the Matrix

[kwag]

hey kwag..why don't u have some samples of SVCDs done by other ppl at vcdhelp on your website as a comparison..

let the ppl decide that way

Samples must be done from the same source. As I did.

I did have samples posted over a month ago. It was from the movie "The Mummy Returns". The fight at the beginning of the movie.

And both samples looked basically the same. Just like the last two I posted.

There was almost no visible quality difference ( unless you looked at the screen from 6 inches away ), but the file size in the MPEG-1 kvcd was more than a megabyte smaller than the SVCD.

Viewed on a HDTV at a normal 6 to 10 feet, you couldn't tell the difference. Just like the "The Matrix" samples posted before. If you burn them and watch them on a TV, you can't tell the difference.

Unless you use binoculars in your livingroom to watch tv, then you're really sick

kwag

[kwag]

the kvcd does look pretty good, but you've got to realise the next step in general public video is dvd-r. vcd is has pretty much been maxed out. svcd looks good, its just the hassle of having to have 2 or 3 cd's per movie is a pain.. once dvd-r rolls around, you'll be able to move your svcd's on to dvd-r.. hey ya may even be able to squeeze 2 or 3 movies on to one dvd-r, and have a nice little movie selection menu.. sorta like the old multigame nintendo cartridge's..

kwag, is it your dream that one day on the outside of an apex box it will read "dvd/vcd/svcd/mp3/kvcd" ? heh

No exactly .
It's my dream that all future generations of DVD players will play correctly 352x480(576) .

kwag

[Kdiddy]

You got to be kidding if you don't understand this one. This is plain math. Take for example a movie of 240 minutes. 4 hours. If you use a VBR calculator, like the one in FitCD, the average bit rate for a 4 hour movie is about 329Kbps at an audio rate of 128Kbps for a 80 minute CD-R. How do you think that movie is going to look? Like crap!

Dude chill out, no one is attacking you for starters. I understand the math completely, however your statement did convey exactly what you just stated. Which is why I asked what you meant, instead of assuming something. But is funny that you say it looks like crap, because that is EXACTLY what others have ben trying to tell you. That a 2-hr movie at 776 kbps looks like crap to one done at 1681 kbps.

I'm sorry, but the new version of TMPEG blew away CCE. I've tested 2-pass and 3-pass VBR in CCE and CQ in TMPEG has, for your information, better quality/fidelity/stability/motion estimation algorithms, etc.

Im sorry that you are sorry. But my test have proven otherwise. Which is why I fall back on when done correctly with CCE. You own quoted website even states otherwise. Or are we now to not believe the same source that you sought for validation? Again a video done at bitrate A/2 is not the same or can compare to a video done at bitrate A in terms of fidelity. If CQ was so vastly superior to that of VBR, then explain why DVD video is encoded using VBR, instead of CQ?

Oh, and before anyone says that still 480x480 is sharper than 352x480, technically speaking, yes. Visually speaking NO.

That is a matter of opinion. One man's trash is another man's treasure.

There is a huge difference between CQ and VBR. They are NOT the same, and I think some people still just don't understand this issue.

Again, not according to your own quoted source, to again I ask, are we suppose to disregard the information from that source?t I requote in case you missed it. "CCE SP's one remaining advantage, multi-pass VBR, is worth it. This mode gets you the quality of CQ mode, while still giving you complete control over the bit rate. If you must have the highest quality but can't tolerate unpredictable bit rates, this is the only encoder that will do"

[kwag]

You got to be kidding if you don't understand this one. This is plain math. Take for example a movie of 240 minutes. 4 hours. If you use a VBR calculator, like the one in FitCD, the average bit rate for a 4 hour movie is about 329Kbps at an audio rate of 128Kbps for a 80 minute CD-R. How do you think that movie is going to look? Like crap!

Dude chill out, no one is attacking you for starters. I understand the math completely, however your statement did convey exactly what you just stated. Which is why I asked what you meant, instead of assuming something. But is funny that you say it looks like crap, because that is EXACTLY what others have ben trying to tell you. That a 2-hr movie at 776 kbps looks like crap to one done at 1681 kbps.

I'm sorry, but the new version of TMPEG blew away CCE. I've tested 2-pass and 3-pass VBR in CCE and CQ in TMPEG has, for your information, better quality/fidelity/stability/motion estimation algorithms, etc.

Im sorry that you are sorry. But my test have proven otherwise. Which is why I fall back on when done correctly with CCE. You own quoted website even states otherwise. Or are we now to not believe the same source that you sought for validation? Again a video done at bitrate A/2 is not the same or can compare to a video done at bitrate A in terms of fidelity. If CQ was so vastly superior to that of VBR, then explain why DVD video is encoded using VBR, instead of CQ?

