I was wondering what the difference is between the best Blu-Ray (or HD-DVD) discs and the prints they use in actual theaters? What I wanna know is the difference in tech specs. Does anyone know the resolution and video/audio bitrates for original theatrical prints that are completely uncompressed is? I have always wanted to know this. Thanks. :)

I was wondering what the difference is between the best Blu-Ray (or HD-DVD) discs and the prints they use in actual theaters? What I wanna know is the difference in tech specs. Does anyone know the resolution and video/audio bitrates for original theatrical prints that are completely uncompressed is? I have always wanted to know this. Thanks. :)

If you are asking about the difference between a release print and a Digital Cinema presentation:

http://www.etconsult.com/papers/Technical%20Issues%20in%20Cinema%20Resolution.pdf

If you are asking for the raw uncompressed data that is retrieved during a 4K transfer from film:

http://www.soundandvisionmag.com/features/2624/blade-runner-how-great-hd-is-made.html

If you are asking about the difference between a release print and a Digital Cinema presentation:

http://www.etconsult.com/papers/Technical%20Issues%20in%20Cinema%20Resolution.pdf

If you are asking for the raw uncompressed data that is retrieved during a 4K transfer from film:

http://www.soundandvisionmag.com/features/2624/blade-runner-how-great-hd-is-made.html

Thank you! That is very very helpful. But I was wondering if you knew how to get information on the differences of bitrates in Blu-Ray and the theatrical versions of movies. I didn't see anything in the links about that.

Thank you! That is very very helpful. But I was wondering if you knew how to get information on the differences of bitrates in Blu-Ray and the theatrical versions of movies. I didn't see anything in the links about that.
Well film is film, if it hasn't been scanned and digitally altered in any way and is just projected on a film projector, it won't have any bitrate.

These days they often scan it digitally to colour correct it/colour grade it/composite etc.

If they project it digitally, and someone will hopefully correct me if I'm wrong, but I think they mostly use 2K projectors (I'm not sure) - some might be 4K. And I think they output it using something like Jpeg2000 (more like mjpeg than mpeg) and with a lot higher bitrate and I think more colour depth. Blu-ray is up to around 40 megabits per second for the video and digital cinema will be lots higher (around 250 Mbit/s according to this link http://en.wikipedia.org/wiki/Digital_cinema - I'm not sure if that's for 2K or 4K or both)

So I think:
Blu-ray = up to 40 Mbit/s for video (VC1 or AVC/Mpeg4 or Mpeg2)
Digital Cinema = up to 250 Mbit/s for video? (Jpeg2000)

Well film is film, if it hasn't been scanned and digitally altered in any way and is just projected on a film projector, it won't have any bitrate.

These days they often scan it digitally to colour correct it/colour grade it/composite etc.

If they project it digitally, and someone will hopefully correct me if I'm wrong, but I think they mostly use 2K projectors (I'm not sure) - some might be 4K. And I think they output it using something like Jpeg2000 (more like mjpeg than mpeg) and with a lot higher bitrate and I think more colour depth. Blu-ray is up to around 40 megabits per second for the video and digital cinema will be lots higher (around 250 Mbit/s according to this link http://en.wikipedia.org/wiki/Digital_cinema - I'm not sure if that's for 2K or 4K or both)

So I think:
Blu-ray = up to 40 Mbit/s for video (VC1 or AVC/Mpeg4 or Mpeg2)
Digital Cinema = up to 250 Mbit/s for video? (Jpeg2000)

WOW, huge difference, damn! I wonder when we will get those kind of bitrates on home formats...? Also, any idea in regards to the sound bitrate? I really wanna know that. :D (and yes, I meant the versions that are played in the nicest theaters)

WOW, huge difference, damn! I wonder when we will get those kind of bitrates on home formats...? Also, any idea in regards to the sound bitrate? I really wanna know that. :D (and yes, I meant the versions that are played in the nicest theaters)
Actually I think, reading that page, that I was wrong and that the Digital Cinema bitrate of up to 250 Mbit/s includes the soundtrack too (in uncompressed WAV format I think - it says it might be 1 of 2 different sampling rates and it doesn't mention number of tracks etc.) - I think other people on this forum (like FilmMixer) will know a lot more about this.

So I think - for Video+Audio (and subtitles) its:
Blu-ray: up to 48.0 megabits/second (VC1 or AVC/Mpeg4 or Mpeg2)
Digital Cinema: up to 250 megabits/second (Jpeg2000 & Wav audio)

WOW, huge difference, damn! I wonder when we will get those kind of bitrates on home formats...? Also, any idea in regards to the sound bitrate? I really wanna know that. :D (and yes, I meant the versions that are played in the nicest theaters)

It's not such a huge difference, really, and if there is any, I'd bet BD's bitrates are actually more favorable. Remember that VC1 and AVC (and MPEG2 for that matter) are far more efficient than JPEG2000 (all I frames, if I'm not mistaken).

It's not such a huge difference, really, and if there is any, I'd bet BD's bitrates are actually more favorable. Remember that VC1 and AVC (and MPEG2 for that matter) are far more efficient than JPEG2000 (all I frames, if I'm not mistaken).
No ways, at least I don't think so :eek:
But I do think VC1/AVC etc. are more efficient codecs, though with Jpeg2000 it wouldn't matter how much motion there was between the frames, and you'd get less banding as it's higher bitdepths for the colour etc.

