AOD vs Blu-ray revisited...

No amount of "postprocessing at the decoder" is going to fix the problems stemming from dissolves, fade in/out, dark scenes approaching black level, and detail corruption in high motion scenes.

Actually post processing is a necessity anyways (as is pre-processing at the CCD/CMOS capture level). But all these problems that you complain about have been or are currently solvable and addressable. In fact, today, these are largely human decision problems (outside the limitations constrained on the storage medium).

FWIW, my impression is that the new techniques implemented in MPEG4 (tracking moving entities and such) are simply addressing additional ways to compress data. I am unaware of any refinements which specifically address video quality issues under demanding/extreme circumstances (specifically, conditions that give the current formats trouble).

That's because we're talking about compression, not filtering. You're talking about filtering. The moment you go digital, it immediately becomes a problem of how much data can you handle (upwardly limited by how much data you can sample). Sure it'd be nice if we could watch uncompressed sources from something like a Thomson Viper rig or something like that, but lets be reasonable. The amount of data generated simply makes it unfeasible, so invariably it becomes a problem of how much data can you toss without degrading the quality enough to be disruptive (which is a difficult boundary to define).

"Special conditions" that you bring up are special because they're difficult to compress. If it's difficult to compress then chances are it's going to become a data problem and the storage format is going to determine how much head-room you're going to have to deal with the problem. Unfortunately the featureset of a particular decoding device imposes another limit on the language you can use to describe your problematic data (or scene if you will).

Obviously you're going to have to define a feature set of techniques that a data decoder is going to be able to understand and stick with it if you want to set anything resembling a standard. What MPEG4 gives you (say for instance over MPEG2) is a much more verbose language to describe the data in question. Special conditions start becoming less special because you now have a wider range of options in attacking the data problem.

Well getting to the root of the problem is easier said than done wouldn't you say? I mean what alternatives are there?? Laserdisc Ver. 2?

That still wouldn't get to the root of the problem, that lies further upstream. To a point you can argue if there even "a problem". Fundamentally you're dealing with the process of reproduction. With reproduction comes interpretation... And that is truly the problem (that is if you believe we should all see things the same way).

The problem now is that the digital video paradigm has been a "race for more compression" where it should be a "race for better quality".

Actaully it's been more a race for tradeoffs, not compression... Compression is easy, deciding and managing tradeoff is hard.

One is only asking that digital be called upon to fulfill its full potential in being a truly superior successor to it, not just a get some/lose some/but it is newer proposition.

Well for the most part it *has* proven to be so. But there is no truly superior solution, since new formats always involve trade-offs... However the benefit has truly been greater than the drawbacks...

It's the marketing department which determines the final quality more than the codec developers, they dont determine what bitrates to use.
Actually most *marketting* people wouldn't understand bit-rate other than more == better quality spaek... Bit-rate is determined more on a practicality basis. Now if we're talking broadcast, then it's a bit different because available bandwidth is somewhat more variable.

I keep repeating this, but no one seems to get it. The problem isn't necessarily bit rates.

Yeah and it's you!

The problem isn't necessarily bit rates. You can throw as much bit rate as you like at an MPEG-based video. There will still be problems in certain basic situations. The one problem where you get pixelation under inadequate bitrates is only one of the problems. It's an easy fix, as well

No, but bit-rate gives you a basic window to work within the problem. However since encoders aren't defined (only the decoder), the quality can be variable from encoder to encoder...

The choice of MPEG2 vs. MPEG4 is irrelevant,

Ah, well this is finally something you *are* spot on about... (although since MPEG4 is a logical improvement/evolution over MPEG2, it does bring more benefits specifically from a feature headroom point of view)

As an industry spec, do you honestly think they will use the extra data headroom in MPEG4 to ease back on compression? Of course not! They will just compress stuff twice as much as MPEG2, and you end up with a result not that much better in quality than what you had before MPEG4.

Answer? YES!!! and NO!!! It all really depends on the situation... MPEG2 encoders the past 5 years *have* improved and their effect on DVD has been noticeable. Some of this has come from improved mastering processes as well, but DVDs mastered today have the benefit of encoders that give you more headroom in difficult scenes than they had even just 2 years ago.

As far as production goes, improvements in digital compression have allowed video cameras (e.g. DVCAM/DVCPro, IMX, HDCAM) to capture higher and higher quality scenes while still maintaining manageable volumes of data. And it's trickled down into consumer space (e.g. DV, MicroMV)...

OTOH, it can also be a tool in which producers use it to broadcast more, often tilting in the direction of more channels at the cost of quality...

