and protons.

indeed the technology exists to make a 300 megapixel sensor at the size of a 35mm sensor but the images would require massive computer interpolations and digital noise reduction to produce something of a useable image---at the level of the technology used in satellite photos.

At the same time, the more pixels permit a higher resolution but at a cost of more potential noise.

And those shotos posted above by roover 2002, are clearly showing DIGITAL NOISE, as much if not far more, than film grain.

Seen too much of it from top of the line Canons and Nikon cameras, shots that were taken with the highest quality of lenses of various subjects (outdoor and indoor with excellent lighting) and at various ISO in real life--ie not attempting to reproduce an image from a negative

Here is a very basic article discussing the nature of the problem towards the bottom of the article---
http://www.cambridgeincolour.com/tutorials/digital-camera-pixel.htm


Edit--I forgot to mention that the top of the line 7,200 dollar 22 megapxel Canon 1ds markIII (and older mrkII) has an internal lens right in front of the sensor that softens the focus ever so slightly to cut down on noise issues-------

Sort of off-base.
DSLR mostly have anti-aliasing filters in front of the sensors and this has NOTHING whatsoever to due with cutting down on noise. It is to remove moire and jaggy-stair step artifacting. You do need to apply a little bit of sharpening in post to bring the crispness back (but not with halos and way overdone like many of the disc mastering monkeys seem to play with).

Unfortunately, all sophisticated sensors and processors have some form of digital noise reduction hard wired into them.

How much is hard to say, as those things are a closely held trade secret........but when canon feels compelled to a put a softening lens right over the sensor to cut down on noise issues (also related to moire issues, but still the same thing---and probably represents what some are seeing in the first portion of Braveheart, where they complain about the jaggies), that should tell you something, if reading the cambridge article does not

(Do not get me wrong--I still beleive 100 iso film, properly processed and then prints developed thorough pure chemical processing with highest quality paper, still produces the best images, and i think the same for film. I am a 'analogue" film lover--remebr the 70mm film of Lawerence of Arabia!!!!yes oh yes!!!)

BTW--many folks will say, look negatives do not produce any better prints than pure digital and have the prints to prove it. I do not disagree, because what they forget, is when you have NOW have negatives reproduced, the machines scan the negative into digital form, and then it is printed from the digital scan image, using the same dot process as a pure digital image, rather than the pure process of light and chemicals. The latter still produces the best images on paper, by far. There is a smoothness, yet even more detailed image that is produced, than produced by any of the curent printers using digital in some form or fashion..

But it also rapidly becoming a lost art, heckfire, trying finding the highest quality paper to produce black and white images, it just is not being made....like it was a mere five years ago

this is mostly just all wrong

First you are attempting to define noise as only being those specks that sometimes appear randonly, when in fact, the only way to look at it, is that noise is a range of issues, including moire, et al, that are all interelated. Without question, the noise (even as you attempt to define it) can be generated by the details of an image, the dark areas, the redish areas, the very bright areas and the borders in those areas.

Indeed, one problem is dealing with color blurring of a sharp edge between two areas having distinctively different color and light--hence the source of the "jagginess" problem in Braveheart, where the result of the computer interpolation to deal with color blurring at the edge, at some point in this process (what point-I do not know) produced an unpleasant image of "jages" instead of correcting the blurring of color (and because this problem originates in the distance--it probably should be blurred, because blurred background and foreground is what produces a depth of field in a picture and the program is identifying the problem as one of a "color blur" that needs correction, then improperly processes the image)

You also refer to the sensor, when in fact, the sensor is connected to a processer and it is the processor that permits the sensor to work. The sensor is just a surface of pixels that collect the light. When the pixels emitted the electical charge as a result of contact with a photon, the individual electrical signal goes to the processor.

The processor perfoms a function called demosaicking which is used to reconstruct a full color image from the incomplete color samples outputed from the sensor which uses a color filter array to reduce artifacts. Notwithstanding all the false advertizing to the contrary that the RAW image information is "unprocessed", even a "raw" image from a camera has undergone massive processing (including DNR) before being placed on a memory card. The difference is that the raw image undergoes further processing in the computer using software where the image can be further manipulated as to certain variables (and this manipulation is extremely necessary to be able to view the image).

For example, CCC pixels have only changed to become smaller and more numerous, but it is the changes in its accompanying processor, that makes the newer "sensors" to be "better".