Oh, and before anyone says that still 480x480 is sharper than 352x480, technically speaking, yes. Visually speaking NO.

That is a matter of opinion. One man's trash is another man's treasure.

There is a huge difference between CQ and VBR. They are NOT the same, and I think some people still just don't understand this issue.

Again, not according to your own quoted source, to again I ask, are we suppose to disregard the information from that source?t I requote in case you missed it. "CCE SP's one remaining advantage, multi-pass VBR, is worth it. This mode gets you the quality of CQ mode, while still giving you complete control over the bit rate. If you must have the highest quality but can't tolerate unpredictable bit rates, this is the only encoder that will do"

"But is funny that you say it looks like crap, because that is EXACTLY what others have ben trying to tell you. That a 2-hr movie at 776 kbps looks like crap to one done at 1681 kbps."

You still dont get it. CQ is not the same as VBR. You get the "Illusion" that is VBR. In a way it is. but is a "CONSTANT QUANTIZATION"

I'm not even going to bother any more with you, trying to explain this. Just do a damn clip as CQ and another as VBR and look at the result with a bit rate viewer. Maybe if you can read the graph, you'll understand.
But apparently you're sticked with CCE.

"If CQ was so vastly superior to that of VBR, then explain why DVD video is encoded using VBR, instead of CQ?"

"MARKETING" Because they can fit exactly the ammount of video they want in one DVD. If they used CQ, they can't, and they would probably have to re-encode at another bit rate to fit.
It's not about quality, it's about money. For us, it's quality, even if we have to re-encode again!

"That is a matter of opinion. One man's trash is another man's treasure."

So are ex wifes


"CCE SP's one remaining advantage, multi-pass VBR, is worth it. This mode gets you the quality of CQ mode, while still giving you complete control over the bit rate."

I agree, but only if the movie is an average of less than 90 minutes, based on all the tests I've done.
Longer than 90 minutes, TMPEG's CQ curve takes over and at 120+ minutes CCE is crap and TMPEG shines.

Just try a 2 hour film like "The Matrix" with CCE's VBR to fit a 80 minute CD-R, and do the same with the kvcd template at CQ=74.

I did. The file sizes will be about the same. If I remember correctly around 815MB for both encoders.

But the quality difference is almost twice better with TMPEG's CQ than CCE's 3-pass VBR. This is no bologna. I tried it. If your movie is less than 90 minutes, then there's hardly any difference with either encoder.

kwag

[poopyhead]

yikes, it fails by a long shot....have you tried watching any of the kvcd samples on your computer? it's horrible.....maybe if you're watching on a small TV sitting far away (cuz doing that will also make normal VCDs look pretty good)....it may all be subjective, but take a good look at the backgrounds and even the people's faces when you're watching the samples on the computer...really, really bad

Of course it will look like crap on most computers. That's caused by the software mpeg decoder you have. And the close distance you're looking at the screen.
If you use WinDVD, it will look excelent besides PowerDVD.
And if you burn the sample, and view it in a real DVD player, then WinDVD will look like crap.

errr...1) watching on a computer still counts...2) comparing SVCD and kvcd on the same software (winDVD) makes a big difference...BOTH svcd and kvcd are subject to the software mpeg decoder

as for you fooling people about your kvcd being the actual DVD, play your kvcd rip side by side with the actual DVD and i'm sure ppl can tell...of course, if u just show them your kvcd w/o the actual DVD to compare it too, it's harder for people to tell

as for samples, as long as it comes from the same DVD title, two ppl can rip and encode in two different formats and still compare (doesn't have to be just 1 person encoding to both kvcd and svcd)

also, these samples don't really justify the quality cuz the encoder only has a short sample to allocate the bitrate..if you encode the entire matrix on 1 cd, the high action scenes will jack all the bitrate from the other scenes, making those scenes look like crap

finally, you have to get it into your head that lowering the resolution so you can get more bits/pixel doesn't mean the quality is gonna come out better cuz the resolution affects how clear the video will be...

it all boils down to the inadequate comparisons you make between kvcd and svcd samples, if someone watched a 3 disc SVCD sample (done correctly) compared to your 1 disc kvcd sample, there will be a big difference (it seems that some of the people who are using your template have had problems getting the quality they like w/ SVCD...so they don't have a good quality SVCD to compare kvcd to)

i.e....back in my pre-dvd ripping days, the only source of dvd rips i had was dling those highly compressed .avi smr rips....since those were my only source of dvd rips..i thought they were "awesome" cuz i had nothing else to compare it to....once i discovered svcd (again...correctly done), it totally blew away the compressed smr crap...hence, if you let all those "fans" of kvcd compare to the original DVD or a well done SVCD, i'm sure they won't be that big of a fan anymore

[Kdiddy]

You still dont get it. CQ is not the same as VBR. You get the "Illusion" that is VBR. In a way it is. but is a "CONSTANT QUANTIZATION"
I'm not even going to bother any more with you, trying to explain this. Just do a damn clip as CQ and another as VBR and look at the result with a bit rate viewer. Maybe if you can read the graph, you'll understand.
But apparently you're sticked with CCE.