Keep in mind that the point of the digital cinema standard specifying JPEG2000 for image compression is that the output must be an exact match for what "entered" the mastering system: "The DCP [Digital Cinema Package] arrives at the theater, it is unpackaged, decrypted and decompressed to create the DCDM* [Digital Cinema Digital Master], where DCDM* is visually indistinguishable from the original DCDM (where the original DCDM was the input to the Digital Cinema Mastering Process)". [From http://www.dcimovies.com/DCIDigitalCinemaSystemSpecv1_2.pdf] The digital cinema folks are more concerned with quality and consistency than with storage space (they're currently using hard drives instead of optical disks for storage).

Any idea on audio bit rates? I believe Dolby TrueHD and DTS MA run between 3-7Mbits. I'm wondering how that compares to the audio used in theaters. I know I was given the 250 figure, but I don't know how much of that is for audio...

I think if you were to take a Digital Cinema projector and project it in a home environment on a 120" screen (which is generally bigger than most of us projector lovers use), then you'll find little difference between, digital vs Blu-ray print. And that really should be the bottom line.

I was watching Ratatouille again on my 720p Epson, and I was floored by the sharpness and detail at 10 feet back on my 92". At 1080p, it will look even more phenominal, but then the diminishing returns chapter opens up.

It's great to talk about numbers etc... But in a home environment I don't think Blu-ray can be topped in terms of a price to performance ratio, at least not for many many years.

No ways, at least I don't think so :eek:
But I do think VC1/AVC etc. are more efficient codecs, though with Jpeg2000 it wouldn't matter how much motion there was between the frames, and you'd get less banding as it's higher bitdepths for the colour etc.
JPEG2000 is much less efficient because it can't use inter frame redundancy. So the quality of the compression is not 6 times better because the bit rate is 6 times higher (at the maximum). It's better but not 6 times. More relevant are the different color gamut and the >= 10 bit samples.

Any idea on audio bit rates? I believe Dolby TrueHD and DTS MA run between 3-7Mbits. I'm wondering how that compares to the audio used in theaters. I know I was given the 250 figure, but I don't know how much of that is for audio...

Not sure if the OP understands the difference between analog film and a digital theater so keep in mind that at this point digital theaters make up a very small percentage of overall theaters. Not sure of the ratio in the US but it is probably something like 95% film to 5% digital. Film uses DD encoded on the film strip of 360Kbps or DTS on CD-ROM at 1.5Mbps (SDDS somewhere in between on the film strip?). In a digital theater the audio is likely 24/48 LPCM which would be the same data rate (6.9Mbps for 5.1 24/48) as HDM.

Not sure if the OP understands the difference between analog film and a digital theater so keep in mind that at this point digital theaters make up a very small percentage of overall theaters. Not sure of the ratio in the US but it is probably something like 95% film to 5% digital. Film uses DD encoded on the film strip of 360Kbps or DTS on CD-ROM at 1.5Mbps (SDDS somewhere in between on the film strip?). In a digital theater the audio is likely 24/48 LPCM which would be the same data rate (6.9Mbps for 5.1 24/48) as HDM.


There are approx 25,000 screens in the USA of which 5000 are equipped for digital. 20% - not so small a number.

SDDS somewhere in between on the film strip?.

Yeah, the SDDS track is on either side of the frame on a 35mm print.

DD is inbetween the sprocket holes, and the DTS timecode is next to the analog track...in case anybody was wondering.

Not sure if the OP understands the difference between analog film and a digital theater so keep in mind that at this point digital theaters make up a very small percentage of overall theaters. Not sure of the ratio in the US but it is probably something like 95% film to 5% digital. Film uses DD encoded on the film strip of 360Kbps or DTS on CD-ROM at 1.5Mbps (SDDS somewhere in between on the film strip?). In a digital theater the audio is likely 24/48 LPCM which would be the same data rate (6.9Mbps for 5.1 24/48) as HDM.

Actually DD is 320Kbps on film strip and DTS in a theater is 1.2Mbps.

There are approx 25,000 screens in the USA of which 5000 are equipped for digital. 20% - not so small a number.

WOW that is a much higher number that I would have thought. Do you have any idea what the 2K to 4K breakdown is? Also what about DLP to LCoS ratio? Maybe we are finally near the magical "10 years to go" mark for digital to replace film.

WOW that is a much higher number that I would have thought. Do you have any idea what the 2K to 4K breakdown is? Also what about DLP to LCoS ratio? Maybe we are finally near the magical "10 years to go" mark for digital to replace film.
I would like to know this too. And is the 250Mbit/s the maximum data rate for Digital Cinema the same maximum for both 2k and 4k?

So I think:
Blu-ray = up to 40 Mbit/s for video (VC1 or AVC/Mpeg4 or Mpeg2)
Digital Cinema = up to 250 Mbit/s for video? (Jpeg2000)

You are correct actually

But they also can do FULLY uncompressed TIFF for video. That goes up to a whopping 1 Gb/s.

and specifically


The ingest interface shall comply with either Clause 34 or Clause 44 of IEEE 802.3-2005 for either 1000 Mb/s or 10 Gb/s operation, respectively.


I would like to know this too. And is the 250Mbit/s the maximum data rate for Digital Cinema the same maximum for both 2k and 4k?

Oh heavens no. The maximum bandwidth that D-Cinema is capable of is 10 Gb/s. Quad link - Dual SDI. Why so much bandwidth you might ask?

AES Encryption 128 bit in cipher blocking chaining mode.

Uncompressed TIFF @ 12 bit.