Because noise and other artifacts are reduced as a result of its use-making the job of the processor easier




This is so funny---what DNR you are turning off, is merely "extra DNR". If all DNR (including EE or sharpness) inside the processor was turned off, you would not recognize the image, except with extreme difficulty

Again: Noise and other artifacts are reduced by the processor, not the sensor pixels








there is so much stuff here that is so ridiculously wrong I barely even know where to begin!

How can you lump photon shot noise with sensor read noise with late chain electronics noise with moire??

It is ridiculous to say you wouldn't even recognize the image straight off the sensor, have you ever used IRIS or RAWAnalyze or some PRE-demosaic, AW software??

What the heck does DNR have to do with moire or edge enhancement?

You do realize that many digital cameras output files that have not even been de-bayered right?

RAW files have mostly NOT undergone massive processing, you can even see highly gaussian plots of sensor blackframe readout noise across all frequency ranges, etc. Granted some cameras like the Sony A900 do force a fair amount of processing even on RAW output, but that is pretty rare from DSLR.

The sensels do not directly feed the image processing DSPs you seem to have forgotten the entire ADC/amp stage....

The Braveheart stuff doesn't even make any sense....

The processor perfoms a function called demosaicking which is used to reconstruct a full color image from the incomplete color samples outputed from the sensor which uses a color filter array to reduce artifacts. Notwithstanding all the false advertizing to the contrary that the RAW image information is "unprocessed", even a "raw" image from a camera has undergone massive processing (including DNR) before being placed on a memory card. The difference is that the raw image undergoes further processing in the computer using software where the image can be further manipulated as to certain variables (and this manipulation is extremely necessary to be able to view the image).
Those DNR processors must be doing a pretty bang-up job of hiding the processing, since if I run my RAW files through a raw converter that doesn't demosaic the bayer data, my Canon EOS-30D puts out noise that looks decidedly random. I've definitely heard of "cooked" raws with some cameras but it's not universal.

Noise and all the rest of the artifacts of digital reporduction are simply different sypmtoms arising from the same problem of pixels not reacting as intended to light photons. To deal with these problems, the signal is being sampled (or processed). The result of faults with the processing produces a range of problems that consititute moire, alaising, and those little specks (that you think is all that noise constitutes)--Perhaps it is best to use the term distortion, so you will not get so confused, but they are all so interrelated as to causation as to be indistinguishable.

The fact that you and the other poster (bronx) are clueless and simply mouthing stuff you do not understand is easily demonstrated by the following:


[QUOTE=mhafner;17136279]Noise in a sensor has several sources but noise is not just any image anomaly. That is not a valid definition of noise. Aliasing is not noise. Show me the definition of aliasing as noise in some standard reference source, please. I would like to see it.

(With DeNiro voice): Are you talking to me? If you are, please don't make a hopeless mess out of different image processing concepts and mix it all together into one big noise pot. That is not helpful. Jaggies are not noise and noise are not jaggies. The connection is lose at best.



QUOTE]
Now I cite the simple and easy to understand references available to anyone who knows how to read (in addition to those cited earlier):

Page 10, the references to periodic noise, as demonstrated by moire.....i cited this earlier, if you had bothered to read, but here it is again:

http://campratty.com/4tooltours/kz/pdf/DigitalCameraNoiseRemovalFilter2006.pdf


Then there are simpler articles (if you want the academic, those are not easily found on the web, and you can buy your own copies and try to read them, if you like) --and please try to understand that "sampling" in one form or another is the fundamental method for dealing with noise and all the other problems of digital processing:

http://wiki.panotools.org/Aliasing

Aliasing is a process that results in "jaggies" or Moire effects when an image is sampled or resampled.

The word "aliasing" comes from frequency-domain analysis. Slow periodic sampling of a signal causes high frequency components to be "aliased" into low frequency ones that corrupt the sampled signal.

To avoid aliasing, a signal must not contain frequencies that are higher than one half the sample rate. Likewise, an input image must not contain details that are smaller than one pixel in the output image.

This can be accomplished by appropriately filtering the signal or blurring the image before resampling it.

http://en.wikipedia.org/wiki/Aliasing

In signal processing and related disciplines, aliasing refers to an effect that causes different signals to become indistinguishable (or aliases of one another) when sampled. It also refers to the distortion or artifact that results when the signal reconstructed from samples is different than the original continuous signal

For example the blurring reduces the noise generated by the sharp edges of the images---a well known source of "noise" --at least known to those who have actually dealt with same.
Now for those more interested in advanced discussions: http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p4.html


which discusses more advanced methods of sampling and averaging. And as pointed out above, it is the sampling errors that lead to the presence of all thse sypmtoms that the low pass filter is designed to eliminate.