This is always your response we people point out the fallacies in your statements. First, I dont need a bitrate viewer, I can look for myself to tell that a clip done at 1/2 the average bitrate level (regardless of quantization) will also look substandard that of its conterpart done at full bitrate level. And yes Im stuck on CCE just as much as you are on TMPG, because CCE does the job. I'll use the same statement that you used early in this post "Well you obviously are doing something wrong in your encoding, because that's not the feedback I get daily from the people that are using it exclusively."

"MARKETING" Because they can fit exactly the ammount of video they want in one DVD. If they used CQ, they can't, and they would probably have to re-encode at another bit rate to fit.
It's not about quality, it's about money. For us, it's quality, even if we have to re-encode again!

When I asked the question, I prayed you were going to come with something better than you did when I asked before. Sadly, you didn't. Yes the use multi-pass VBR becuase it not only provides the same quality as CQ, but you can accurrately predict filesizes. Marketing has no play in this, these companies do not have time to "dilly-dally" with the unpredictability of CQ. When the same can be accomplished with multi-pass VBR. If quality was no concern of theirs, they woud have stayed with VHS, where the tapes wear out and consumers would continually have to buy newer tapes, as oppose to the one-time purchase of DVDs.

Longer than 90 minutes, TMPEG's CQ curve takes over and at 120+ minutes CCE is crap and TMPEG shines.

You are losing sight of the debate. CQ for one disk would be better quality than multipass VBR in 1 disk. Im not arguing that point. I countering your claim that this stands true as well in 1 disk CQ vs. 2+ disk multi-pass VBR scenario, which is not true.

See this is where I have issue with your statments Kwag, you make statements, then go to quote sources agreeing with you. Only to find that sources always disclaim what you are trying to prove. You want to say that you have a template that does any movie very good on 1 disks , thats fine. It works in your scenario, a scenario which others may find themselves. No one will argue with "I do it this way because thats the way I like to." The problems come when you try justify what you are doing is being better than some other method with inaccurrate facts.

[jeex]

kwag, is it your dream that one day on the outside of an apex box it will read "dvd/vcd/svcd/mp3/kvcd" ? heh

thats ******* hilarious! heh

[kwag]

It's really sad to see people that are "tighter" than a nail driven in a cement wall.

All I can say is that TMPEG 2.54a is better than CCE. I've tried them both. Even the latest CCE.

If CCE was better, I would be using it. I dont get "married" with a brand.
I don't "stick" with something like a "fanatic".

As you guys with the "MPEG-2 better than MPEG-1", which I supplied facts and graphs from industry experts. But it was really a waste of time.

We'll keep doing a single kvcd with the complete movie and complete chapters.

And as for quality, well, lets end this with a big question:

If SVCD is so much higher quality than the samples we are discussing here, is SVCD really 3 or 4 time higher in quality than a kvcd sample?
I don't think so.
And in a computer monitor, at 5 inches or even one feet, you might see some artifacts ( just as an SVCD ) but maybe just a little bit more in the KVCD.

But when we watch TV, even on HDTV's as I do, the original DVD to a KVCD viewed from a logical distance of 10 feet, there's no difference.

But hey, you want three "bathroom breaks" because your whole movie is in 3 o 4 SVCD's? Good for you. Not for me. I want to be able to change or jump to different chapters in the movie, maybe after I've seen it once.

Can you do this with 3 or 4 SVCD. Absolutely NOT.

So you see, there is a big advantage to the single CD-R issue.

And as I said before, just to rub it in, if you increase the CQ to about 84, you'll blow away the SVCD.

But I guess it's worthless trying to repeat myself.

And remember. Only 20% of the DVD players out there play SVCD.
But 80% play XVCD ( and KVCD's included because that's an XVCD ).

