Good news for Sony : Digigami MPEG-2 Encoder destroy h.264

[Mr. Hanky]

This means you did not bother to know that detail before championing the article (since you are asking)? You don't think cbr vs. vbr would have been an important factor in forming a context on this article? Very sloppy work on your part, and your promoting of the article to us w/o putting on context with this detail is disingenious work upon us.

 

[scooby_dooby]

This means you did not bother to know that detail before championing the article (since you are asking)? You don't think cbr vs. vbr would have been an important factor in forming a context on this article? Very sloppy work.

lol, you're a funny guy.

That information is not disclosed einstein!

What it does show is you didn't even bother to READ the documnent before cheerleading for One and declaring it a 'very good point'. LMAO, how's that foot in your mouth taste?

For all you know it WAS a VBR test!

VBR could make a difference, it could not, it could favour mpeg2 more, or it could not. Without any studies or evidence to back up your assumptions both of your are grasping at straws at this point...

 

[Mr. Hanky]

I think you're just fixated with bit-rate.

No more than you being fixated on compression factor, evidently.

 

[Mr. Hanky]

lol, you're a funny guy.

That information is not disclosed einstein!

It's not disclosed, hence it becomes unimportant to find out? Can't find out? Any sane, honest person would deduce that this article is cannot really give an accurate conclusion to the efficacy of codec series and bitrates on a hd-disc medium, without this piece of data. Simple matter, really. Yet, you couldn't resist on using it for your argument because it happened to go in the direction you preferred. It's all very clear now! Y-o-u...a-r-e...e-x-p-o-s-e-d.

 

[Mmmkay]

Oh, and BTW, care to point out exactly where it says these tests are performed using a CBR? How do you know it's not VBR at 24mbps average?

DVHS Emulation implies that the MPEG-2 content was CBR. Also, DVHS is limited to 1080/60i and would be presented as such, in comparison to the H.264 1080/24p streams.

 

[scooby_dooby]

It's not disclosed, hence it becomes unimportant to find out? Can't find out? Any sane, honest person would deduce that this article is cannot really give an accurate conclusion to the efficacy of codec series and bitrates on a hd-disc medium, without this piece of data. Simple matter, really. Yet, you couldn't resist on using it for your argument because it happened to go in the direction you preferred. It's all very clear now! Y-o-u...a-r-e...e-x-p-o-s-e-d.

No. That study goes in the direction of something called the truth.

You don't want to believe it, so you're looking for any reason to ignore the results of that test. I guess because for some reason you prefer to stick with codecs that are a decade old. Blind love for Sony or what is it?

Whatever the reason, you have those blinders on REAL tight, so I won't try and persuade you. Mpeg2 is clearly the best compression technique ever created.

 

[TrungGap]

It's not disclosed, hence it becomes unimportant to find out? Can't find out? Any sane, honest person would deduce that this article is cannot really give an accurate conclusion to the efficacy of codec series and bitrates on a hd-disc medium, without this piece of data. Simple matter, really. Yet, you couldn't resist on using it for your argument because it happened to go in the direction you preferred. It's all very clear now! Y-o-u...a-r-e...e-x-p-o-s-e-d.

Do you know how automaker test seat belts? Is it unimportant for you to find out? Most of the time, you have to leave it to the experts to tell you. By your argument, any sane/honest person would not trust automaker/insurance/experts and not wear seat belts?

In this case, I don't see why scooby_dooby needs to become an expert just to explain his opinion on this matter. He's putting his faith in the experts to tell him that...if you have issue with expert/test, you should raise those issue instead of trying to attack scooby_dooby on his honesty.

 

[randycat99]

Uhhh, so are you saying vbr is irrelevant to how mpg2 will look at the tested bitrates? It seems like you are pretty well cornered, but you don't want to admit it. At the very least, you have to concede that the polled results are not fully representative of what to expect on a vbr encoded disc format, no? No doubt, mpg4 will perform better at lower bitrates, either way, but the projected bitrates for mpg2 in vbr mode should be quite sufficient to for optimal consumer hd playback, don't you think? If that is the case, then it is really irrelevant what mpg version that is used, unless you are still promoting a capacity limited format for hd playback such as standard DVD media. Is this what you are implying?

At any rate, you should just let it go for this topic discussion, as the gist of it all is that mpg2 just made a step closer to mpg4 quality, which makes your point even less relevant, altogether.

 

[randycat99]

In this case, I don't see why scooby_dooby needs to become an expert just to explain his opinion on this matter.

