or Connect
AVS › AVS Forum › Display Devices › Flat Panels General and OLED Technology › TV Signal Compression In The Real World
New Posts  All Forums:Forum Nav:

TV Signal Compression In The Real World - Page 5

post #121 of 270
Thread Starter 
Holy crap on a cracker. I had to go back and read my original post to even remember what I asked. In referring to my last sentence in that post, regarding TV broadcasts, is there anything on the horizon that we can look forward to that will provide better PQ from the provider or are we stuck with the status quo because of financial restraints?
post #122 of 270
Quote:
Originally Posted by andy sullivan View Post

Holy crap on a cracker.

 

In the spirit of diversionary OT replies, let me take a moment to thank you for my new phrase for the year.

post #123 of 270
Thread Starter 
I think I first heard that phrase on "Everybody Loves Raymond".
post #124 of 270
Quote:
Originally Posted by gregzoll View Post

...

I just purchased the remastered version of Beetles "Help!". They did a good job cleaning it up, leaving some of the original film noise, original pseudo stereo mix that it was filmed in, but still is not a true 1080p/24 master. A true 1080p/24 disc, will be mastered from original 24fps stock or digital filming, which keeps it 100% the same all the way.

....

Who are these "Beetles" you speak of? Insects who needed Help? Not the only one who needs help...

Help! was originally filmed with 35mm film at a 1.75:1 aspect ratio at 24 frames per second. It was transferred to Blu-Ray with a 1080p resolution with a 1.67:1 aspect ratio at 24 frames per second. Music was mixed from the original multitrack tapes into 5.1-channel surround. Dialog and effects were also remixed from the original mono analog film recordings. No pseudo stereo was used.

So Help! started at 24 fps and shows up on my projector at 24 fps with a 24 Hz refresh rate.

When I create a Blu-Ray disc I encode it so that every 1/24th of a second there is data so that a new frame is drawn. That is what is meant by a Blu-Ray is encoded at 24 frames per second. If a player of any variety then modifies that, it does not change that the information on the disc is encoded at 24 fps.
post #125 of 270
Quote:
Originally Posted by andy sullivan View Post

Holy crap on a cracker. I had to go back and read my original post to even remember what I asked. In referring to my last sentence in that post, regarding TV broadcasts, is there anything on the horizon that we can look forward to that will provide better PQ from the provider or are we stuck with the status quo because of financial restraints?

Let me try to bring this thread back on topic.

Anyone who was around during the launch of HDTV broadcasting (late 90's, early 00) can tell you that picture quality was at its peak during those years. It has steadily declined since then as a result of more channels and sub-channels being added. Early on, there was also no stat-muxing. Each channel had dedicated bandwidth that it didn't have to share. I did a lot of HDTV satellite recording during those years and you could literally record the same program multiple times and it would be bit-for-bit identical to a previous broadcast. The providers were simply passing through what came from the studios without messing with it. In theory, a channel like HBO could pre-encode a non-live show/movie using advanced multipass offline encoders and splice that into their live broadcast. The result would be a similar increase in quality to what you see in other non-live downloadable content.

Of course those days are long gone and I'm not sure if we'll ever return to that level of broadcast quality again. Even if we were to switch to HEVC, the providers would eventually cram in more crappy channels and we'd be right back where we started. They would also be using real-time encoders which are not as efficient as something that could be done offline at the source.
post #126 of 270
Thread Starter 
I imagine you could be correct but i hope we may have reached a saturation point of potential crappy channels.
post #127 of 270
Quote:
Originally Posted by andy sullivan View Post

I imagine you could be correct but i hope we may have reached a saturation point of potential crappy channels.