WOW that is a much higher number that I would have thought. Do you have any idea what the 2K to 4K breakdown is? Also what about DLP to LCoS ratio? Maybe we are finally near the magical "10 years to go" mark for digital to replace film.

The number of 4K PJ's is a fraction of the 5000 because it seems that only Sony is selling a 4K PJ at this moment. But JVC has invented a 4K chip to be installed in a DC PJ.

Based on what I have read - maybe 100 or so 4K PJ's.

Almost all of the DC PJ's are based on the DLP chip (2K PJ's)

You are correct actually

But they also can do FULLY uncompressed TIFF for video. That goes up to a whopping 1 Gb/s.



Oh heavens no. The maximum bandwidth that D-Cinema is capable of is 10 Gb/s. Quad link - Dual SDI. Why so much bandwidth you might ask?

AES Encryption 128 bit in cipher blocking chaining mode.
Thanks - that's tons more. I wonder why they don't use that instead of Jpeg2000 more. Maybe it's because it would cost too much for the hard discs, or other stuff might be more expensive too maybe.

Maybe they need to update the http://en.wikipedia.org/wiki/Digital_cinema page because they only seem to mention the 250 Mbit/s rate and not the 1 to 10 Gb/s rates.

PS: Although digital cinema can use Uncompressed TIFF, is that what most films presented digitally now use - or do they use the Jpeg2000 compression?

Approx how much more efficient is VC1 or Mpeg4/AVC than Jpeg2000. Assuming VC1 or Mpeg4/AVC had the same colour bitdepth and colour scheme as digital cinema, what bitrate would be needed to match the average cinema using Jpeg2000 (obviously uncompressed Tiff at 10 Gb/s at 4K would be a lot better though :)).

So is it up to 1 Gb/s uncompressed Tiff for the actual video and the other 9 Gb/s just for the encryption? - or is that for the video/audio in some way too?

Thanks - that's tons more. I wonder why they don't use that instead of Jpeg2000 more. Maybe it's because it would cost too much for the hard discs, or other stuff might be more expensive too maybe.

Maybe they need to update the http://en.wikipedia.org/wiki/Digital_cinema page because they only seem to mention the 250 Mbit/s rate and not the 1 to 10 Gb/s rates.

Well remember, Wiki can be updated by anyone. Usually the company or in this case, the DCI, probably does not update that. :)

They are using Jpeg2000 more but the bandwidth requirements are still the same. That is as of March 7, 2008, Digital Cinema Specification 1.2.

as for costs. They just need space for the movie/s.

SO say that the theater in question has 8 room. Each with 4k worth of display (EXAMPLE).

Their NAS/SAN would need to be roughly about 40 TB (377.374 GB total for a three hour movie with the following , One three-hour feature,
20 minutes of pre-show material at the same resolution, 16 channels of uncompressed audio at 48 kHz at 24 (AES3, this means encrypted) bits, 3,000 sub pictures in PNG file format, 3,000 Timed Text lines).

EMC, NetApp, Hitachi make NAS/SAN that can go to about 140 TB capacity :)

btw an update. Seems like the DCI has actually heard thier IT folk.

the 10 Gb/s connection are now specified as fiber and not the quad-link-dual SDI that I posted before.

A slight side note: this theater is just down the road from me.

Rave Motion Pictures Hickory Creek 16 - http://www.ravemotionpictures.com/theaters/tx_hickorycreek16.html

The Rave Motion Pictures chain is a 100% Digital theater chain (every screening room has both a digital cinema setup and a standard film projector). If I have the opportunity, I now go exclusively there.

Here's a link to their entire theater list: http://www.ravemotionpictures.com/theaters/default.html

There are approx 25,000 screens in the USA of which 5000 are equipped for digital. 20% - not so small a number.
This sounds awfully high. Do you have a link? I live in very affluent suburb of Northern VA and among 50 or so nearby screens, I think only about 10% are digital. If the number is truly 20%, then most must be also dual digital/analog because it's very difficult to find digital projection for anything other than blockbuster of the week.

Some current info on D cinema rollout at today's display daily:

http://displaydaily.com/

Some current info on D cinema rollout at today's display daily:

http://displaydaily.com/

So according to this article, it's 4,000 digital screens out of 37,000 in the US. This is little over 10% which is still a little more than expected.

I think if you were to take a Digital Cinema projector and project it in a home environment on a 120" screen (which is generally bigger than most of us projector lovers use), then you'll find little difference between, digital vs Blu-ray print. And that really should be the bottom line.



Not on your nelly!. Blu-ray is VIDEO . Relative to film it has a limited gamut and a severly compromised intensity range : it also exhibits dynamic range compression from shot to shot and scene to scene by virtue of the necessary telecine its gone through.

Peak whites never look all that acceptable to me on video (they always look harsh and clipped even if you calibrate to show everything up to 255). Blacks are generally quite flat although they don't really exhibit the same harsh clip that the whites have. Video is basically all mid tone.

As for the dynamic range compression : think about the night time modes on audio processors and apply that to imagery and thats basically what video means.

Not on your nelly!. Blu-ray is VIDEO . Relative to film it has a limited gamut and a severly compromised intensity range : it also exhibits dynamic range compression from shot to shot and scene to scene by virtue of the necessary telecine its gone through...

You should reread what you are quoting. It is about commercial digital projection and not film projection. So both are video that was shot on film. The BD is just more compressed and uses a different codec. Also the commercial could be 4K or 2K. 2K would at least be a little higher resolution that my 1080p projector.