He is disregarding a specific detail which puts a substantial context on the test results with regard to how hd will be encoded on a disc format. This is not to say that the test results are invalid. They are just not directly correlating to how well vbr mpg2 will perform on a disc format. He doesn't need to be an expert, but he needs to be minimally conversant on the relevant test conditions to really determine if the study has a bearing on the argument he is posing. Short of that, he might as well choose studies by the title to prove his arguments. That's a might bit pointless, don't you think?

 

[TrungGap]

He is disregarding a specific detail which puts a substantial context on the test results with regard to how hd will be encoded on a disc format. This is not to say that the test results are invalid. They are just not directly correlating to how well vbr mpg2 will perform on a disc format. He doesn't need to be an expert, but he needs to be minimally conversant on the relevant test conditions to really determine if the study has a bearing on the argument he is posing. Short of that, he might as well choose studies by the title to prove his arguments. That's a might bit pointless, don't you think?

I didn't mean to imply that the test result are valid or not. However, I'm saying that Mr. Hanky should argue against the test result and bring up his case to show that vbr mpg2 can out do that test result. Beside, I believe S_B point was that for h.264 the HD sweat spot is 16mb. The test validated his claim. But instead, Mr.Hanky just didn't address the validity of the test as to 16mb is whether the sweat spot or not and whether 8mb is adequate for HD, instead he goes go off saying S_B is dishonest.

Maybe I shouldn't have spoken on this matter as it's totally off topic, but I rather see a debate on the issue than personal attack. I think Mr Hanky could have said, 16mb is not the sweat spot because...or you shouldn't trust the test because...Anyway, I'll leave it be.

 

[randycat99]

Well, that's just the point- the test isn't directly representative of the way vbr mpg2 will behave on disc formats. C'mon- you know the only reason dooby is remotely interested in mpg2/4 comparisons is not for codec performances, in general, but because of their use in upcoming hd disc formats. This has to be obvious to you, right? If it was the former, there wouldn't be a reason for him to make this line of inquiry in the console section in a topic that specifically implies Sony. The software section would be the correct place for that discussion, and that is where he should have started a fresh, new topic, if this interest was the case. ...but we DO know his intent is the same as it has always been- to discredit the validity and utility of an upcoming BR format, since it is being back by Sony, and is not being supported by MS. I don't think there is much ambiguity over this, do you?

 

[Magnum PI]

- I said 700mb DivX are near DVD quality (this is from my own experience)

Then I suggest you to either experience more 700 MB DiVX movies and/or DVDs, immediatly see an eye physician or fix your TV set.

There is definitively something wrong.

DivX is probably superior in it's HD compression than h.264 is

As long as you make such ridiculous claims, there is no point debating about video codecs with. Knowing a little bit what you are discussing wouldn't hurt.

 

[Gubbi]

On our website we have compressed material which supports this assertion. A recent example is a 400MB H.264 720p video blog that we recompressed to 172MB MPEG-1 VBR. In our testing, only highly saturated, brightly colored material (atypical of most content) is improved by H.264 - owing primarily to the use of 4:2:2 color. It amuses us that our MPEG-1 VBR encoder can also match and outperform H.264 on many progressive encoding tasks at HD frame sizes. MPEG-1 is 6 years older than MPEG-2 and even more widely adopted, reliable, proven and trustworthy."
Improves Widely Used MPEG-2 Rather Than Hopping H.264 Bandwagon

Without disclosing the resolution and the bitrate of the H.264 stream the above statement is completely worthless.

Anyone could easily make a H.264 stream with excessive amounts of I-frames which could then be recompressed to a much lower size using just about any codec.

Cheers

 

[Griffith]

DivX 6.0 is not even comparable to X264 or H264

http://compression.ru/video/codec_comparison/pdf/msu_mpeg_4_avc_h264_codec_comparison_2005_eng.pdf


I believe that there are two major codec, really suitable, H264 family and WMV-HD

H264 for High Cap. Disc (as HD-DVD)
WMV-HD for DVD9

I've downloaded some real wmv-hd videos from microsoft site, just to taste in practice how much it compress without loosing quality

the videos are eye candy 720p and those are the results:

Dust_to_Glory_720.wmv
242 MB - 3 min 19 sec

Amazing_Caves_720.wmv
67 MB - 1 min 27 sec

Amazon_720.wmv
81 MB - 1 min 42 sec

and so on
__________________________________

so 242 MB for 3 min and 19 sec

means
3,28 GB for 1 Hour
6,56 GB for 2 hours
8,7 GB for 2 hours and 40 mins

so an average film of
1h 30min ........ 2 h, have some space to additional content (other languages, 'a make of')

