DemoCoder said:
This is the whole point of "tweak where needed" and it has nothing to do with GOP.
Not if you are doing a multiple generation rendering.
If you encode a video in mpeg2 and after that you take themovie and recompress it again to "tweak it " the Long GOP format make you lose a lot of quality.
The mpeg2 codec is not made for multiple rendering, every time you render again the same movie, you have a quality loss, and this have all to do with the gop .
It has to do with the fact that automated workflows produce suboptimal quality. An engineer's job is to find scenes with suboptimal quality and adjust the bitrate or other parameters to fix the problem.
Not true.
More comprerssion = more artifact = more work
When you use less compression you have less artifac, and you wont need to find scenes with suboptimal quality and adjust the bitrate or other parameters to fix the problem if there are no problem to fix.
Using an enough hi bitrate like 80mbit sec, you wont have any artifact at all .
So, who was claiming that you could? You're arguing with yourself.
You quoted this from another post putting it total out of his original context.
H.264 can produce better quality than MPEG-2 at both high bit rates and low bit rates. It is better across the board. Objective PSNR studies show it. Subjective perceptual tests show it. Both the SMPTE and ISO/ITU accept this.
If you talk about hi bitrate condition, this is a professional realm and you have alredy professional codec used in those condition.
Dvcpro hd , hdcam sr , avid hd , all of those professional codecs blow h.264 out of the water in temr of quality, and of course blow mpeg2 also.
So why waste time trying to adapt h.264 to do what you can alredy do with those codecs and with a better result ?
It still hasn't stopped people from encoding video with MPEG-2 at 1Mbps or less. There are plenty of PVRs out there that consumers buy in droves which record at <2Mbps with horrible artifacts.
This dont change the fact that it is an mpeg2 limitation , it can't have enough quality with this
bitrate, this is what i stated.
Yes, it is correct. GOP is irrelevent. If an engineer finds that a particular transistion or scene is causing problems, he can either increase VBR bitrate for that section, or apply pre-filters to the video before compression to try and alleviate the problem.
Before the compression, not after. He never stated before the compression, he said
after the compression is alredy done.
here what aaronspink said "
When the compressors are done, they then review the actual result and tweak where needed."
If you recompress again the alredy compressed mpeg2 movie you lose a lot of quality
because of the GOP compression so it is relevant
Again, you run off arguing a strawman. No one claimed you'd tweak an "already compressed mpeg-2 video". What's being claimed is that in the process of encoding the master to MPEG-2, many steps are taken. It is not "click and go". The output of a given encoding run is typically reviewed, and then recompressed using the tweaks.
Hum , no , aaronspink stated clearly he was talking about an alredy done compression :
When the compressors are done , they then review the actual result and tweak where needed."
Only because you continue to not slow down, read, and use your brain. You keep making blanket assumptions, like the assumption I am arguing that mpeg-2 artifacts will be fixed by constantly recompressing mpeg-2 recusively instead of going back to the master each time? I mean, how the hell do you come up with such drivel?
If you say "When the compressors are done " to me means when the work is done.
So? The fact that you can see artifacts doesn't mean that video artifacts are worse than audio artifacts. Audio artifacts are in fact, far more psychologically unnerving.
It mean that is pretty easy to see a video artifact when the video became all pixellated,
and with the same compression ratio you may not hear any difference in the audio but you
can tell the difference for the video.
And not only that : if you compress an audio file 10 times with the same bit rate , it will still sound good,
but if you recompress the same video with the mpeg2 codec 10 times it will look like shit, and it will be a lot more easy tell that the video have a visible artifacts and not easy to tell that the audio file was compressed 10 times.
With a multiple rendering video lose a lot more quality than audio.
Compare this to loss of sound, pops and squeaks, and frequency shifts in audio.
This is due to a poor source it is not an effect of the compresssion.
audio shifts are due to a poor mixing, is not an effect of the compression.
Is like if i say compare this to the fact that the video became all green...if the video became green because i made a bad recording this is not a compression defect.
I can't live with mpeg-2 artifacts whether it's small screen or large screen. You practically have to put it on a 9" TV to hide it. Perhaps you mean *POOR QUALITY SMALL SCREEN*. On a 15" LCD screen, viewed at THX recommended distances, the artifacts are still visible.
You dont have any video artifact with mpeg2 at 80Mbit/sec. I had seen by myself a lot of Dcinema movie and i never seen any type of artifact.