Also I think most commercial projectors have a much lower contrast range and the lamps are far too bright to use on a small (10' wide) screen.

You should reread what you are quoting. It is about commercial digital projection and not film projection. So both are video that was shot on film. The BD is just more compressed and uses a different codec. Also the commercial could be 4K or 2K at least which is a little higher resolution that my 10870p projector.

Dcinema is not VIDEO . Do your own research.

As for the dynamic range compression : think about the night time modes on audio processors and apply that to imagery and thats basically what video means.
I'm not sure. The dynamic range as in smallest versus largest value can be as good at home from HD video as in the cinema. You can have even a larger range at home with lower blacks and the same 16 ftl peak white. The difference is what happens between this bottom black and peak white. There a lot less steps at home than in the cinema, so it's a resolution difference within that range.
In audio though range compression actually lowers peak values and rises the lowest values. The range is made smaller.

I'm not sure. The dynamic range as in smallest versus largest value can be as good at home from HD video as in the cinema. You can have even a larger range at home with lower blacks and the same 16 ftl peak white. The difference is what happens between this bottom black and peak white. There a lot less steps at home than in the cinema, so it's a resolution difference within that range.
.

I usually refer to it as intensity range (steps) but no-one knows what that means round here. You take my point though.

Here is a quote from Home Cinema Choice where a similar test was conducted back around late 1999.

http://i262.photobucket.com/albums/ii100/BourneUltimatum3417/DVDVS35mm.jpg

Here is a quote from Home Cinema Choice where a similar test was conducted back around late 1999.


not as sharp or with as much contrast

Little innocently worded phrase. "Looked like crap in comparison" would have been nearer the mark.

I remember this from a few years ago. It must have taken a lot of effort to get that many idiots in one place at the same time.

I take it you read Home Cinema Choice or have you stopped buying it because I haven’t brought any since 2001 I think. Many of those idiots are professionals LOL, so which ones are you referring too?:D

I take it you read Home Cinema Choice or have you stopped buying it because I haven’t brought any since 2001 I think. Many of those idiots are professionals LOL, so which ones are you referring too?:D


BKTS (usually referred to as the blue rinse brigade).

I moderated my post as I was venting my spleen about how moronic HCC is a little too readily.

I notice the only quoted comments are from Snell and Wilcox guys essentially demoing their interpolator.

So scaled projected SD dvd looked okay but not as sharp or with the contrast range of 35mm: hugely profound conclusion that . And shutting your eyes is darker compared with opening them....great. ( through a £100000 interpolator as well remember)

HCC strikes again. I bet they'd say exactly the same thing if they used 1080p versions of the films instead of dvd.

BKTS (usually referred to as the blue rinse brigade).

I moderated my post as I was venting my spleen about how moronic HCC is a little too readily.

I notice the only quoted comments are from Snell and Wilcox guys essentially demoing their interpolator.

So scaled projected SD dvd looked okay but not as sharp or with the contrast range of 35mm: hugely profound conclusion that . And shutting your eyes is darker compared with opening them....great. ( through a £100000 interpolator as well remember)

HCC strikes again. I bet they'd say exactly the same thing if they used 1080p versions of the films instead of dvd.


Hey if you like me to re-edit the post after you made the comment, but hey that’s 9 years ago now.:D


The Dolby SR-D issue about dirt between the modified video camera system, where the digital bits are translated and decoded into separate audio channels and watching small particles of dust fluttering around near the Dolby cat 700 when I was in the cinema back around late of 1998 for Warner Bros Village cinemas UK, didn’t worry me. I vaguely remember, I think it was the optical pick up head that had the red LED on the projector.

http://www.kinoteam.de/kt/image/big/reader-dd.jpg

http://www.dolby.com/images/products/catno702/702_angle.jpg

http://www.dolby.com/professional/motion_picture/catno702_01.html

The whole tone that the differences were subtle or or minor is what gets me .

The comments about the differences in sharpness and contrast is like a huge massive gulf. Contrast and sharpness differences between images are like the differences between a paddle steamer and the space shuttle.

As to the comments about the sound head : yes I think they just screwed up the install to be honest.

You should reread what you are quoting. It is about commercial digital projection and not film projection. So both are video that was shot on film. The BD is just more compressed and uses a different codec. Also the commercial could be 4K or 2K. 2K would at least be a little higher resolution that my 1080p projector.

Also I think most commercial projectors have a much lower contrast range and the lamps are far too bright to use on a small (10' wide) screen.

the problem is that film is still film even if it is "virtual" (ie the film is on digital hard drives with encryption etc). Media for home use is VIDEO hence having to do the 3:2 pull down.

the problem is that film is still film even if it is "virtual" (ie the film is on digital hard drives with encryption etc). Media for home use is VIDEO hence having to do the 3:2 pull down.

Well its more down to the colourspace and dynamic range of the imagery . 24p video is still video . The Dcinema release formats are closer to a DI type colourspace. although I don't think they have quite the same headroom.

Its not video though think of it as a digital film print.

Well its more down to the colourspace and dynamic range of the imagery . 24p video is still video . The Dcinema release formats are closer to a DI type colourspace. although I don't think they have quite the same headroom.

Its not video though think of it as a digital film print.

I understand

I have DCI spec 1.2 in my hands ;)

12 bit colour space and all the X Y plots too boot :) Hence why they still are using uncompressed Tiff but have the option of using Jpeg2000.

...Its not video though think of it as a digital film print.

