battle of the CBR/VBR/CQ modes

[Martyn1980]

I encoded my first film sometime on March and i cant claim to be all knowledgable, but i use what knowledge i do have to the best of my abilities, when a film is encoding i usually print off as much as i can, and im really out of my league in this argument too, but i only state my opinion, and nothing of what i have said has been lies, i just wish i could post the results, but for me my way is superior to any other way because im encoding for one machine specifically, within its limitations, and setting 2300 as a Maximum is way below what it can handle, CQ might care about Quality rather than filesize but with a Maximum of only 2300 theres only so much quality you can have, the level of blocks i saw using Kwags new CQ template were unacceptable to me, even if the low motion stuff did look great, i dont see this post heading anywhere, and its pointless continuing it while people are seeing it as going on peoples sides, when did it become a contest.

[Martyn1980]

Just thought id say Triple X trailer, CCE 3 pass min 300 maximum 6000 average 1722, great quality far better than any of those two samples, final filesize for video 17.5mb, works fine in my DVD player, and Kwag you did say any settings, as long as it matched the filesize well i beat it by 3mb and the quality is much better, not a block in sight, i know its way out of standard but it plays in mine and all i had was a filesize to beat.

[injunpana]

hhhmmmmm....why is it, that a member with 37 posts...who has been on this board roughly 2 1/2 months feels the need to throw himself into this argument between two of the smartest guys on this board and kwag, the wily newcomer who is opening eyes and changing minds on this board...

@injunpana- you comments are uneeded and make no sense...I have been on this board longer than you...and even though I started this little X-pass vs. CQ thing, I bowed out to watch...and so should you...please

The smartest guys? I doubt it. It just means they are very vocal and passionate about what they believe in.

About your two smartest guys and the wily newcomer comment, frankly I've been more inclined to listen to someone who conducts a battery of tests day and night to back up what he's saying than the two so-called experts who may have done battery of tests long time ago but have stopped having found their so-called optimum settings or template.

Things have changed, and there are newer versions of encoders but I doubt if they are still conducting tests to once again look for that sweeter spot. They are resting on their laurels and doesn't realize they are stuck in a status quo.

Contrast that to that wily newcomer, kwag, who is not content with his last encode even if he achieved the highest quality he can get from a source. He is always trying to push the envelope, not content to accept the defaults or set standards in Tmpgenc. He changes the GOPS, the quantization matrix, the resolution, always trying to find a better template for smaller filesize but with same or better quality. He is the one breaking gounds here, or trying to, while your two experts are now content spouting textbook information, and their interpretation of it may not even be right.

Whether I'm a newcomer or not is immaterial. This is a board, if I feel I have something to say then I will say it. If you feel the opposite, that's your choice but don't impose it on me. About my comment being nonsense, that's your opinion, one other did not think so.

Frankly, I am content to just be quiet and read along but sometimes it just gets frustrating. Your smartest guys may not be as smart as you think because they keep on bringing up conditions that have been addressed before, can't they read? Look at those conditions that vitualis are asking for a fair trial. For god's sake, kwag just said he has done that some previous posts before,lets not go over that again!.

[Martyn1980]

Kwag is only pushing the envelope in one field and that is in Tmpeg and mostly Mpeg1, most DVD players will accept Mpeg2 these days and if your after a good one for XSVCD then you dont have to spend much at all, i have a DVD player for DVD's and one for my own copies, and the cheap one plays them better than the more expensive one.

Even if you keep within SVCD standards, and only increase the Max to 2520 you will still produce better quality than the Tmpeg CQ sample of the Triple X trailer, everyone who has CCE and Tmpeg try these tests and tell your results, it wont take long, even on 3 or 4 pass VBR, for Tmpeg use Kwags new 704x480 CQ_VBR template and for CCE use the same resolution but use Min300 Average 1800 Max2520 Bicubic Resize, Image Priority 17, no anti noise, and tell everyone the final filesize then which one you think looks better.

[Martyn1980]

Slight correction an average bitrate of 1722 not 1800, was just being lazy 1800 will make CCE clip 1.4mb larger

[Martyn1980]

Or maybe even 1700, im experimenting, i dont have a Mpeg2 bitrtae viewer, so im trying to estimate it 1722 makes it still about 700k too big.

[kwag]

Kwag is only pushing the envelope in one field and that is in Tmpeg and mostly Mpeg1, most DVD players will accept Mpeg2 these days and if your after a good one for XSVCD then you dont have to spend much at all, i have a DVD player for DVD's and one for my own copies, and the cheap one plays them better than the more expensive one.

Even if you keep within SVCD standards, and only increase the Max to 2520 you will still produce better quality than the Tmpeg CQ sample of the Triple X trailer, everyone who has CCE and Tmpeg try these tests and tell your results, it wont take long, even on 3 or 4 pass VBR, for Tmpeg use Kwags new 704x480 CQ_VBR template and for CCE use the same resolution but use Min300 Average 1800 Max2520 Bicubic Resize, Image Priority 17, no anti noise, and tell everyone the final filesize then which one you think looks better.