Of course, I could perform the calculations for various situations, but that would be too tedious, you two guys (that includes Bronx) would quckly get lost, and something I do not do for free.

And the fundamental principles are the same, regardless of whether it is a CCD, CMOS or Foveon sensor or a 3CCD sensor (although Foveon sensors have their own set of unique problems)

Those DNR processors must be doing a pretty bang-up job of hiding the processing, since if I run my RAW files through a raw converter that doesn't demosaic the bayer data, my Canon EOS-30D puts out noise that looks decidedly random. I've definitely heard of "cooked" raws with some cameras but it's not universal.

The raw files are being partially processed and de-noised before they hit your memory card--and further processed by the program that reads the RAW data, but there really is not such completely random pattern to noise, although it appears random, because if it were totally and truly random, DNR would not work at all. DNR seeks a certain pattern and bases it sampling on the typical pattern. There is a difference between so-called random and periodic patterns--see page 10 of 34 of : http://campratty.com/4tooltours/kz/pdf/DigitalCameraNoiseRemovalFilter2006.pdf

(referenced earlier in another post)

As pointed out in the Chicago article, there are theoritical methods or models to noise elimination, but they have yet to implemented in the real world....

HVisone,

I just wanted to thank you for taking the time to keep up with this thread. Reading your posts has been informative and refreshing.

A really, really dumb question for you ... The color depth of BD is 8 bit I think, what is it on a top of the line DSLR? If this is a technically meaningless question, please say why.

HVisone,

I just wanted to thank you for taking the time to keep up with this thread. Reading your posts has been informative and refreshing.

A really, really dumb question for you ... The color depth of BD is 8 bit I think, what is it on a top of the line DSLR? If this is a technically meaningless question, please say why.
Two response:\

1. you are just playing a game and asking questions which you already know the answer--as it is written like my old grad profs would write

or

2. It is more of a question as to reality or practicality of human vision. Because there are those who would mantain that the human eye can not distinguish between the higher bit colors, why bother?

Slight color differences are somewhat easy to notice where compared side by side but when a subject is presented the same variation by a card, and then a second card, and then asked is there a difference, the responses become random. They will say there is a difference when there is none, and they will say there is not difference when there is a difference.

For this reason, the increase from 12 bits to 14 bits of the Nikon 300D, (and 14 bits present in the Canon 1 ds mark III) despite adding a considerable higher number of possible color variations, has been said to be meaningless.

Even Canon, in one of its technical articles admits the same:

http://www.usa.canon.com/dlc/controller?act=GetArticleAct&articleID=2748


Photoshop in its default mode, automatically converts images to 16 bits in rgb, where there are some monitors and computers that will not display 16 bit color.

It is possible to go 48bit and above, resulting in trillions of color variations.


Blu ray is not so limited as you imply in the question, except by industry protocal.

The so called Deep Color which some blu ray players falsely claim to use, increases the number of bits available for transmission, using 12-bit or higher compatibility, to provide at least 4,096 shades of each of the three primary colors for a total of 68.7 billion possible color variations. Some will even go to 48 bit.

Speaking of more false advertizing----Unfortunately 1080p discs and all other std DVDs are limited by protocal to 8 bits--but technically, a blu ray (or 1080p resolution) could be made to play with 16 bit color, which would come close to the gamut of colors that an LCD or plasma screen will currently support, esp with a 1.3 HDMI.

Movie makers hate the idea as this would make the deep color blu ray dvd into being an identical or virtual clone of a movie, So the idea of a "DEEP COLOR BLU RAY PLAYER" is currrently FALSE as no blu ray will do the "deep color" and with movie owner Sony holding the rights, it ain't likely in the near future

So all you folks who spent $$$$$$ to upgrade to deep color and bought those 1.3 cables, weeelll, err, err, err.....(and same is true for those who bought LCD or a plasma because of those incredible numbers......well......someday maybe, you will get a source that will really let the 'true colors come shining through"......

There is one advantage to higher bit colors in that when properly used, amnay of the artifacts between areas are not such a problem as the colors blend together better without all the tendency to have jaggies or other such noise or distortion (read the canon article before some of you others start sprouting off...