It's a vicious cycle until people wake up and realize the "race for more compression" should not be the primary issue.

People aren't going to wake up, because there is no race for compression... Ever since somebody figured how to quantize an image it's been a problem of tradeoffs (arguably one of quality)...
 
Perhaps, but if a digital format needs to rely on such computationally extensive "postprocessing" to compensate for its inherent ills, that does not strike me as a worthy digital format to perpetuate.

Post (and pre) processing is simply a fact of life when you digitize something. Cameras have to perform limited pre-processing on scenes recorded by their CCD/CMOS (CMOS in particular) simply to reduce the noise captured and correct errors sampled by the sensor. With compression you're tossing data so a certain level of error is going to crop up(and possibly noise depending on how you quantize a picture down), post filtering simply helps the codec along (actually pre-filtering). In fact most codecs do a small amount of post processing anyways as the raw output would not be too pleasant.
 
You keep switching back and forth. You refer to "postprocessing", but then you give an example which is "preprocessing". When I was referring to postprocessing, I was referring to anything that happens after it has been encoded to some lossy MPEG format (essentially something that happens on the end user part of the chain). Naturally, there is all sorts of preprocessing that goes on to prepare a sample for encoding/compression.

I realize there is some confusion going on because we are referring to pre and post, but from different stages of the production process. The postprocessing for one stage would actually be the preprocessing of another stage. WRT everything I've brought up, it's all preprocessing if it is happening before data compression. Once it has been compressed, any data lost or artifacts that crop up become difficult to undo. That's the point I'm getting at.
 
WRT everything I've brought up, it's all preprocessing if it is happening before data compression. Once it has been compressed, any data lost or artifacts that crop up become difficult to undo. That's the point I'm getting at.

Because quite often (in video) you're already working with "compressed" data before it hit's the encoder (e.g. color space conversion, chroma sub-sampling), so there's already data loss before you've really hit the encoder...

In any case, the instance that you speak of (post compression), post processing isn't used to "reproduce" data. It's merely used to make the output more presentable (like I mentioned before, it's rarely presentable in it's unprocessed form)... Obviously the level of processing can (to downright alteration) can vary...
 
archie4oz, what do you think about M-JPEG quality-wise? Is it better than MPEG2 and MPEG4. Also do you think M-JPEG 2000 will be a superior solution in terms of quality compared to the MPEG4?
 
PC-Engine said:
archie4oz, what do you think about M-JPEG quality-wise? Is it better than MPEG2 and MPEG4. Also do you think M-JPEG 2000 will be a superior solution in terms of quality compared to the MPEG4?

PSOne used M-JPEG :)
 
Panajev2001a said:
PC-Engine said:
archie4oz, what do you think about M-JPEG quality-wise? Is it better than MPEG2 and MPEG4. Also do you think M-JPEG 2000 will be a superior solution in terms of quality compared to the MPEG4?

PSOne used M-JPEG :)

Yes I know and so did the PC-FX which came out before the Playstation ;)
 
Yes I know and so did the PC-FX which came out before the Playstation

Yes but ironically the Playstation was physically complete piece of hardware before the PC-FX was...

archie4oz, what do you think about M-JPEG quality-wise? Is it better than MPEG2 and MPEG4. Also do you think M-JPEG 2000 will be a superior solution in terms of quality compared to the MPEG4?

M-JPEG isn't really all that great as a distribution format. It is nice for NLE and offline video editing because it's frame accurate (no interframe dependencies), very resolution scalable (8x8 tiles are it's only real restriction), and it's cheap on processing. It's fine down to about 2Mbps for NTSC and PAL resolutions but that's about it.

Nowadays alot of the tasks that were/are handled by M-JPEG are now being picked up by I-frame MPEG implementations.

M-JPEG2000 OTOH is a bit of a different beast. It has all the benefits of M-JPEG, while delivering much better results at lower Q-values (and degrades more gracefully). Still, MPEG specifications give a lot of leway to the implementation of a codec and their applications have been far wider ranging. M-JPEG2000 might get a nice niche in offline editing and such, and perhaps video conferencing (low latency and no inter-frame dependencies make it very suitable in that purpose)...
 
Blu-Ray Group under investigation by feds...

Justice Department Probes DVD-Standards Group

Sun Jan 25,10:22 PM ET

In an increasingly contentious battle to create a global standard for the next generation of digital videodiscs, the U.S. Department of Justice (news - web sites) has started a preliminary inquiry into activities of one industry group, led by Sony Corp (NYSE:SNE - news) (news - web sites). (NYSE:SNE - News) , Matsushita Electrical Industrial Co. and Philips Electronics NV , that is vying for the lead, people familiar with the matter told The Wall Street Journal.

