PQ differences: Codec or Mastering related? (was "AVC better than VC-1???") (1 Viewer)

[Jari K]

I fully agree. Since people throw these "numbers" in a different threads, here´s one from me; 80% of people claiming that one codec is *clearly* better than the others - and just make the film look "better" just like that - are basing their info mostly on online rumors and stuff that they´ve read in another threads ("VC-1 is so cool!" etc). Sure, there are mediocre releases (in both formats, I might add), but in most of them the codec *alone* is not the reason. Don´t believe the hype.

 

[DaViD Boulet]

???

For whatever reason, AVC titles, on average, look better than VC-1 (not only the reviews at dvdfile, but when Dan talleyed the scores from other review sites, he found the same trend with AVC looking better overall). Now, this isn't a proof that the codec is responsible... it could be source material or studio-mastering related even across such a large sample. But it certainly would seem odd if AVC was filtering by default considering that Sony's and Disney's AVC titles have more fine detail (on average) than most VC-1 discs from WB.

Not to mention that the whole reason I started this thread was because the AVC-compressed scenes on Shrek 3 looked BETTER than the VC-1 compressed feature film on the same HD DVD. Again, this doesn't proove that one codec is better/worse because there could be other factors (and probably are). However, it would strongly indicate that no "filtering" is being applied defacto to AVC given than, in this case, it's more detailed than the VC-1.

 

[Shane Martin]

David,
Your comparing apples and oranges really. I think I know where you are headed with it but I think it's a reach really.

Wasn't The Prestige encoded on VC1 for HD(overseas) and AVC here in the states(if memory serves)? The people who have compared the 2 said there is no difference. That's an apples to apples comparison. The Paramount discs are another. AVC(BR), VC1 HD DVD.

Even MPeg2 can look great. It's all down to who does the encoding and the source. Not much else.

 

[DaViD Boulet]

Please don't assume you know where I'm headed.



Naturally encoding methodology makes a world of difference with any codec. That doesn't mean all codecs respond the the same way to the same source material at the same bit-rate. It's possible that given optimal conditions, both AVC and VC-1 are identical. It's also possible they aren't.

I think it's pretty telling that the BONUS MATERIAL on Shrek (which happens to be AVC) LOOKS BETTER than the FEATURE FILM (which happens to be VC-1). In my very opening post I stated that this difference was not conclusively a codec issue for all the reasons cited. My intent was to open a good topic for conversation and questioning given this specific example.

If it turns out that the fault of these differences lies in the hands of compressionists who filtered away HF detail from the feature film (which just happened to use VC-1) but left HF detail in-tact on the bonus material (which just happened to use AVC), then that's ALSO cause for some debate.

Is everyone just comfortable that your bonus material looks better than the same scenes in your feature presentation????

We put up with that on DVD... Phantom Menace and The Incredibles are just two examples with feature films that had image quality below the level of what was seen with the *same* scenes in the suppliments. They're also two examples of DVDs that were lavished with praise for being so visually perfect when they were first released despite being obviously compromised in video for anyone who bothered to compare with the suppliments (Yes, take The Incredibles DVD off your shelf and compare with its own suppliments... the movie is filtered substantially compared to scenes in the bonus material!).

Sounds like an important topic to discuss no matter what the cause. And asking questions about it is the first step to discovering what was the problem.

 

[ppltd]

If it does not prove anything (which subjective views never do) just what was the reason for starting a thread that basically states that AVC is better than VC1 even though it really can't be proved?

 

[Douglas Monce]

Have you seen the Shrek disc yourself or are you just relying on some review from the internet?

You think supplements look better on The Incredibles than the feature? I don't know I find that strange. I never noticed any difference, and I've watch the feature and the extras over and over again. It's one of my favorite movies. Frankly I think The Incredibles is one of the best looking DVDs ever released.

Doug

 

[Shane Martin]

David,
If you keep going down the road of apples and oranges, then we'll go nowhere. If you want to discuss real world examples like Flags of Our Fathers or The Prestige which are primo examples of "one format has VC1 and the other has AVC" then we're comparing apples and apples and we can certainly go down that road.

Comparing different video codecs would be like me saying "The special features sound better in DTS than the movie in DD".

 

[DaViD Boulet]

How is comparing the same scene from the same movie any more "apples and oranges" if it's on the same disc versus if it's two different releases? If your example of flags of our fathers is "apples" when why isn't the bonus scenes on this disc?

I've already said, in my first post and again and again, that nothing is proved by this or any other example: only that it raises questions that are good to ask.


I haven't seen the disc myself. Have you compared the bonus material to the feature film on a wide-angle system to take issue with Dan Ramer's assessment? Dan Ramer is format-agnostic and calls the shots like he sees them on his Sony Ruby projector running 1080p. His reviews are consistently on-the-mark as far as picture and sound quality impressions are concerned.