And this 8 bit conversion down to the Blu ray dvd to have the info imbedded on the disc, is another source of inaccuracy and distortion....

Actually it was - 3. An attempt to get a discussion of color depth from you. I succeeded :D

Is it fair to summarize you by saying that 12 bit would be a legitimate improvement (visible and meaningful to human vision) as well as a technical improvement (fewer artifacts as per your last two paragraphs) but that it won't happen because it would look too good?

If it is fair, do you think it's fair of me to say that current consumer HD media (i.e. BD but could just as easily be HD-DVD, I don't care about the format) is less than it could be using current technology?

That's been my view, I just don't know if it's technically valid. It would mean that a lot of the arguing that goes on here at AVS, which is premised often on the assumption that BD is some kind of ultimate, purist format to be treated with loving hands, is pointless.

Which isn't to say that those motivated shouldn't hold the studio's feet to the fire to get a better product, but that product is a consumer product that costs 20 bucks and is being hawked by the biggest greed mongers on the planet. We're just lucky there's at least some techies down in the cave that care enough to do their job and give us those nice digital images when it all works.

But I guess that's where the arguing comes in - when does it work? I just viewed in the last few days a movie on BD that was universally praised in the two main "rate it" threads here and I thought the PQ was ... annoying I guess is a good word. Razor sharp everywhere, made me squint (I view on a calibrated 50" plasma from 6" or so.)

I guess that's why I don't get involved, I never seem to agree with anyone lol.

Noise and all the rest of the artifacts of digital reporduction are simply different sypmtoms arising from the same problem of pixels not reacting as intended to light photons. To deal with these problems, the signal is being sampled (or processed).
The result of faults with the processing produces a range of problems that consititute moire, alaising, and those little specks (that you think is all that noise constitutes)--Perhaps it is best to use the term distortion, so you will not get so confused, but they are all so interrelated as to causation as to be indistinguishable.
The fact that you and the other poster (bronx) are clueless and simply mouthing stuff you do not understand is easily demonstrated by the following:

I understand this stuff pretty well. I had to because I was working in the field. ;)

Page 10, the references to periodic noise, as demonstrated by moire.....i cited this earlier, if you had bothered to read, but here it is again:
http://campratty.com/4tooltours/kz/pdf/DigitalCameraNoiseRemovalFilter2006.pdf

I disagree with calling aliasing noise. Periodic or not. I find it not helpful to describe any anomalies and artifacts in the sampled version of an analogue source noise. The most important attribute of noise is that it's random, and exactly not a deterministic artifact due to sampling at too low a rate. In addition the party analogy is misleading. The sounds of the world are what they are, not full of artifacts and noise. What is noise and what not at the party is a completely subjective decision. Noise at the party is what you don't want to hear and the rest is useful signal. The technical noise problem does not arise at all unless you start recording the sounds and there are differences between what you recorded and what the recording actually captures (and sounds like when played back). A high fidelity recording of the party actually has to record all that 'noise' in detail or the recording system would not be high fidelity.
The rest of the article is about filtering details of this specific filter for random noise. The filter is not meant to deal with aliasing of all kinds. No wonder, because it's not (random) noise. And the proper solution is to avoid violating the sampling theorem, not running a 'noise' filter over the picture after the fact.

and please try to understand that "sampling" in one form or another is the fundamental method for dealing with noise and all the other problems of digital processing:

Sampling is the fundamental operation to convert from analogue to digital and resampling is from digital to digital when the number of samples has to change. Old news. Aliasing is a sampling artifact, not a noise artifact. Nothing to be gained by calling it noise.

Now for those more interested in advanced discussions: http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p4.html

Describes ways to reduce noise in the sensor that are not caused by aliasing but physical implementation details of the sensor. And confirms what I said. Methods local to single pixels that are not the kind we see in video processing. The kind my still camera has too and which does not add visible artifacts to my stills. The non local methods are clearly separated in the document and it's pointed out that they (can) add artifacts.

which discusses more advanced methods of sampling and averaging. And as pointed out above, it is the sampling errors that lead to the presence of all thse sypmtoms that the low pass filter is designed to eliminate.

Sampling errors as in different pixels on the sensor not all reacting the same to photons and noise generation during read out etc. Not sampling errors as in aliasing by violating the sampling theorem. Big difference!