The inquiry is part of a battle involving powerful companies from the consumer-electronics, high-technology and movie industries, over a new format for DVDs. The companies are jockeying to establish a new DVD format capable of playing movies in crystal-clear high-definition video, which requires far more storage space than the standard movies on today's discs.

In a replay of past format battles -- notably the Betamax-VHS fight in the budding videocassette-recorder market of the 1970s -- two powerful camps are quarreling over the design of the successor to the DVD, which they hope will drive sales of a new wave of entertainment gadgets. Also at stake in the battle are potential royalties that could provide significant income to the companies behind the standard that ultimately prevails.

The Justice Department is looking into a coalition of companies known as the Blu-ray group, people familiar with the matter say. Founded by Sony, Matsushita, Philips and seven other major electronics companies, the group is promoting the Blu-ray disc, a format that can store about six times as much data as a conventional DVD, enough for more than four hours of high-definition video. A person close to the DVD Forum, an official standards-setting body for DVDs that includes hardware and software companies, said the Justice Department is looking into whether the group's members potentially acted in concert to impede the forum's technical progress.

A Justice Department spokeswoman declined to comment. A U.S. spokeswoman for Japan's Sony said the company wasn't aware of any inquiry by the Justice Department. "We know nothing about this," she said. A Matsushita spokesman in Japan also said he wasn't aware of such an inquiry. A representative for Philips didn't return calls.

Wall Street Journal Staff Reporters Nick Wingfield In San Francisco, John R. Wilke In Washington and Phred Dvorak In Tokyo contributed to this report.
 
Wow, that AOD group is pretty underhanded.

If blu-ray offers more storage space than AOD, how is it technically inferior? Maybe not supporting the new compression format is, but as a pc user and a gamer, I'd rather have blu-ray, as mpeg 4 compression does nothing for non movie/audio data. Though, as a movie watcher, I'd probably rather have AOD, since it could have longer movies and special features.........wait I don't care about special features 99% of the time, give me my increased storage space, use 2 disks if needed.
 
Fox5 said:
Though, as a movie watcher, I'd probably rather have AOD, since it could have longer movies and special features.........wait I don't care about special features 99% of the time, give me my increased storage space, use 2 disks if needed.

I don't think there will be a problem, see my and cthellis's post which not only confesses his undying love for Will Smith, but states the BD-ROM capabailities wrt this topic, which are pretty good.

Cthellis42 said:
It's a good thing they're looking into... uh... Just WTF are they looking into?

I just love how our tax's go towards investigations like this, gotta love big corperations playing hardball. :rolleyes:
 
Fox5 said:
So blu-ray is perfectly fine for any movies and special features at full quality?

Aslong as it fits, which I, personally, don't know of any that wouldn't, within these bounds - but I'm sure some movie buffs could:

  • Single Layer BD-ROM:
  • 135m of HD movie data, plus 3 languages of Dolby Digital, and 3hrs of bonus, SD video content
  • 135m of HD movie data, plus 3 languages of audio in Dolby Digital, DTS and PCM, and 70 minutes of bonus, SD content.

    Dual Layer BD-ROM
  • 3hr of HD movie data, plus 3 languages of Dolby Digital and 9 hours of bonus, SD video content
  • 3hr of HD movie data, 3 languages of audio in Dolby Digital, DTS and PCM, plus 2 hr of bonus, HD video content.
 
Vince:
I just love how our tax's go towards investigations like this, gotta love big corperations playing hardball.
So you know specifically which practices, and the legalities of such, are under inquiry from the Justice Department here?
 
...

So you know specifically which practices, and the legalities of such, are under inquiry from the Justice Department here?
The industry collusion to prevent the adaption of AOD as HD-DVD standard...
 
Yeah, I finally found that part. Took them a while to get around to it, though... Silly me, I thought the FIRST sentences were usually the ones used to encapsulate the whole situation. o_O

Regardless, I rather doubt anything will come OF this even if there's something TO it. (Anyone know the bylaws of the DVD steering committee offhand? ;) ) Netscape was killed long before anyone seriously looked at Microsoft's actions (and hey, look at the end results of THAT). The HiDef DVD format will go through its paces long before anyone can take a serious look at THAT, either. (If it even gets past this stage.)
 
Email floating around Sony is to *NOT* delete emails or destroy documents that even mention Blu-Ray. Mind you, nothing from the Justice Department has been officially served to Sony (yet).
 
Back
Top