The vast majority of programming now is "Reality" shows which are dirt cheap to produce. They can crank out enough of this type of crap to fill hundreds of channels. If an "a la carte" subscription model ever takes off, maybe we'll see those rarely watched channels disappear - I'm not holding my breath waiting for that to happen.
Edited by Wizziwig - 8/25/13 at 1:33pm
post #128 of 270
Quote:
Originally Posted by Wizziwig View Post

If an "a la carte" subscription model ever takes off, maybe we'll see those rarely watched channels disappear - I'm not holding by breath waiting for that to happen.
There is nothing technologically that prevents this to happen, but greed. Cable companies keep insisting that it will be 'more expensive', just to hide the fact that now we are subsidizing with our fees 90% those stupid channels. Of course, what I call stupid might be great shows for others, but that's what really makes markets evolve - demand and offer. Not Nielsen make-believe.
People voting with their wallets.
post #129 of 270
Every time I see someone invoke the OWS-esque "greed" mantra as if it were somehow a bad thing I want to scream "401k" at the top of my lungs...
post #130 of 270
Thread Starter 
Yea I agree. In our society we have always adhered to the mantra, "bigger is better", and "more is better". Doesn't matter what constitutes the "more" Have you seen the Geiko commercial with the adult sitting around the table with a bunch of little kids asking them if big is better? If you offer most of us 120 or 150 channels for the same price most will jump on the 150 without even asking what the extra 30 channels are or even if they are giving up some channels they like in the process. Sometimes greed just equals good business.
post #131 of 270
Um, that's an AT&T commercial for cellphones and cellphone coverage. Geiko is the little green gecko (or the pig) who sells insurance.

So, apparently neither companies' advertising is very effective!
post #132 of 270

I guarantee that the grandparents who are surviving on whatever mutual funds they were able to create during their lifetime are (rightfully) hoping their investments are in companies managed by the geediest bastards who ever lived.

 

....moving on....
 

post #133 of 270
SD analog on satellite use to look a lot better than SD digital now, And HD channels now are just like an average widescreen DVD, That's why I kept my blu-ray monthly subscription to blockbuster, it's all about quality, I hope blockbuster don't go out of business because I will have to pay extra for them with netflix and usually their blu-ray library is poor compared to blockbuster.
post #134 of 270
Quote:
Originally Posted by latreche34 View Post

SD analog on satellite use to look a lot better than SD digital now, And HD channels now are just like an average widescreen DVD,

 

That's a bit of exaggeration I think.  My HD 1080i horribly compressed channels look better than the DVD's 480i.

post #135 of 270
Thread Starter 
Horribly compressed is the key phrase here I think. What is/can be done to rectify this problem? What would have to happen to allow for a totally uncompressed signal? What provider is delivering the best signal to it's subs today?
post #136 of 270
Quote:
Originally Posted by andy sullivan View Post

Horribly compressed is the key phrase here I think. What is/can be done to rectify this problem? What would have to happen to allow for a totally uncompressed signal? What provider is delivering the best signal to it's subs today?

 

You'll never have a "totally uncompressed" video signal available to the mass market.  Besides, it's not needed and very much something I'd rather not see.  4:4:4 is good for computer use, but IMO complete overkill elsewhere, and the result of trying to support it would be to move 4K to the back burner.  4:4:4 would throttle down the availability of HFR too (another thing that eats bandwidth like candy).  Totally a lose-lose.  Keep in mind that even Blu-Ray is chromasubsampled.

post #137 of 270
Uncompressed just ain't happening.tongue.gif

Uncompressed RGB video requires a byte indicating the intensity for each primary in a pixel. So file size = width * height * 3 bytes per pixel * frames per second. That comes to ~ 112 GB per hour just for SD (720x480). Way too much data there, no? And HD will be even bigger.

Two ways to reduce that:

YUV color system. Y is grayscale brightness, while U and V are color. Since the eye has less resolution for colors, color data can be safely (?) reduced. So, for example, with 4:2:2 subsampling, for every 4 Y values, there are 2 U and 2 V. That means only about two bytes per pixel. Therefore you need about 2/3 the data in the above case, or ~ 75 GB per hour. Still too much. So there's 4:2:0 subsampling, which gets it down some more.