So really an XVCD is more flexible than a SVCD.

kwag

[Kdiddy]

ahhh the famous Kwag "backpeddle & shuffle"

It's really sad to see people that are "tighter" than a nail driven in a cement wall.
All I can say is that TMPEG 2.54a is better than CCE. I've tried them both. Even the latest CCE.
If CCE was better, I would be using it. I dont get "married" with a brand.
I don't "stick" with something like a "fanatic".

And who is the one going around pushing off their template on others when that wasnt the initial question of the poster?? Who is the one the named a template after himself based upon an encoding application. Theis preverbial "pot calling the kettle black", you have no room suggesting anyone is fanatical about any idea. It is simply a matter of the inaccuracies of your statements.

is SVCD really 3 or 4 time higher in quality than a kvcd sample?

So now is this an admission to the fact that there is a quality gain?..lmao, yourself even said the gain is not linear is it not? So it is quite obvious that the answer is no. Is there a significant gain, yes. However, it is up to the "user" to determine whether or not that gain is worth the multiple disks. I personally feel it is.

But hey, you want three "bathroom breaks" because your whole movie is in 3 o 4 SVCD's? Good for you. Not for me. I want to be able to change or jump to different chapters in the movie, maybe after I've seen it once.
Can you do this with 3 or 4 SVCD. Absolutely NOT.
So you see, there is a big advantage to the single CD-R issue.

Again, stick the subject, never said this was not advantage. The trade-off to this advantage is a sacrificing of quality. Personally, no movie is worth 3 disks, if I need that much quality, then I'll buy the DVD.

And as I said before, just to rub it in, if you increase the CQ to about 84, you'll blow away the SVCD.

In a single disk KVCD vs. single disk SVCD scenario, I won't argue that.
In a single disk KVCD vs. multi disk SVCD scenario, simply not true.

But I guess it's worthless trying to repeat myself.

Yes it is, when some of the statements you make are untrue.

And remember. Only 20% of the DVD players out there play SVCD.But 80% play XVCD ( and KVCD's included because that's an XVCD ).

Thanks for the specifications on DVD players. Yet this has no bearing on the quality level of multi disk SVCD being higher than that of KVCD.

[kwag]

Ah Kdiddy, I guess I hit a nerve
Just what I wanted!.

You remind me of PacMan, Waca Waca Waca

As I said a long time ago. You must be a frustrated lawyer.
You excelent with words, the problem is they dont make any sense.
You just love to twist words around to your advantage. Great fun!

It's been very amusing, as usual, for me and everyone here.
Take your nightly pill, so that you don't blow up

Hasta la vista baby!,
kwag

[e-z-e]

this is way off the original topic...lol...i guess thats THE MATRIX for you.

[Kdiddy]

No you didn't hit a nerve, just laughing at how predictable you are, I find it quite funny and entertaining at best. "just what you wanted"..what to shift the topic off to teenage name calling to hide your false statements, lol, you tried that game already. Even Adam recognized your silly game by saying

It's not hard to recognize the point at which one's lack of evidence forces them to resort to petty name calling.

[adam]

That Farscape clip has no noise reduction used in it at all... Adam and Kineera, I could put a wager on it that that Film as seen on my 19inch philips 4:3 t.v carries no discernable difference when viewed from 10ft (ave viewing distance ??)as compared to the original dvd.

There are many other factors which can contribute to a soft image. The fact that the softening of your picture wasn't intentional doesnt make it any less detrimental to the overall quality of your film. Judging from the sample, the most logical assumption I could make was that noise reduction was used. Even though this is not the case I still feel that the image is way too soft. There just isn't nearly enough detail to that picture.

I would love to take that bet bilbogod. To be brutally honest I cannot understand how anyone could compare the quality of that farscape sample to a dvd. The sample looks very poor to me...thats not to say there is anything wrong with your encoding method. Like I said, it looks exactly as I would expect with a bitrate that low.

with CQ mode, you can't predict the final size but the quality will be better than an X-pass VBR"

I don't want to turn this thread into a CQ vs x-pass vbr debate but your conclusion is incorrect. CQ only results in more quality because the encoder is being so liberal with the bitrate. Compare a CQ encode and a multipass encode at the exact same filesize/avg bitrate and CQ will always lose, and isn't that the only test that actually counts for anything?If you take any CQ encode and determine the avg bitrate and then use that avg in x-pass vbr then the multipass vbr encode will ALWAYS be better...either it will be higher quality or it will be of equal quality and a smaller filesize. This is not a subjective issue it is an inherant fact. If comparable bitrate settings are used then CQ and multipass vbr are LITERALLY doing the exact same thing except that x-pass vbr has the benefit of scanning the video before encoding. I wish the creator of TMPGenc would post here, (does he even speak english?) it would clear up a lot of the misconceptions about TMPGEnc.