...and I have a PDF of a $100 bill so can I think of that as money too:D (guess I can but can I spend it?)? The industry may want to use a fancy term like digital film but it ain't film once it is scanned and stored as 0's and 1's on a hard drive.

...and I have a PDF of a $100 bill so can I think of that as money too:D (guess I can but can I spend it?)? The industry may want to use a fancy term like digital film but it ain't film once it is scanned and stored as 0's and 1's on a hard drive.

Its 12bit. It has a different whitepoint , gamut and curve than video. Its designed to mimic and be reasonably transparent to a print film dynamic.

Its not video .Its not a simple question of nomenclature. Go call it video around people who actually use it and they will think you are a moron.

Its 12bit. It has a different whitepoint , gamut and curve than video. Its designed to mimic and be reasonably transparent to a print film dynamic.

Its not video .Its not a simple question of nomenclature. Go call it video around people who actually use it and they will think you are a moron.
Mr D do you think it would be good if there were Blu-ray players available to consumers that could play back content with the same 12 bit per colour channel bitdepth, gamut, white-point etc. and could play back content at 4K resolution for a reasonable price (and there were compatible consumer HDTVs too with such specifications, and assuming the players and TVs could handle any necessary increase in bitrate)?

Mr D do you think it would be good if there were Blu-ray players available to consumers that could play back content with the same 12 bit per colour channel bitdepth, gamut, white-point etc. and could play back content at 4K resolution for a reasonable price (and there were compatible consumer HDTVs too with such specifications, and assuming the players and TVs could handle any necessary increase in bitrate)?

I don't think 4k is important from a resolution perspective with regard to a domestic format.

I do think increasing the bitdepth AND increasing the dynamic range would be benificial but bear in mind most affordable displays won't even handle 8bit video transparently at the moment.

1k (1024x768) 8bit log encoded film material still looks better than 1080p 8 bit video.

I'd have been happy with 720p 8bit log encoded material over 1080p video I suspect .

4k is not viable for home use. The somewhat subtle improvement thats noticable between 720p and 1080p is way more noticable than the difference between 2k and 4k.

I can't reliably identify the difference between 2k and 4k resolution on most film imagery without the abilty to AB . I can spot 1080p video a mile away.

Not sure if the OP understands the difference between analog film and a digital theater so keep in mind that at this point digital theaters make up a very small percentage of overall theaters. Not sure of the ratio in the US but it is probably something like 95% film to 5% digital. Film uses DD encoded on the film strip of 360Kbps or DTS on CD-ROM at 1.5Mbps (SDDS somewhere in between on the film strip?). In a digital theater the audio is likely 24/48 LPCM which would be the same data rate (6.9Mbps for 5.1 24/48) as HDM.

So the bitrates and dynamics in the audio tracks on Blu-Ray would be similar to what's in the actual theaters? That's awesome! We finally match something in tech specs in regards to HT vs. Cinema. :D Well, we definitely win in room acoustics - and HT equipment can blow cinema equipment out of the water. All we need now is the optical formats to match what is shown in the actual theaters.

[...](AES3, this means encrypted) [...]

AES3 in the context of digital audio does not mean "encrypted", i.e. it is not some type of AES encryption.

"AES3", or commonly refered to as AES/EBU, is a digital audio transmission standard.

It is equivalent to IEC 60958 Part 4.

S/PDIF was derived from AES3 and is basically a consumer version of that standard.

Here's some info about AES3 in the context of MPEG2 Transport Streams:
"The SMPTE 302M standard specifies the transport of linear PCM digital audio in MPEG-2 transport streams. MP2TSME allows you to configure up to 16 SMPTE 302M elementary streams in each program. By definition, each SMPTE 302M elementary stream may convey 2, 4, 6, or 8 channels each pair of which are derived from a two-channel data source like an AES3 stream, PCM data, or other digital audio data like Dolby-E.

If more than 8 channels are required, a second SMPTE 302M elementary stream may be configured to carry the additional channels. A channel_identification field in the SMPTE 302M header is used to indicate the channel number of the first channel in the SMPTE 302M elementary stream. Note that each SMPTE 302M elementary stream must be assigned a unique PID in the transport stream."

1k (1024x768) 8bit log encoded film material still looks better than 1080p 8 bit video.

Can you elaborate on this? Because of the different color gamut? With 8 bits there are not many more steps than on video (assuming this uses <= 256 steps while video uses ~220).

Can you elaborate on this? Because of the different color gamut? With 8 bits there are not many more steps than on video (assuming this uses <= 256 steps while video uses ~220).

It has far more intensity variation towards black and a much softer roll off towards peak white. It can posterise with certain imagery ( the opening titles to Gladiator were done at 8bit log I was told and it has banding) but probably not any more than 8bit video . Probably the film grain helps disguise some banding but likley not to any greater extent than it does with 8bit video.

And there is no colour downsampling which helps. I don't really lock on to the improved color gamut ( certainly not in isolation) but certainly that softer roll off into black and white is one of the main things that I prefer.

Now bear in mind thats 8bit log thats encoding a full negative density 0-1023 in 10bit land) and its usually fine on most imagery when it comes to posterising. If you didn't encode all that headroom ( ie: you graded it and then lowered the white point : 685 is the nominal "white ref with 10bit log so you'd maybe want to encode a stops worth of headroom above this ) then you free up your 8bits somewhat and minimise posterising further.

Essentially what you end up with is a cineon or dpx type film scan image thats encoding a pseudo negative dynamic with a LUT applied to give you a viewable print dynamic.