Of course, I could perform the calculations for various situations, but that would be too tedious, you two guys (that includes Bronx) would quckly get lost, and something I do not do for free.

Don't be so condescending. I worked on digital image processing research at university level and wrote my own filters. :rolleyes:

I understand this stuff pretty well. I had to because I was working in the field. ;)

I disagree with calling aliasing noise. Periodic or not. I find it not helpful to describe any anomalies and artifacts in the sampled version of an analogue source noise.

So now you disagree with those others who would call it noise, having been nailed on your BS "find me a reference that refers to it as noise" ....TOO funny:rolleyes:


Sampling is the fundamental operation to convert from analogue to digital and resampling is from digital to digital when the number of samples has to change. Old news. Aliasing is a sampling artifact, not a noise artifact. Nothing to be gained by calling it noise.

Only for those that fail to understand the sampling process, and one of the primary purposes of sampling is to reduce or eliminate noise, noise of all sorts, and where the process fails, the result is noise in various forms such as those specks you see as being random (these specks are not random, but are generated in direct response to certain conditions as well as sampling failures just the same as moire, jaggies and so forth....


Don't be so condescending. I worked on digital image processing research at university level and wrote my own filters. :rolleyes:


Just trying to make it simple for everyone, esp those who think they know, but do not.......all i can say, if I wanted to be polite, those who do, do, the others write at a university, or I could say that well, your challenge has been met and destroyed, knocking out the foundations for your claims:eek:

Actually it was - 3. An attempt to get a discussion of color depth from you. I succeeded :D

Is it fair to summarize you by saying that 12 bit would be a legitimate improvement (visible and meaningful to human vision) as well as a technical improvement (fewer artifacts as per your last two paragraphs) but that it won't happen because it would look too good?

If it is fair, do you think it's fair of me to say that current consumer HD media (i.e. BD but could just as easily be HD-DVD, I don't care about the format) is less than it could be using current technology?

That's been my view, I just don't know if it's technically valid. It would mean that a lot of the arguing that goes on here at AVS, which is premised often on the assumption that BD is some kind of ultimate, purist format to be treated with loving hands, is pointless.

Which isn't to say that those motivated shouldn't hold the studio's feet to the fire to get a better product, but that product is a consumer product that costs 20 bucks and is being hawked by the biggest greed mongers on the planet. We're just lucky there's at least some techies down in the cave that care enough to do their job and give us those nice digital images when it all works.

But I guess that's where the arguing comes in - when does it work? I just viewed in the last few days a movie on BD that was universally praised in the two main "rate it" threads here and I thought the PQ was ... annoying I guess is a good word. Razor sharp everywhere, made me squint (I view on a calibrated 50" plasma from 6" or so.)

I guess that's why I don't get involved, I never seem to agree with anyone lol.

48 bit "Deep Color" would be a clear improvement over 8 bit, as would 12 bit over 8 bit....Indeed, there is a dramactic pereceivable improvement in 12 over 8 bit......the break even point though seems to be no higher than somewhere in the 12 v 14 bit area as to actual ability of humans to perceive differences or so current data seems to indicate in terms of percievable differences

(I should say the average person, there are those exceptions of exceptional people, and then there are those who are medically speaking partially or completely color blind)

So you do have a point that the blu ray std as currently set is NOT the highest perceivable standard available (and blu ray probably should be at least be at 16 bits)---but that might well look too good in the eyes of the movie makers--and given industry standards as well as greed, is not likely to be happening any time soon--sort of like the long and difficult battle for HD tv, even though the japanese had it for years and years before it became available in the USA

Funny as another example of false advertizing (and why BD SHOULD BE AT 16 BITS) is how Photoshop will claim that they upscale to 16 bits as some sort of higher quality over 12 or 14bits, but that is not the real reason. It has to do with binary operations of a computer and 16 bit is far easier to computate then is 14bit or even 12 bit. It has to do with binary operations---the optimum levels are 8, 16, 32, and 64 for bits close to those sizes (ie 30 bit would be tougher than 32 bit and so on)

as to color, everyone is talking Braveheart as to its fidelity to color. There is no way any reveiwer can know the fidelty, unless they have the exact positive negative to compare it to, in side by side comparisions for reasons stated above.....although my guess is that it is very close

(would add that 16 bit does take up more memory space, and one reason for its absence in DS cameras and perhaps even in the current disc size limitations of blu ray)

HVisone:
>Noise and all the rest of the artifacts of digital reporduction are simply >different sypmtoms arising from the same problem of pixels not reacting as >intended to light photons.