 

[troy evans]

I will agree that I have noticed variable differences between supplemental material shots and feature dvd presentation. The one that stands out significantly for me is B5 series sets. The preview trailers of episodes look better than the episodes themselves. Now, is this because of codec differences, no. "Phantom Menace" is horrible compared to all other Star Wars films as far as PQ. So, if it's not because of codecs on sd dvd,then, how does all this relate to the "AVC better than VC-1" in regards to HDM ?

 

[Cees Alons]

Fact? And if so, reason?


Cees

 

[DaViD Boulet]

Oh I don't know Cees. It's been happening ever since the first time I suggested that Sony might actually be able to replicate dual-layer Blu-ray Discs. Come to think of it, the same thing happened back in 1999 when I suggested that 4x3 interlaced TVs shouldn't be considered the "reference standard" with 16x9 component DVD software.

I say we hold the studios accountable to deliver the highest qualtiy feature presentation a format is capable of delivering. If a feature's suppliments look better, there's a problem. If it's CODEC related or mastering related, we should discuss and raise awareness of the discrepency until such problems are understood and cease to exist.

 

[RobertR]

You shouldn't conflate the fact that that some of the things you say are debatable (and therefore debated vigorously by others) with personal attacks so much, David. There's way too much of that on the Internet.

 

[ppltd]

I did. See my comments above on the uselessness of tallying PQ numbers of differing titles. Absolutely no conclusion about video codec capabilities can be derived form this.

 

[Douglas Monce]

Trying to find out if AVC or VC-1 is better based on internet review scores is unscientific at best, and frankly just plain silly.

First, with out doing blind tests you never know if the reviewers bias is knowingly or unknowingly slipping in. Many people who are reviewing these discs have an agenda.

Secondly, how do we know that the reviewers in question even know what to look for or even know what a film is supposed to look like? Many of them seem to be enamored a grainless, highly processed HD video look.

Doug

 

[DaViD Boulet]

agreed. Can the mods change the topic to something more like "PQ differences: Codec or Mastering related"?

 

[Grant H]

Also different to compare the two, especially for BD people, since our VC-1 titles tend to be low bitrate WB titles. Some (2001) look awesome. Others (Troy DC) look a bit soft, with occasional artifacting in dark backgrounds. (Face it; there's a lot crammed on that disc.) And 2001 doesn't exactly have a lot of motion to deal with, does it? Probably easier to compress.

Deja Vu looked quite awesome, but it's VC-1 at a rare, high bitrate.

Some of the VC-1 titles I've seen look a lot more like broadcast HD to me than BD, but I'm guessing the softer, less detailed, maybe flatter appearance in these cases is more bitrate-related than codec. Especially, as the quality varies throughout the films. Gets tough to judge when you know sometimes it's the film itself's limitation.

It's my guess the codecs are about equal, though people used to say AVC actually had the potential to be more efficient, though the application of VC-1 had been more or less perfected, hence better.

If AVC can grow the way MPEG2 did for DVD, what a treat we have in store for us.

 

[Mark Zimmer]

The discussion at AVS regarding AVC was a mess, but here is a paper that talks about the AVC loop filter that's part of the codec process:

http://ee.uta.edu/Dip/Courses/EE5359...eyas_Desai.pdf


Specifically:

I'm no electrical engineer, but this seems to say that filtering is applied whenever there's an AVC encode.

 

[DaViD Boulet]

Though the "adaptive" part, coupled with the fact that AVC HD transfers happen to incorporate the most detailed HD signals on any consumer media, would imply that the result is less of a compromise then whatever is happening with many VC-1 transfers/masters. Does anyone have the skinny on VC-1?

Even highly-touted reference VC-1 encodes like King Kong aren't quite as detailed as reference AVC titles (Pirates and Ratatoulle). Again, no hard conclusions can be drwawn given different sources and mastering. However, when the best AVC titles do outshine the best VC-1, it would suggest that AVC isnt' being over-filtered by default.


Sounds like the "filtering" you're talking about above (mentioned in the paper a few paragraphs before this) is talking about a signal stored as a reference in a memory buffer to compare against with original image to make sure real detail isn't being lost. In other words, it sounds like its purpose is to preserve perceptible picture information, not remove it.

"Filter" is a word that can mean many different things in a technical discussion (ie: an "oversampling filter" in a CD player isn't a "fitler" in the sense that it removes information... it actually interpolates new information to round-out the edges of the 44.1 kHz waveform).

 

[Douglas Monce]

I would say that this is a HIGHLY debatable conclusion on your part.

Doug

 

[DaViD Boulet]

Not debatable on my part. But perhaps on yours?



Dan Ramer has also noticed the same thing on his 1080p front-projection rig. BTW, if you find any VC-1 encodes that look as good/better than the best AVC encodes to your eyes, let me know so I can make sure I test them on my system to see how it looks on my gear. Thanks!