[poopyhead]

And remember. Only 20% of the DVD players out there play SVCD.
But 80% play XVCD ( and KVCD's included because that's an XVCD ).

So really an XVCD is more flexible than a SVCD.

kwag

errr...as mentioned by jeex in an earlier post

the kvcd does look pretty good, but you've got to realise the next step in general public video is dvd-r. vcd is has pretty much been maxed out. svcd looks good, its just the hassle of having to have 2 or 3 cd's per movie is a pain.. once dvd-r rolls around, you'll be able to move your svcd's on to dvd-r

don't think that arguement applies...as for SVCD compatibility, it's more than 20% of the dvd players that can play SVCD (i've even heard up to 50%)

[kinneera]

Can't resist just a few more comments here, since there's actually something to discuss again...

As you guys with the "MPEG-2 better than MPEG-1", which I supplied facts and graphs from industry experts. But it was really a waste of time.

Would these be the graphs that you completely misinterpreted? And would these be the same "industry experts" who think the compression limit of MPEG1 is 5Mbps and don't know the resolution of broadcast television? It was a waste of time because you're incapable of proper critical analysis (or an intelligent response when someone points it out - last post case in point)...

You are losing sight of the debate. CQ for one disk would be better quality than multipass VBR in 1 disk. Im not arguing that point

I don't actually disagree with anything Kdiddy has said here in spirit, but its worth pointing out that CQ could only match multipass VBR for 1-disk purposes at best. If the limiting factor is space, there is an absolute maximum number of bits that can be allocated. PERIOD. Multipass will always determine the best allocation. PERIOD. CQ will be superior if space is NOT the limiting factor, and you don't care how high your bitrate spikes can go.

You still dont get it. CQ is not the same as VBR.

Actually, it is. It's just that the rate-limiting factor is different. And inherently unpredicatable.

[adam]

Adam I still cant help feeling that your criteria for acceptability is above what would realistically interest the masses that frequent this forum. I quote

Your probably right about the lines not showing up on the protagonists face (I haven't checked the dvd source) but who the hell is gonna notice that sat 10 feet from the screen??? Common guys common sense please !!! I honestly believe that a passion for a hobby like this serves to distance you from the reality of what the lay person thinks. There's nothing wrong with wanting the best, but remember that something less than that does not equal crap.......

The softness of the picture was only ONE aspect of the video that reduced its quality. Would the average person feel the quality was lower because you can't see the lines in the faces of the people...no but what about the lines in every single surface in the movie? What about the reflectiveness of surfaces, the opaqueness of objects like smoke and water? Depth perception? Colors? All of these and more are lost with a soft image and yes I would guess that the average viewer would definitley notice this, especially when compared to a high quality source such as the source dvd.

The real quality issue here is the lack of bitrate and the infinite artifacts it creates and the loss of detail that results. Can you really look at that matrix sample at 660kbits and honestly say that you can sit through an entire movie that looks like that? If so then KVCD and 2+hrs on 1 cdr is for you and there is nothing wrong with that but I would not hesitate for a second to say that the vast majority of the people on this site do not share this opinion. If mpeg could achieve acceptable levels of quality at that resolution and at that bitrate then sometime in history there would have been a format made to utilize it. Obviously the mpeg consortion did not feel it viable to ever utilize bitrate at 500-600kbits at any resolution, and I wholeheartedly agree with them. If you disagree then by all means you and Kwag can start a revolution, just not HERE.

I feel that the low bitrate matrix sample, and any other similiar looking material, ARE utter crap but that is an entirely subjective opinion and I never claimed otherwise. I don't feel they are crap because they are less than the best or even because they are sub par in quality. I feel they are crap because I personally would never sit through a movie encoded like that and therefore they have as much value to me as crap would.

[stanwebber]

kwag, what is the upper limit of your template? you talk about the quality of 2hrs per disc as being absolutely amazing so what about situations where quality is not so critical. i have several 4-5hr dvds i want to backup with your template....how much quality would i have to sacrifice to get it to fit onto a single 80min cdr? what settings do i need to change to get acceptible quality? i am interested in preserving the original chapter points so multiple discs is simply out of the question

[bbb]

just like Palestine vs Israel

when will the brutality end ... when?

[poopyhead]

just like Palestine vs Israel

when will the brutality end ... when?

when someone drops the bigger bomb?