@20.500 kbps Video and Windows Media Audio 9 Pro @ 440 kbps, 48 kHz, 5.1 channel 16 bit 1-pass CBR

miles and miles and miles better than any 480p dvd film out of here


all this to say a simple thing:

for me strange non-standard disc format are boring
If prefer to take a WMV-HD Film on DVD9 and put it in my pc, see the film without others $$ to pay for a reader hd-dvd or bluray, then go to my girlfriend with the film, put it in his 360, see the film without problems with her, then maybe I give the film to my friend that have pc or 360, then I sell this, and buy another

in my thoughts there's no extra ( a lot ) of money to support some strange format that may not be standard in future (read this as "wasting lot of money")

if Bluray uses MPEG2, wtf, my thoughts become ever and ever strong.


edit: to correct a typo and to say that my perspective is of average film viewer not an expert, so I believe that the mass can think as I think, not as the 'video tech geek', sorry if I was wrong, this are only my poor 2 cents

 

[Shifty Geezer]

He is disregarding a specific detail which puts a substantial context on the test results with regard to how hd will be encoded on a disc format. This is not to say that the test results are invalid. They are just not directly correlating to how well vbr mpg2 will perform on a disc format.

But that information is moot, isn't it? Surely it's taken that the same techniques were used unless stated otherwise? In the case of the study, if VBR was used, we can assume it was used across the board. If CBR was used, we can likewise assume it was used across the board. Either way the results showed that 16 Mbps h.264 (CBR or VBR, whatever they used) looked better than 20 Mbps MPG2 (CBR or VBR, whatever they used).

The only way this comparison isn't valid is if they used VBR on the h.264 encoding and CBR on the MPG2, giving a higher bitrate for h.264 when needed. I'd be surprised if they did this without reporting their findings. They were happy giving out other settings (4:2:0, 1080p, 8x8 matrix).

Of course, it's not proving anything. Most people on this board are probably smart enough to know a single study isn't flawlessly absolute in it's findings. But it adds some weight to an argument and gives some validation to an opinion. This makes a wonderful change from all too often people giving opinions without supporting them with tangible figures or information! If the article is missing points, if the information is dubious, or if the conclusions drawn don't follow in logical progression from the information at hand, there's room for discussion. But if you are to refute a point with some experimental backing, you ought to argue on the same level.

The problem is he has accepted it as fact, hence we are obliged to accept it as fact

You're not obliged to accept anything. Scooby uses these numbers in support of his argument. If you disagree, explain why the numbers are not valid. Provide a link to a paper showing a different study that shows a different appreciation of different codecs and bitrates. Link to a website that list comparative differences between CBR and VBR codecs, or a blog entry that explains in the experiment referred to, VBR was used for h.264 but not MPG2. Saying only 'you can't draw conclusions as the information isn't complete' doesn't help a discussion (except where someone sticks to some information and ignores other and needs to be corrected). We all of us make decisions based on partial information, because none of us are omniscient! As long as they're not life threatening decisions or choices with equally severe repercussions when badly made, there's nothing wrong with being wrong in your opinions.

Scooby has found data supporting his arguments. It befalls those who disagree with him and want to present a counter-opinion to present their case, preferably supporting their views with more than just criticisms of the other argument. 'We are right because you prevented information and it's flawed' isn't as convincing as 'though you prevent this information, we present this and this, including these explanations and clarifications, that show areas your information were unclear on and which in light of this new information, shows opinions are probable more accurate.'

 

[bloodbob]

While i wouldn't say a 700 meg divx is as good as a dvd 5 gig dvd I prolly would say a 1.4 gig divx is prolly pretty close but remeber you have to make the divx and the DVD from the same source.

Most of the codec comparisons are done from a already compressed DVD which is frankly why they never look as good as the DVD.

Most of the Mpeg 4 AVC codecs ( h.264 ) are pretty young in developement aren't quite up to the full potential yet. However its pretty obivious they if they don't surpass Mpeg 4 ASP ( h.263 what divx uses ) they will soon. At low bit-rates there is simply no competition between either of the Mpeg 4 video codecs and Mpeg 2 frankly if you believe mpeg 2 is better your mad or blind. When you start to get to exteremly high bit-rates things might change but well you might wanna start looking at losses codecs then anyway.

Now, DivX is probably superior in it's HD compression than h.264 is, so I would imagine a movie running at ~10mbps DivX quality would be extremely good, and probably better than a MPG2 stream in the mid 20mbps's.

Now frankly this is a silly comment the less processor and time consuming simpler codec out preforming the more advanced codec developed in the same standard so pretty well reviewed by the same people dosen't quite make sense.

 

[Simon F]

this is great, now we have our format war cluttered up with a codec war. all of this is insignificant in the grand scheme of things. as long as the quality is there, the features are there, and all of it is accessable it doesn't matter.