Compression. This is the process of removing redundancies in digital data to reduce the amount that must be stored or transmitted.
Lossless removes no more data than is consistent with full restoration of the original data. That's still going to result in too much data.
Lossy compression. This sacrifices additional data to achieve smaller file size. And that's what you'll get.

Last time I checked, the current preferred broadcast *submission* format for 1920 x 1080 is XDCAM-HD. (If, for example, you're submitting an advert). That is, 50 Mb/s MPEG-2, 4:2:2. Still the industry standard, at least for the USA and UK (BBC). What further processing the content providers/broadcasters do with it after that is their business. Re-encoding at lower bitrates and changing the color subsampling to 4:2:0 can further reduce bandwidth requirements. BTW, Blu-Ray is either MPEG2, AVC, or VC-1, 8 bit, subsampled at 4:2:0.
Edited by fritzi93 - 8/29/13 at 10:10am
post #138 of 270
Quote:
Originally Posted by fritzi93 View Post

Two ways to reduce that:

YUV color system.

 

I'm an American Supremacist :), so of course you're referring to SD NTSC, in which case you might mean YIQ.  I think PAL uses YUV.  YIQ was super intriguing to me and I believe absolute genius.  It's where I first came across the

.30  .59  .11
.60 -.28 -.32
.21 -.52  .31

matrix.  (Most importantly the R/G/B luminance equivalents of 30%/59%/11%

post #139 of 270
I'm no engineer, merely a video hobbyist. I know a thing or two about capturing and filtering/encoding though.

Just giving a gloss from memory on why uncompressed is impossible as a delivery format. I'm afraid I'd have to bone up on the subject to continue the conversation. tongue.gif
post #140 of 270
Quote:
Originally Posted by tgm1024 View Post

You'll never have a "totally uncompressed" video signal available to the mass market.  Besides, it's not needed and very much something I'd rather not see.  4:4:4 is good for computer use, but IMO complete overkill elsewhere, and the result of trying to support it would be to move 4K to the back burner.  4:4:4 would throttle down the availability of HFR too (another thing that eats bandwidth like candy).  Totally a lose-lose.  Keep in mind that even Blu-Ray is chromasubsampled.
I don't see why 4:4:4 colour would be a problem with compressed video - it might help compression since it should be easier to track objects and have the full quality image to compare against when tracking/compressing and I don't think compressed HFR would need a lot more bandwidth. The more frames per second, the less difference between frames there is, so it should be easier to compress (assuming there isn't a big increase in noise) and they can use bigger GOPs.

Also there's been EBU tests of HFR UHDTV recently up to 240 fps:
http://tech.ebu.ch/Jahia/site/tech/cache/offonce/news/subjective-testing-confirms-importance-o-07aug13

with both uncompressed and compressed,
Quote:
Early results give a clear indication that higher frame rates are appreciated by the observers, to a significantly greater extent than increased resolution. This supports the position that future UHDTV systems should include higher frame rates
I think at the same bitrates, the higher frame rate one will usually be rated higher (better quality) and so HFR won't normally need a higher bitrate to appear better quality (in terms of UHDTV picture accuracy) than the lower fps version.
Edited by Joe Bloggs - 8/29/13 at 2:13pm
post #141 of 270
Quote:
Originally Posted by Joe Bloggs View Post

Quote:
Originally Posted by tgm1024 View Post

You'll never have a "totally uncompressed" video signal available to the mass market.  Besides, it's not needed and very much something I'd rather not see.  4:4:4 is good for computer use, but IMO complete overkill elsewhere, and the result of trying to support it would be to move 4K to the back burner.  4:4:4 would throttle down the availability of HFR too (another thing that eats bandwidth like candy).  Totally a lose-lose.  Keep in mind that even Blu-Ray is chromasubsampled.
I don't see why 4:4:4 colour would be a problem with compressed video - it might help compression since it should be easier to track objects and have the full quality image to compare against when tracking/compressing

4:2:2, or 4:2:0 are themselves forms of compression.  Your point is lost on me: you either use more space storing 4:4:4 or you don't, and you do.