Interesting that you don't even realize that the photons themselves are the result of much of the noise. Shot noise is a pretty important aspect to have overlooked.



>The result of faults with the processing produces a range of problems that >consititute moire, alaising, and those little specks (that you think is all that >noise constitutes)--Perhaps it is best to use the term distortion, so you will >not get so confused, but they are all so interrelated as to causation as to >be indistinguishable.

haha, really? They are all so interrelated that the causation is indistinguishable between ADC noise, amp noise, shot noise, moire/aliasing, etc.?


>For example the blurring reduces the noise generated by the sharp edges of >the images---a well known source of "noise" --at least known to those who >have actually dealt with same.
>Now for those more interested in advanced discussions: >http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p4.html
>which discusses more advanced methods of sampling and averaging. And >as pointed out above, it is the sampling errors that lead to the presence of >all these sypmtoms that the low pass filter is designed to eliminate.

yeah.... sure.... that is EXACTLY what that site is saying.

> Of course, I could perform the calculations for various situations, but that >would be too tedious, you two guys (that includes Bronx) would quckly get >lost, and something I do not do for free.

Too bad. I am sure I will be hopelessly lost now without your expert guidance. I guess I will just have to go back to my quantum field theory research, more my speed.

Two response:\

1. you are just playing a game and asking questions which you already know the answer--as it is written like my old grad profs would write

or

2. It is more of a question as to reality or practicality of human vision. Because there are those who would mantain that the human eye can not distinguish between the higher bit colors, why bother?

Slight color differences are somewhat easy to notice where compared side by side but when a subject is presented the same variation by a card, and then a second card, and then asked is there a difference, the responses become random. They will say there is a difference when there is none, and they will say there is not difference when there is a difference.

For this reason, the increase from 12 bits to 14 bits of the Nikon 300D, (and 14 bits present in the Canon 1 ds mark III) despite adding a considerable higher number of possible color variations, has been said to be meaningless.




It is actually more about maximizing DR and giving room to apply various level and tone curves and nothing much to do with trying to better cover the gamut of human vision or anything. It has been said to have been meaningless for most cameras to have gone from 12bits to 14bits simply because virtually none have good enough sensor/electronics to take advantage of that many bits. They virtually all have under 12bits of DR. The Nikon D3X, perhaps, is able to make use of 14bits.

Just trying to make it simple for everyone, esp those who think they know, but do not.......all i can say, if I wanted to be polite, those who do, do, the others write at a university, or I could say that well, your challenge has been met and destroyed, knocking out the foundations for your claims:eek:
I'm really convinced now. You win. ;)

The whole discussion is interesting nonetheless as it shows the (big) difference between the noise reduction or elimination going on in sensors which is local to individual pixels and the noise reduction we usually talk about here on AVS which is not local to single pixels and requires neighbouring pixels (spatial and/or temporal). The noise in the sensor is not noise of the source itself but noise which is an artifact generated by the sensor sampling the source. The noise on our sampled video is both, source noise and sampling noise, with the former (far) dominating the latter on state of the art transfers. The noise reduction methods local to single pixels aimed at removing sampling/sensor noise don't leave the artifacts behind we are used to seeing from filters not local to single pixels.
If we look at sampled 16/35mm film frames from state of the art sensors the 'noise' we see is to a very large degree source noise (grain), not sensor noise.

Got my disc, looks pretty spiffing to me

Got my disc, looks pretty spiffing to me

How's the grain?:D

Grainy

Grainy

You sure that's not just sensor noise now?

It might be jitter that is changed by moon rays

Ok just watched some of it, and it looks like light DNR with a layer of fake grain added on top.
The EE is still there

Ok just watched some of it, and it looks like light DNR with a layer of fake grain added on top.
The EE is still there

Yuck the EE is the killer for me. I thought it looked like added fake grain ( not saying it is ) in the screenshots provided in this thread and mentioned it in an earlier post and i know they have the tools to do that sort of thing but who knows for sure.

I'm pretty sure the grain is real but the EE and whatever equipment they used to scan the film has exaggerated the hell out of it

Maybe it does not look natural in motion tho :(

Why would the director want it to be that way?I cannot see the point and of course the studios don't even allow to degraded the PQ cause the disc can't be saled.