[bilbogod]

Well I'm still sticking to my guns. I have to strongly disagree with the notion of what "normal acceptable quality" means. One could easily argue the toss that mp3 forces an inferior format on to the masses as compared to the wave file, however I think you'll find that its discernable difference will mean that most simply do not care. I earlier advocated KVCD as being used by myself for a variety of purposes. This in no way inferred that I found it exclusively useful and I gave examples of where it shines as a format. Having frequently used DVD2SVCD -4pass I am fully aware of any visible gains that this could provide.
I will say again that it seems ever so easy for people to lose sight of what your average newbie could wish to attain, and to this end it is arrogant for any of us to pigeonhole there wishes. There is however overwhelming evidence(over 600 replies in the original post) that many have been pleased (as compared to other formats) with the choice that Kwag has offered to them. This includes ALL of my friends who acussed me of being anal over needing 3cds when shown 'the gains'. that they could expect to receive. Newbies are NOT cultural dupes that will be influenced either by what Adam or Kwag says, rather they can make up their own mind having been offered the choices.

[adam]

Which is EXACTLY why I asked that kwag present a sample at ~600kbits because THAT is the representative quality that you will achieve when squeezing 2+hrs onto a single cdr. Yes the samples at 1500kbits looked pretty good, but they didn't prove anything.

You never answered my question...Would you be able to watch an entire movie that looked like that low bitrate Matrix sample? If so then that is fine but if you think it is arrogant to assume that the average person on this board would find that unacceptable then YOU are the one being arrogant, and closed minded as well.

Your mp3 and wav comparison is ludicrous. Mp3 is a highly lossless codec, mpeg is exactly the opposite. Many audio experts cannot tell the difference between mp3 and wav. I don't think there is a single person in the world that would have any problem discerning differences between video encoded below 1MBper sec compared to an uncompressed avi created from a flawless video source.

The purpose of (x)vcds, (x)svcds, dvds etc it NOT to fit as much content onto a disk as possible or to see how far you can push the format in one particular direction ex: compressoin. The point is to create something that you can sit down and watch. I do not feel the average person would want to sit down and watch video encoded at ~500kbits, if you disagree than I really think you have a pretty warped view of society.

If you want to use this format great, if you want to promote this format even better. But do not make false claims that will mislead people. Regardless of what template you use, if you squeeze 2+hrs onto one disk than you will NOT have quality as good as the source dvd, something which you claimed about your Farscape dvd. You will NOT have quality equal or better than svcd, something that kwag claimed. In fact you will NOT even have quality comparable to vcd. This format is only viable if compression is of the utmost concern and you are willing to sacrifice a great deal of quality. Some people will have this agenda, but it is idiotic to assume that the majority of people will.

[baker]

This has turned ionto another of kwahs fouroums xo I didn't read the wholr lot as I aint got the time.

FOR CCE USERS:
Try setting the complex flat settings down to 25 as 40 is way too high.

Baker

[energy80s]

Your mp3 and wav comparison is ludicrous. Mp3 is a highly lossless codec, mpeg is exactly the opposite. Many audio experts cannot tell the difference between mp3 and wav.

Adam, that is garbage!

MP3 audio is piss poor below 192kbps. I have worked with MP3 (and indeed MP2) audio for years and I know just how shite 128k MP3's really are!

[adam]

So raise the bitrate then. What exactly is your point? Mpeg Audio layer III compression is not even remotely as lossy as mpeg video compression. Is this really that hard to understand? Riase the mp3 bitrate up to 256kbits and it will be very close in quality to an uncompressed wav file but it will be significantly smaller in filesize. Can the same be said for mpeg1 @ 500kbits vs uncompressed avi? Will the quality difference between that mp3 and the wav be even remotely proportional to the quality difference between the mpg and the avi? HARDLY!

You cannot compare mp3 to uncompressed wav in the same way you compare mpeg1 to uncompressed avi. Its comparing apples to oranges.

[energy80s]

Adam, my point is that there is a tremendous difference between raw audio/video formats and MPEG versions, to say otherwise is nonsense. At 128k, 89% of the audio is lost. THAT is noticeable!

[adam]

And I never said otherwise. MY point is that you cannot compare a highly lossless codec to a highly lossy one.