You have artifact only at low bitrate with mpeg2.
Don't try to change the discussion because you were proven wrong. Digibeta is irrelevent to the discussion. The discussion is MPEG-2 vs H.264. I compared MPEG-2 to H.264 on a large screen using a large venue digital projector that I own, that outclasses the projectors in many dcinemas,
Sorry but what projectior di you have that outclass the $80,000 D-Cine Premiere DP100 ?
http://www.barco.com/digitalcinema/en/products/product.asp?element=1721
and the result is that H.264 delivers less artifacts than MPEG-2 at the same bitrate. Period.
It will never have less artifact at 80Mbit/sec, period, any dcinema movie i seen never had a single video artifact.
Star wars Episode 3 , I robot, costantine, no one had any visual artifact.
If you say that the Dcinema movies have video artifact what you are saying is pure bullshit.
Irrelevent. No one is going to ship dcinema encoded optical formats.
Very revelant, it is the same codec and a dcinema movie can fit on a 100bg blu ray disk.
Star wars episode 3 is encoded in an 2048x1080 mpeg2hd file with 80mbit/sec, and it take 80 gigabyte.
So with a 100gb blu ray disk you can have a Dcinema movie quality.
Very revelant
They won't even ship 4:2:2 encoded lossy video.
You absolutly have enough space to allow this since star wars episode 3 take about 80 gigabite.
Wuth a 100bg blu ray disk you can fit the same dcinema movie, and if you use half the bitrare to 40mbit/sec, you can store the movie on a 50gb disk.
You are not very clear about what your point is. First you're arguing about how superior MPEG-2 is for Sony in BluRay,
No i argued about how the less compressed mpeg2 give you a better quality at an hi bitrate.
I argued about how a more compressed codec dont give you more picture quality
Sony Pictures’ senior vice president of advanced technology Don Eklund apparently said:
"Advanced (formats) don’t necessarily improve picture quality"
but Sony is shipping bog-standard MPEG-2/4:2:0 at a slightly higher bitrate than ATSC broadcasts. It's not compelling.
Hum no, not in the long period with a 50 and a 100gb blu ray disk availbe in the market.
Yea, outsourcing your work never is.
My work ? what do you think post production studios exist for ?
He did not claim that Sony was going to ARCHIVE their video libraries to MPEG-2 masters. He was talking about the fact that studios have a huge amount of experience starting with analog film material and generating DVDs from them.
But is is absolutly not clear. If you say you move "to digital medium" since most of the library is on 35 film stock, to me means that you move your analog librady to the digital medium, and you digitalize it compressing it to mpeg2.
I am not saying you are wrong, but he was not clear on this point, and add to this the fact that he also said you tewak the video after that compressor have done all the work,
it is even less clear.
Well, that's your poor English ability unable to grasp the nuances of the language. When I wrote computer programs, if I find a bug or issue, I can go back and "tweak it". That means recompilation. It doesn't mean hex-editing binary EXEs.
If you say "When the compressors are done"
Once the compression is done is done.
Not wrong. Disney's desire to for larger media is separate from the codec they to use (Disney wants to put lots of extras and interactive media on the disc), which is going to be based on ECONOMIC considerations, like production costs and availability, not to mention POLITICS. In case you missed it. Disney until recently was firmly in Microsoft's camp, a big supporter of HD-DVD and VC-1 codec. Even though Disney claims they will now support both BluRay and HD-DVD, Disney's IP stake in iHD with Microsoft, and their interest in getting content on Windows Vista and XBox360, virtually assures VC-1.
Nice theory. If you want ot use Vc-1 you wont need a 50bg blu ray disk.
And not only they want that, they stated that they want starting release video only to the the 50bg .
50GB is a pipe dream for launch. Studios haven't even be able to be assured that scratch resistant spin-coated single layer discs can be mass manufactured at affordable prices.
"Panasonic kicks off 50GB Blu-ray production"
http://www.engadget.com/entry/1234000293071193/
It wont take a log time to see the 50gb begin used.
Of course, because Sony doesn't want to pay patent fees for H.264 of VC-1, whereas Sony owns IP stakes in MPEG-2. It's also cheaper to produce MPEG-2 at the moment.
Cheaper because it is less compressed, it show less artifact and it dont have need a long encoding wher you have to fight aganist the codecs problmes.
Higer is the bit rate, less artifact can appear , more fast cheaper the encoding