Think of it like a domesticated DI enviroment. I think I posted the Kodak lad chart on here somewhere downsampled to 8bit but with no telecine type color correction ( essentially an 8bit log jpeg).

http://library.creativecow.net/articles/oconnell_pete/cineon.php

Here is a little article about cineon/log file formats ( cineon as a format is becoming superceded by DPX but the colourspace and bitdepth are to all intents and purposed identical).

Bear in mind that the "corrected" images in this article represent probably something like an sRGB type colorspace so its still not representative of what a cineon type image looks like when its lutted into a print film dynamic on a video monitor. (essentially you are still looking at a "telecined" video type image rather than something that represents an improvement over video).

And its based around after effects which has all sorts of stupid things that it thinks you want to do but the explanation of the colorspace is reasonably easy to digest.

Mr. D:

This paragraph:

So Cineons look washed out because 35mm film is very sensitive to darkish values and less and less sensitive to brighter values. In other words, 35mm film has an exponential (AKA logarithmic) sensitivity to light, which is why cineon files do.

Am I understanding this correctly? Film is more sensitive to low light detail as opposed to high light detail? We can see more detail in shadow areas then we can in bright sunny areas?

Mr. D:

This paragraph:



Am I understanding this correctly? Film is more sensitive to low light detail as opposed to high light detail? We can see more detail in shadow areas then we can in bright sunny areas?

The Human visual system is more sensitive to intensity changes towards black than towards white . Film and video to a lesser extent exploit this with their non-linear distribution curves. In bitdepth terms its about putting more bits where you are more likley going to notice the banding than just evenly distributing throught the image (linear intensity representation is generally a bad thing).

The problem with an 8bit video range ( with regard to film originated imagery) is that the film range has to be compromised into the video range. This involves trading off clipping against crushing: the more range you try to cram into 8bit the more chance of banding you get.

If you want to minimise banding you have to limit the range : clipping off information towards white and or black.

Sometimes you get banding , sometimes you get flat clipped blacks sometimes ( all the time really) you get harshly clipped whites. The only way you can fit more into video is to change the curve but again nothing is without cost.

The easiest way to fix all this would be to bump everything up a couple of bits (or even 1!) staying with 8bit all you can do are tricks which won't work as well for some imagery as others.

Essentially what I'm suggesting is increasing the range by changing lengthening the curve( call it gamma if you like) but its probably going to introduce banding on some imagery albeit with the benefit of a softer roll off into black and white. However I suspect its would look better than current video most of the time.

You can't really get round the bit depth issue although I do feel a 10bit strictly video format for domestic use would be a waste unless you used it more like a cineon colourspace.

Its 12bit. It has a different whitepoint , gamut and curve than video. Its designed to mimic and be reasonably transparent to a print film dynamic.

Its not video .Its not a simple question of nomenclature. Go call it video around people who actually use it and they will think you are a moron.

"Video is the technology of electronically capturing, recording, processing, storing, transmitting, and reconstructing a sequence of still images representing scenes in motion (http://en.wikipedia.org/wiki/Video)..."

It IS video no matter what you or anyone in the industry call it. Film is the actual film and any electronic version IS a form of video regardless of the format.;)

"Video is the technology of electronically capturing, recording, processing, storing, transmitting, and reconstructing a sequence of still images representing scenes in motion (http://en.wikipedia.org/wiki/Video)..."

It IS video no matter what you or anyone in the industry call it. Film is the actual film and any electronic version IS video regardless of the format.;)

So when video is a digitally encoded intensity step map what do you call video then?

A film scan , DI , and the usual Dcinema format is NOT video.( why don't you just read the article).

When you describe something as "video" you can infer one thing about it: its notionally a gamma corrected image designed to be displayed on a display with a CRT type intensity response ( ie about 2.5). That is all you can infer from the term "video".

If you describe an image as "linear" you can infer its a linear intensity scale and would require viewing on a display with an end to end gamma of 1.

If you describe an image as "film" ( unfortunately "log" has become a popular term even though it doesn't actually describe anything bar the encoding scheme) you can infer its going to have a simplified gamma around 1.7 and probably has a certain amount of headroom above a white reference.

You can have images which are described as being of the above that have never been anywhere near a realworld capture device or medium.

Interesting back-and-forth. My assumption is that some movies are captured as RGB data (film scans, or electronic-cinema cameras), and that others might be captured in a video format (such as heavily filtered Sony HDCAM tapes, although the HDCAM camera could output RGB data to a hard drive).

RGB data, AIUI, has the advantage of not undergoing the filtering and PQ reduction required for typical video codecs. D-cinema cameras can output either data (usually used) or sometimes video signals, or both simultaneously. Have the Feb. issue of the HDVideoPro before me, a magazine that covers both D-cinema and non-cinema items; assume that, editorially, they spell out recording formats when the distinction is important. -- John

Interesting back-and-forth. My assumption is that some movies are captured as RGB data (film scans, or electronic-cinema cameras), and that others might be captured in a video format (such as heavily filtered Sony HDCAM tapes..........
Hi John,
Just so people don’t get the wrong impression and leap to conclusions, from…………
http://forum.blu-ray.com/showpost.php?p=574477&postcount=38

“HDCAM SR records 1920 x 1080 linear OR Log, in 10 bit. It does not
pre-filter any input signal, nor convert Log input signals to linear for record, or otherwise alter the Log input signal in any way. It records what you give it using intra-frame compression for progressive pictures and intra-field compression for interlace.”