[bilbogod]

Well actually I did not claim that the kvcd quality of Farscape is as good as dvd. I questioned your claim that the discernable difference is so obvious and unacceptable as a result. I think it isn't, you think it is. But let me ask this. Do you honestly think that all of those (and there are many as you can see) who have tried the format and are happy with it have never tried the standard templates in TMPGENC for vcd and svcd ??? These are extremely easy to use and I simply refuse to believe that people use KVCD for lack of knowledge of other "superior" outcomes. There is nothing arrogant about my assumptions, on the contrary I advocate taking an "Average Joe" appraisal and not a scientific one. In fact it would be in all our interests if the original poster tried out the various formats and provided us with relevant feedback. Any openmindedness will result in greater learning, which for all of shiZZZons faults was actually one of his virtues. He took nothing at face value and would check it for himself. Other posters such as Vhelp have demonstrated an enthusiasm to increase all of our learning curves by discovery, rigorous testing and a refusal to simply accept the status quo.
As for the question about the Matrix film. My honest answer is I dont know. I haven't encoded this onto 1 cd and hence haven't seen what it looked like in its entireity. I have encoded "Moby -Play", Massive Attack, Farscape, Lolita (black and white), The Shining and Dr Strangleglove (black and white). The outcomes were acceptable (and that means good) for both myself and others that viewed them. The only film that I didn't like the look of was The Grinch, although that was noteably my first attempt and I chose the wrong resolution.
Let people try out these methods and tell us what they think, cause in my opinion there is a vocal minority here that are pissing in the wind..... Peace

[kwag]

bilbogod:

Check your mail. There's something cooking in the oven for later today

kwag

[MITSUI_1]

Wow...

Lots of "my dick is bigger" shit going on in this thread. Also lots of knee-jerk but uninformed, IMHO, criticism too.

Point, the first: You guys who are making generalizations on the "effective average" bitrate of a CQ encode using Kwag's approach are a little off the mark.

Quality in terms of viewability is a funciton of quantization, not of bitrate. Quantization is the process, by the encoding software, of throwing information away in order to meet the constraints set by the user. If an encoding process can get the number of bits way down, but ALSO has an acceptable average quantization on a macroblock by macroblock basis, then that reduction in bitrate did not cost anything in terms of viewability.

What many here don't seem to grasp is the importance of the long GOPs , and the structure of that GOP sequence, in Kwag's template. One of the common tools (I think it might be BBMPEG) has a "verbose" mode that lets you watch the actual file space taken by each frame of a video, and tells you whether the frame in question is an I, P or B frame. I urge you to take a look some time. From a random MPEG faq I've got on my hard drive from somewhere:

MPEG video encoding consists of three components: I-frames, P-frames, and B-frames. During MPEG encoding, some video frames are compressed to I-frames, some to P-frames, and some to B-frames. I-frame compression can yield up to 6:1 compression before noticeable artifacts appear. Using P-frames along with the I-frames, results in further compression without noticeable artifacts. However, it is the B-frames that enable compression ratios up to 200:1. B-frames are typically 15 percent of the size of I-frames and less than half the size of P-frames. I-frames remove spatial redundancy in the video data and the P and B-frames remove temporal redundancy in the video data, as explained below.

Now, based on those numbers we can estimate that a clip in which the average I frame at a particular quantization level equals 100k will result in P frames of 30k and B frames of 15k. This ratio is 100:30:15 or, essentially, 6.666:2:1

Ok. Take a look at the regular DVD GOP for NTSC film footage, which is

IBBPBBPBBPBB

for a 12 frame GOP.

With Kwag's template GOP, on the other hand (at least his most recent version with the 1-12-3-1 settings in TMPGENC), looks like this:

IBBBPBBBPBBBPBBBPBBBPBBBPBBBPBBBPBBBPBBBPBBBPBBBPB BB

for a 52 frame GOP.

Now, just for comparisons sake, lets do some math on the two gops. I'll use 48 frames worth of each GOP structure, which gives us two seconds of video at 24 fps NTSC film rate.

Regular
IBBPBBPBBPBBIBBPBBPBBPBBIBBPBBPBBPBBIBBPBBPBBPBB
IBBBPBBBPBBBPBBBPBBBPBBBPBBBPBBBPBBBPBBBPBBBPBBB
Kwag

Now lets do a comparison. We'll use the frame size ratio above, but here lets be a little closer to real world and suppose that in our clip (at a specific constant _acceptable_ quantization level(i.e., no visable mpeg artifacts) the average I frame at some is 200k, the average P frame is 60k and the B frames are 30k in size. (this meets the 6.6666:2:1 ratio).

Regular GOP:
Regular I frames = 4 = 4x200 => 800k
Regular P frames = 12 = 12x60 = > 720k
Regular B frames = 32 = 32x30 = > 960k
=================================
TOTAL K for 2 secs of reg GOP video = 2480k total for 48 frames (i.e., 1240kb/s)

KVCD GOP:
KVCD I frames = 1 = 1x200 => 200k
KVCD P frames = 11 = 11x60 => 660k
KVCD B frames = 36 = 36x30 => 1080k
====================================
TOTAL K for 2 secs of KVCD GOP video = 1940k total for 48 frames (i.e., 970 kb/s).