It gets worse... no one's mentioned VC1 yet

 

[supervegeta]

But that information is moot, isn't it? Surely it's taken that the same techniques were used unless stated otherwise? In the case of the study, if VBR was used, we can assume it was used across the board. If CBR was used, we can likewise assume it was used across the board. Either way the results showed that 16 Mbps h.264 (CBR or VBR, whatever they used) looked better than 20 Mbps MPG2 (CBR or VBR, whatever they used).

The only way this comparison isn't valid is if they used VBR on the h.264 encoding and CBR on the MPG2, giving a higher bitrate for h.264 when needed. I'd be surprised if they did this without reporting their findings. They were happy giving out other settings (4:2:0, 1080p, 8x8 matrix).

Of course, it's not proving anything. Most people on this board are probably smart enough to know a single study isn't flawlessly absolute in it's findings. But it adds some weight to an argument and gives some validation to an opinion. This makes a wonderful change from all too often people giving opinions without supporting them with tangible figures or information! If the article is missing points, if the information is dubious, or if the conclusions drawn don't follow in logical progression from the information at hand, there's room for discussion. But if you are to refute a point with some experimental backing, you ought to argue on the same level.

You're not obliged to accept anything. Scooby uses these numbers in support of his argument. If you disagree, explain why the numbers are not valid. Provide a link to a paper showing a different study that shows a different appreciation of different codecs and bitrates. Link to a website that list comparative differences between CBR and VBR codecs, or a blog entry that explains in the experiment referred to, VBR was used for h.264 but not MPG2. Saying only 'you can't draw conclusions as the information isn't complete' doesn't help a discussion (except where someone sticks to some information and ignores other and needs to be corrected). We all of us make decisions based on partial information, because none of us are omniscient! As long as they're not life threatening decisions or choices with equally severe repercussions when badly made, there's nothing wrong with being wrong in your opinions.

Scooby has found data supporting his arguments. It befalls those who disagree with him and want to present a counter-opinion to present their case, preferably supporting their views with more than just criticisms of the other argument. 'We are right because you prevented information and it's flawed' isn't as convincing as 'though you prevent this information, we present this and this, including these explanations and clarifications, that show areas your information were unclear on and which in light of this new information, shows opinions are probable more accurate.'

Hi Shifty Geezer, i don't understand one thing.

Ins't the Digigami Mpeg2 encoder performance aganist the article scooby posted ?

The test say 16 Mbps h.264 look better than 20Mbps mpeg2, but basing on what digigami say, they can make mpeg2 perform better and have a 16mbps mpeg2 with the same quality or look better than the 16mbps h.264.

If a better encoder give a different result i think that no test can be really rapresentative.

 

[supervegeta]

You're not obliged to accept anything. Scooby uses these numbers in support of his argument. If you disagree, explain why the numbers are not valid. Provide a link to a paper showing a different study that shows a different appreciation of different codecs and bitrates. Link to a website that list comparative differences between CBR and VBR codecs, or a blog entry that explains in the experiment referred to, VBR was used for h.264 but not MPG2. Saying only 'you can't draw conclusions as the information isn't complete' doesn't help a discussion (except where someone sticks to some information and ignores other and needs to be corrected). We all of us make decisions based on partial information, because none of us are omniscient! As long as they're not life threatening decisions or choices with equally severe repercussions when badly made, there's nothing wrong with being wrong in your opinions.

Scooby has found data supporting his arguments. It befalls those who disagree with him and want to present a counter-opinion to present their case, preferably supporting their views with more than just criticisms of the other argument. 'We are right because you prevented information and it's flawed' isn't as convincing as 'though you prevent this information, we present this and this, including these explanations and clarifications, that show areas your information were unclear on and which in light of this new information, shows opinions are probable more accurate.'

Hi Shifty Geezer, i don't understand one thing.

Ins't the Digigami Mpeg2 encoder performance aganist the article scooby posted ?

The test say 16 Mbps h.264 look better than 20Mbps mpeg2, but basing on what digigami say, they can make mpeg2 perform better and have a 16mbps mpeg2 with the same quality or look better than the 16mbps h.264.

If a better encoder give a different result i think that no test can be really rapresentative.

 

[Shifty Geezer]

I have no idea! I don't know the ins and outs of their investigations. As this Digigami release is promotional material, it's validity is questionable. Regardless, this all has nothing really to do with consoles. BRD supports 54 Mbps which is a bitrate where difference between codecs is supposedly negligable. BRD is going to feature in PS3 whether people want it to or not. If people feel they're being screwed with HD media needing a whole new player where a DVD9 can support HD 1080p24 just fine with some codec or other, the place to talk about that is AV forums.