Mr. D, if you plan on attending the upcoming Digital Cinema Summit in Vegas shoot me a PM. B.T.W, I recently linked your insights onto another thread here over at the Blu-ray software forum…………………

http://www.avsforum.com/avs-vb/showpost.php?p=13320140&postcount=269

^^^^
Actually, it was a follow-up to the above, this is it…………………..

http://www.avsforum.com/avs-vb/showpost.php?p=13320303&postcount=271

Mr. D, if you plan on attending the upcoming Digital Cinema Summit in Vegas shoot me a PM. B.T.W, I recently linked your insights onto another thread here over at the Blu-ray software forum…………………

http://www.avsforum.com/avs-vb/showpost.php?p=13320140&postcount=269


Too busy doing Bond at the moment : some of us have to do the work and don't get to go to the junkets!

RGB data, AIUI, has the advantage of not undergoing the filtering and PQ reduction required for typical video codecs. .John

RGB or not is a seperate issue from the codec really. Most capture devices operate in terms of RGB ( as ultimately do displays: we never really look at anything that isn't RGB).

You can encode to "component" essentially losslessly if you wanted though so its not really a case of RGB itself being intrinsically higher quality on its own terms.

Hi John,
Just so people don’t get the wrong impression and leap to conclusions, from…………
http://forum.blu-ray.com/showpost.php?p=574477&postcount=38

“HDCAM SR records 1920 x 1080 linear OR Log, in 10 bit. It does not
pre-filter any input signal, nor convert Log input signals to linear for record, or otherwise alter the Log input signal in any way. It records what you give it using intra-frame compression for progressive pictures and intra-field compression for interlace.”

Agree. In mentioning HDCAM for movies, was thinking mostly of the early ones using that format. HDCAM and HDCAM SR indeed differ greatly. -- John

Too busy doing Bond at the moment : some of us have to do the work and don't get to go to the junkets!
Are you allowed to tell us anything about it :) (it's ok not)
Are you doing compositing/fx/grading etc? What res do they scan/composite at?

Are you allowed to tell us anything about it

No.

No problem :)

It has far more intensity variation towards black and a much softer roll off towards peak white. .
Then it must move resolution out of the middle part of the curve and look worse there than video.
I understand the desire to move more steps into the highlight area but not so much into the blacks. The blacks and dark parts look pretty good to me from video provided the display has real low black levels. The brightest parts look relatively bad to me on video (too harsh and flat).

Too busy doing Bond at the moment : some of us have to do the work and don't get to go to the junkets!
:p :p :p

And may you henceforth solely work in the 4K digital domain with the associated increased xxx rendering times to keep you off the forums and out of the pubs. :p

You do realize that you just alerted other astute members as to where you are now employed.

No.
Good answer.

When I’m not working, I go on vacation and just *mill* around myself……….
http://forum.blu-ray.com/showpost.php?p=698273&postcount=1753

Good answer.

When I’m not working, I go on vacation and just *mill* around myself……….
http://forum.blu-ray.com/showpost.php?p=698273&postcount=1753

Is that what you do when you are on the vacation? :)

4K=good, but there are other things to do with rendering that are also good/important. eg. I watched a film which was made in the USA (which shall remain nameless) and, even on SD DVD the big metalic thing in it looked like it was drawn in pastel colours unlike the rest of the shot that was/at least looked mostly real. So I think whatever rendering algorithms they use (like for metal structures/whatever cgi etc.) need to be right too.

Good answer.

When I’m not working, I go on vacation and just *mill* around myself……….



Nope I don't work for the Mill.

...I do feel a 10bit strictly video format for domestic use would be a waste unless you used it more like a cineon colourspace.Agreed!

State of the art film stock can have greater than 10 f stops of latitude (~1000:1), this made possible do to a lazy S pattern sensitivity curve which tracks color logarithmically like the human eye. Film has the ability resolve and record detail (contrast) in an over- or under-exposed frame.

Print film on the other hand (bleach by-pass process) has a nominal dynamic range of 200:1, but a print still represents said log properties, hence why a pristine film print to this day is still perceptually superior to 8 or 10 bit video.

As you ( Mr. D) have pointed out increasing the “video” bit depth to 10-bits will not really gain much improvement. This is due to the Rec. ITU-709 transfer curve that clips highlights and throws away the over-exposure headroom. Also a video gamma curve is really a tonal pre-emphasis, unlike the logarithmic nature of film. Such tonal intensities aka 'digital steps' are essentially linear and the abligitary artifacts/limitations of gamma corrected video reamain.

Gamma (non- lineararity) in video is really a voltage domain compensation for legacy CRT ext. and these digital intensity steps are effectively linear although perceptually shifted in a zero sum game. The “non-linear” ingredient only exist in the analog domain via a given LUT.

Nope I don't work for the Mill.
I know that.
I was speaking about where I hung out during part of the time I was across the pond………….last year.

Plus, I’ll take my canines ………………
http://forum.blu-ray.com/showpost.php?p=698273&postcount=1753

over your peoples’? canine --at least for imagery………….
http://www.youtube.com/watch?v=9beQh1yH5uU
Not to mention the location of the steering wheel.

Gotta run, as I have Arsenal supporters :) over to watch the Gunners match this morning and they’re already spilling beer on my carpet. :eek:

If you are asking about the difference between a release print and a Digital Cinema presentation:
http://www.etconsult.com/papers/Technical%20Issues%20in%20Cinema%20Resolution.pdf
...

Has anyone directly compared a release print to a blu-ray disc?