Ah ha...

So the KVCD GOP structure takes less bits for a given constant quantization. 22% less in our example.

Let me repeat. Assuming that all of the frames are encoded by the same encoder at the same quantization, they should be exactly the same quality. That means you should not notice a visual difference between the two sets of frames.

This is, of course, subject to I and P frame redundancy problems in the encoder, which is a topic I address now. In the past, most MPEG encoders had problems accurately preparing the P and B frames without making them ugly. More to the point, the frames just after the I frame in a sequence were most accurate, but then the accuracy degraded. This was becuase of difficulties in motion prediciton and in frame differencing, both of which are necessary for the P and B frames to look right. When these problems exist, you get ghosting and fuzzy areas that "pulse." Why do they pulse? Well, the picture is basically degrading as the errors build and build on themselves through subsequent P and B predictions until finally the encoder sticks in another I frame which, of course, is freestanding and not based on any predictive calculations and is much more accurate with respect to the source. Kwag argues that TMPG has evolved to where it is so good at preparing accurate and visually appealing P and B frames that you can go a really long time without using an I frame. Hence his long GOP.

His view is probably true based on the results I've seen, and it is not out of the ordinary from the current state of the art in MPEG encoding (those of you who have tried to convert a TIVO mpeg stream to DVD know that they use a really long GOP as well. )

Finally, a word on why the temporal smoother helps so much here. Again, its all a question of how mpeg works. Here, again, Kwag seems to be exploiting the P and B frames. As you know, if two pixels in a frame and then its subsequent frame are identical, that pixel will be considered a "static" area and will not be encoded in the second frame (assuming that second frame is a P or B frame). Well, what the temporal smoother does is look at two pixels in frames A and B and at the same x,y coordinates and see whether they are so close to each other in color and intensity (i.e, the 1,1 settings used by most of us on that filter) that the human eye cannot probably tell they are different. If that is the case, the filter sets the pixels to the exactly same settings, so that the pixels actually are duplicates of each other frame to frame. If you don't do this, and the pixels both close enough that the eye can't tell the difference, but "technically" not identical, then the encoder will have to spend bits encoding both pixels separately. Ergo, the result of the Temporalsmoother filter is to make the P and B frames even smaller.

Finally, two comments, one to the Kwag camp and one to the critics.

To Kwag and his supporters, quit getting so pissy about the criticism. If YOU are happy watching the discs YOU make in YOUR homes with YOUR friends, then who gives a shit if someone thinks there is a better way, or that you cannot possibly be getting the kind of results you are seeing. Just understand that however well those work for you, some us will have incompatible DVD players or will notice some flaw or another that may seem more apparent in the one-disc encodes.

To the critics: If your goal is to be able to see so much detail on your discs that you can actually perceive the wrinkles on the cocks of the male actors wearing tight pants, or perhaps the acne on the teenage actors, even during high action scenes, then by all means, consider sticking with DVD or DVDr, as even SVCDs will not likely placate you. Also, if you like switching discs, that's great too, but there are others who want to keep that to a minimum provided they have a quality of encode THEY can live with. Most of us just like movies and want to have WATCHABLE movies on backup media so we don't have to fork out another $20 every time we scratch a disc. A are recent VHS converts for whom even lower bitrate SVCD or xVcd represent a major improvement over tape. Some of us just like to take inexpensive CDRS with us when we travel, and in that context in particular, the one-disc solutions seem better.

Folks we're all enthusiasts. To each his own, etc., etc. Lets quit with the pissing contests.

[adam]

MITSUI_1 have you looked at that low bitrate matrix example? I don't care what the quantization benefits are of CQ encoding, whatever CQ value kwag used to get that clip to that size then THAT is the quality that is being kept constant thoughout the movie. That is the quantization level and THAT is the quality you can expect when putting a movie like the Matrix on 1 cdr. I don't want to see every detail in the picture, I want an acceptable level of quality. The quality in this clip is VERY poor and I would guess that the average non-critical viewer would agree. Can you honestly tell me that you have no problem watching a movie at that quality?

There is nothing wrong with Kwag's template, I use similar resolution and gop settings all the time. The problem I have with it is squeezing 2+hrs onto one cdr. I do not feel this can be done in an acceptable level of quality. If anyone disagrees than FINE but please just get over yourselves. Use your own settings for your own encodes but stop trying to convert others to KVCD supporters. Tell people what the template does and give them the choice to use it, don't suggest that it is some holy grail because it is far from it.

The original poster asked how to predict the filesize of his encode. How does telling him to put the entire movie on 1 disk help him? In my opinion this advice is a disservice.