 

Quote:

and I don't think compressed HFR would need a lot more bandwidth. The more frames per second, the less difference between frames there is, so it should be easier to compress

 

You have a unique way of looking at things that just strikes me as completely backwards.  Given any degree of non-time based (per frame) compression, there is more data required when there are more frames evident.  If you're looking at something akin to an MPEG stream, where the compression follows forward in both time and space, then there still is more data required for more source frames.

post #142 of 270
Quote:
Originally Posted by tgm1024 View Post

4:2:2, or 4:2:0 are themselves forms of compression.  Your point is lost on me: you either use more space storing 4:4:4 or you don't, and you do.
Imagine someone gave you a video and wanted it compressed to eg. 40 Mbps with the highest picture quality. Would it help if the video they gave you had missing pixels or missing lines of pixels for what the colour should be? Like how encoding progressive is more efficient than interlaced I'm saying give the compressor enough of the original, and allow it to compress more efficiently. If an object moves 1 pixel in any direction, and is otherwise unchanged, with 4:4:4 colour you already have all the required pixel values for it.
Quote:
You have a unique way of looking at things that just strikes me as completely backwards.  Given any degree of non-time based (per frame) compression, there is more data required when there are more frames evident.  If you're looking at something akin to an MPEG stream, where the compression follows forward in both time and space, then there still is more data required for more source frames.
The broadcasters have said there is less difference per frame. We'll be able to see when the EBU give the full results of their tests, but I imaging they will show, based on what they have said already for the UHDTV tests ("Early results give a clear indication that higher frame rates are appreciated by the observers"), that at a given compressed bitrate it will look better with the higher frame rate (up to a point), and even if there is a slight increase in compression artefacts per frame - but with shorter time per frame with higher fps, they will be less noticeable unless in the same place.
Quote:
If you're looking at something akin to an MPEG stream, where the compression follows forward in both time and space, then there still is more data required for more source frames.
But with mpeg you are compressing the differences between frames. Also more fps allows the compressor to track objects better.
With Blu-ray at 40-60 mbps I'd say you'd get better quality, more accurate with compressed higher than 24 fps video than with 40-60 mbps 24 fps video.

eg. a 30 fps 40 Mbps compressed 1080p video wouldn't need more bitrate than a 40 Mbps compressed 24 fps one, but it would look better in terms of realism (though higher fps would look better).

Low fps gives judder/strobing, higher fps is going to look better (such as in the EBU tests) due to the lack of judder/strobing and more realistic look, even with the same bitrates I think.
Edited by Joe Bloggs - 8/29/13 at 3:14pm
post #143 of 270
Chroma subsampling is the last thing we need to worry about in most cases. If you look at what madVR is doing on PC, you can see that its upscaling algorithms are able to create an image that requires 4:4:4 to be displayed without losing any quality.
4:4:4 video would obviously be better, and I would not accept anything less when displaying video games (which use color very differently, and it's obvious if chroma is reduced to 4:2:2 by a display) but the difference between 4:2:0 and 4:4:4 with video content is less than you might think.
Increasing the source resolution would yield much larger improvements, and of course, so would using better encoders, less video compression, higher bit-depths.


And even if you were to just stick to your 1080p display rather than upgrade to 4K, upgrading to 4K source material would bring you similar benefits. 4K using 4:2:0 subsampling would have a 1920x1080 resolution chroma image, so that could be 1:1 mapped. And downsampling 4K video to 1080p would mean that compression artifacts would be 1/4 the size of 1080p video.
post #144 of 270
Quote:
Originally Posted by Joe Bloggs View Post

The more frames per second, the less difference between frames there is, so it should be easier to compress (assuming there isn't a big increase in noise) and they can use bigger GOPs.