There's a chart near the end, which would imply that a blu-ray disc is just plain superior to a 1.85 release print -- horizontal release print resolution being 700 lines/height, vertical release print resolution being 525 lines/height, for an average of 610 lines, which quite a bit less than the 20% line for 2048 digital (blu-ray=1920) being about 900 lines @20%..

The numbers for a 2.39:1 anamorphic release print are horizontal resolution of 550 lines/height and vertical resolution of about 760 lines/height, or an average of 655 lines, due to the larger film size. The blu-ray, the numbers for 2.35:1 would be multiplied by 3/4 or 675. The implication is that an anamorphic theatrical release print has roughly equivalent resolution to the blu-ray movie for the same film.

Any comments?

Which release prints? They can vary quite bit.

Which release prints? They can vary quite bit.
“The better of the 35mm release prints” were measured by objective MTF curve resolution test to determine the best.

Based on the above, “the best 35mm release print” was subsequently displayed to at least 7 expert assessors for subjective scoring in multiple theaters.

See #149……………
http://www.itu.int/md/meetingdoc.asp?lang=en&parent=R00-SG06-C&source=Ch%20SG%206

Any comments?
I just skimmed the white paper by Matt (who b.t.w. was in attendance at the ETC’s DCL test that I linked earlier on this thread) and I’m not sure if he noted it but, the highest observable resolution that the expert assessors could discern in the sharpest part of the screen in the best performing movie theater venue with the release print was about 875 lines/PH.

Was that from flat or scope?

Has anyone directly compared a release print to a blu-ray disc?

There's a chart near the end, which would imply that a blu-ray disc is just plain superior to a 1.85 release print -- horizontal release print resolution being 700 lines/height, vertical release print resolution being 525 lines/height, for an average of 610 lines, which quite a bit less than the 20% line for 2048 digital (blu-ray=1920) being about 900 lines @20%..

The numbers for a 2.39:1 anamorphic release print are horizontal resolution of 550 lines/height and vertical resolution of about 760 lines/height, or an average of 655 lines, due to the larger film size. The blu-ray, the numbers for 2.35:1 would be multiplied by 3/4 or 675. The implication is that an anamorphic theatrical release print has roughly equivalent resolution to the blu-ray movie for the same film.

Any comments?
That linked paper, based on an older U.N. ITU study, further detailed in this pdf (http://www.cst.fr/IMG/pdf/35mm_resolution_english.pdf), shows the measured projected horizontal resolution for four theaters to average ~711 lines/PH for flat release prints after standard processing. That's 711X1.78=1266 lines horizontal resolution (vertical lines) in 16X9 HD terms and 711X1.85=1315 horizontal resolution in film-ratio projection terms.

Joe Kane, at his site (http://www.videoessentials.com/D_TheaterQA.php), wrote:
The majority of movies that Hollywood is making available in the D-Theater format are mastered in 1080p then converted to 1080i before they are mastered and duplicated. The irony of this is that the majority of 1080p film transfers only have enough resolution to support the horizontal capability of 720p. Horizontal resolution of most film masters in 1080p is in the area of 800 to 1300 lines. The horizontal capability of 720p is 1280 lines, the top end of what’s on the 1080p master. An argument could be made that even with the 30% vertical filtering in the 1080p to 1080i conversion the vertical resolution of 1080i is still slightly better than 720p. It’s 756 lines as opposed to 720 lines. Interlaced artifacts in 1080p/24 frame material converted to 1080i/60 are not nearly as bad as video material initially created in 1080i/60.
Believe he was referring to 1080/24p HD-D5 masters at ~270 Mbps, these days likely advanced to HDCAM SR masters, etc. About this time sspears reported spectrum analyzer measurements of 800--1300 lines (http://archive.avsforum.com/avs-vb/showthread.php?postid=2498224#post2498224) maximum effective horizontal resolution from master tapes he measured. This more detailed earlier discussion (http://www.avsforum.com/avs-vb/showpost.php?p=9314235&postcount=24) outlines why I concluded these measurements represent lines per picture width (not height), and why 800 lines might be typical, 1100 lines exceptional, and the maximum, 1300 lines, represent computer-animated images.

So perhaps if 1100/PW lines were exceptional maximum effective horizontal resolution ~5 years back (sspears' measurements), this might represent typical Blu-ray resolution these days. Unless someone plugs Blu-ray outputs into a spectrum analyzer, it seems like mostly guesswork. Supposedly, producing some Blu-rays from 4k D.I. downconversions should provide some scenes exceeding the limiting resolution of HD (about 1700 lines with standard 74-Mhz sampling). But from comments about Blu-ray PQ, it doesn't read like observers feel resolvable detail has doubled to 1600, (assuming 800 to, say, 1000 lines, is typical to most non-disc HD movies). Pick a number. :-) -- John

Mr.D,

What do you think about using S shaped gamma curves? David Richards published a paper while he was at Christie in the digital cinema engineering department that proposed using these to provide more gradual black roll on and white roll off...a typical film look. They also allow using a higher gamma in the middle.

Mr.D,

What do you think about using S shaped gamma curves? David Richards published a paper while he was at Christie in the digital cinema engineering department that proposed using these to provide more gradual black roll on and white roll off...a typical film look. They also allow using a higher gamma in the middle.

I'd rather they just used the correct curve for whatever material they happen to be displaying. You can't bring it back if its not there whatever you do to the curve.

The projector should already have it's own properly calibrated curve so that it can display the content correctly. Let the colorist/dp/director determine the look.