You're right about needing longer GOPs (Group Of Pictures) for HFR, if you're going to keep file size from ballooning. Otherwise your keyframes (I frames) will be (at least) doubled, and that's where a goodly portion of the bits go. But I'm not seeing how this really helps with file size. Your P (Progressive) and B (Bi-directional) frames will be more than doubled by, say, going from 24 to 60, each frame described by motion vectors and error correction. I don't get that bit about subsampling either. You propose going from less precision to more precision. How does that improve compression?

Mind you, I'm all for higher frame rates in principle. cool.gif
post #145 of 270
Quote:
Originally Posted by fritzi93 View Post

You propose going from less precision to more precision. How does that improve compression?
It lets the compressor know more about what the scene is actually like. It will also be able to track (for mpeg compression) objects more easily, even if the target bitrate is the same.
By giving the compressor as much information as possible about the scene, and letting the compressor decide how it can be reduced (instead of throwing bits of the picture away before the compressor has even seen it).

It's a bit like oversampling. If the compressor has a higher quality source, it will be able to produce a better output.
post #146 of 270
Quote:
Originally Posted by Joe Bloggs View Post

It lets the compressor know more about what the scene is actually like. It will also be able to track (for mpeg compression) objects more easily, even if the target bitrate is the same.
By giving the compressor as much information as possible about the scene, and letting the compressor decide how it can be reduced (instead of throwing bits of the picture away before the compressor has even seen it).

It's a bit like oversampling. If the compressor has a higher quality source, it will be able to produce a better output.

Your argument is invalid. Encoders only use the luma portion of the frame for motion estimation.

Ron
post #147 of 270
Quote:
Originally Posted by dr1394 View Post

Your argument is invalid. Encoders only use the luma portion of the frame for motion estimation.

Ron
What about H265 and chroma prediction blocks? Isn't that using the chroma portion for motion estimation?

Isn't it going to help with compression if an object moves 1 pixel in any direction (without any other change) if, with 4:4:4 colour you already know each colour value but with 4:2:0 you won't?
Edited by Joe Bloggs - 8/29/13 at 6:58pm
post #148 of 270
Quote:
Originally Posted by Chronoptimist View Post

... the difference between 4:2:0 and 4:4:4 with video content is less than you might think.

Now that's really interesting. Care to elaborate a little? There seem to be a good many members clamoring for colorspace improvements in UHD.

I confess I never worried much about color space or subsampling. When doing encodes, I scarcely thought about it at all. I recall when using CCE, there was the (internal only) colorspace conversion to YUY2, and TMPGenc was RGB all the way (which made it so slow) until final output. Lots of other things to be more concerned about. Hey, just so the output is compliant.tongue.gif

@TGM: I looked up YIQ, and you're right. It appears to be more efficient than YUV, weighting more bits towards orange-blue (I) than purple-green (Q). To the parts of the spectrum to which the eye is most sensitive, in other words. Anyway, that's for broadcast, which explains why I initially drew a blank on the term.
Edited by fritzi93 - 8/29/13 at 9:21pm
post #149 of 270
Quote:
Originally Posted by fritzi93 View Post

Now that's really interesting. Care to elaborate a little? There seem to be a good many members clamoring for colorspace improvements in UHD.
Colorspace (color gamut) and chroma sampling are two very different things.

Chroma sampling is the resolution of the color component of a TV signal. 4:2:0 means that it is 1/4 the resolution of the luma (detail) component.
4:2:2 means it is half the resolution (horizontally) and 4:4:4 means that it is the same resolution as the luma component.

Colorspace is the range of colors that the display is capable of showing. A wider color gamut (colorspace) means that the content can contain more vivid and more accurate colors. (because there are things which have color beyond the limits of the current BT.709 colorspace)
post #150 of 270
Would 4:0:0 look inferior to us?
New Posts  All Forums:Forum Nav:
  Return Home
AVS › AVS Forum › Display Devices › Flat Panels General and OLED Technology › TV Signal Compression In The Real World