Martyn1980, dont make a fool of yourself.
You encoded with CCE at a MAX of 2,520 and 4,000 on another post, and of course you dont see blocks. I take it you have a memory retention problem? Because I did say to encode anyway you want, WITHIN THE CONSTRAINTS OF MIN 300, MAX 2,300, same resolution. Not above. But hey, you went ahead and encoded over that and then you say that CCE has less blocks. Don't make me laugh. Then you say the file size is smaller and it looks better! Give me a break. Where's your sample? So that EVERYONE here can see it. Please try not to fit the shoes of those that are "blabbercracking" without a single visible sample

-kwag

[injunpana]

...and I have been reading and encoding for over a year...who cares? You probably know more than me or I know more than you...it doesn't matter-the reason I said anything is that that guy's comments don't make any sense and just make me have to scroll down further to read this post (as this post is doing ). He adds nothing new, he just comes in on the side of kwag- as it stands now I am on a side, but do you see me weighing in with unnecessary comments about kwag or virtualis or adam? Nope.

Why not? Don't be shy, or be afraid - choose your pick. If you are on a side, why not stand up and be counted? Because you can't say anything new? Who can? Your champions, Vitualis and Adam? Don't think so, they have been saying the same old things since the first pages. Kwag? Same thing, he's been claiming the same old thing, only he has new samples to offer on the table.

You claim I added nothing new, you might be right but I posted to react to a statement made by vitualis - that is joining the discussion. You came in just to say that you do not like my posts, that's unnecessary. But hey, if it makes you happy, come again.

About my comment on kwag, and the other two guys, feel free to hit the scroll button again because 'm going to expound on it just to irritate you and get you to come back again.

I don't care what will turn out to be the better encoding method, I admit that I am more a fan of kwag rather than adam or vitualis. That's because I like his attitude. he has been the one who was always bending over backwards trying to accomodate the other parties demands. He's been the one doing the tests, samples, and posting them to support his allegation. What did the other side do? Nothing. They just kept on repeating what the textbooks say, much more, of which the interpretation is also subject for debate because the two sides can't agree on it.

A sample, "VBR is better because it allocates more intelligently the bitrates". This sentence alone will create another 10 page thread because the two sides can't agree on what it means.

I am also a fan of kwag because I liken him to one who stirs up the pot. Sometimes he hits the bullseye, sometimes he ends up in a blind alley, but because of him some things are happening. Sure he works mainly with Tempgenc but he does dabble with CCE for comparison purposes. But being described as nincompoop with CCE, I won't be surprised if he decides to put his attention to CCE later.

When one guy early on came and posted a method he discovered of putting a 120 minute movie in one cdr of acceptable quality, he was immediately set upon by everybody, including some moderators, and was literally vilified. Sweeping statements of "you are a liar, a complete 120 movie in one cdr is crap no matter how you do it" was said as gospel truth. Kwag came to the rescue and showed indeed that there's a way. Now, I don't think anyone who will come in with that declaration will receive the same bad treatment as yy22ffly had.

Before, everyone is content to do a standard vcd/svcd or just tinkering around with the bitrates. Now, the also the GOP, and matrices are fair game to some. I don't think Martin180 would have bothered to try the 704x480 resolution in CCE if kwag didn't keep on bandying his 704x480 template for comparison. And now that people will hear the good results from it, more people will be trying it.

Yes, I choose to listen to him more because he hasn't stopped trying, he's always the one who has something new to say. The other party, they think they already have found the answer and stopped right there. Not good, status quo is not a good environment for experimentations and discoveries.

Makes sense? lol
[/b]

[adam]

If you want to do a fair encode you need the highest quality source possible and you need to remove as many variables as possible. I really think my previous suggestions are valid in this regard

Now this statement is what's silly. If you take this argument, up to what point in the quality of the source shall it make it a fair test?

No its not silly. What's silly is running a test on mpeg2 encoding by taking an mpeg2 source then transcoding it to mpeg4, then transcoding it back to mpeg2. The point it not necessarily to have the highest possible source to test with, though this is ideal. The point is to NOT intentionally use a bad, noisy source and I already explained the reason for this. Don't blame me if you didn't understand it.

Maybe you forget that the the test is to see what mode reproduces closest to the original source. .

I swear I could just should just let my opposition do the arguing for me, they can back up my points better than I can. Don't you see that you have just reiterated my EXACT point? The source is very noisy and contains lots of artifacts. You are judging the encoder by how well it preserves the quality of the original source. The BETTER encoder will preserve MORE of the artifacts and noise!

The MPEG-2 CQ mode made with TMPEG is better than the CCE. It's just looks better and has less artifacts than the CCE encode.

See the dilema? Less artifacts should mean better quality but it doesnt if you have a crappy source to begin with. A good encoder does NOT remove artifacts, it PREVENTS them from occuring in the first place.

[kwag]

[quote]
[quote="adam"]

If you want to do a fair encode you need the highest quality source possible and you need to remove as many variables as possible. I really think my previous suggestions are valid in this regard

No Adam. You're completely silly now. You obviously need to freshen up on analog and digital technology, because you don't know the difference. You don't know the difference between analog noise and "embedded, what used to be noise" in a MPEG file. Analog noise would be processed by an encoder ONLY if the source is ANALOG at the time of encoding. The process of TMPEG and CCE encoding from a DivX is 100% DIGITAL. What you see in the DivX, are residuals of what used to be noise in some generations before. The DivX DOESN'T contain any noise, because it's an MPEG file. So both encoders will see a 100% digital image. There's NO analog noise. What looks like noise, is also part of the MPEG stream, and it's also digital. Now if that DivX looks pretty good in it's original form, it CAN be used to evaluate the encoders. Why? Because if you choose an extremely high quality source, the result are probably going to be the same with both encoders. So you're not proving anything with that. If the source is not so perfectly clean, as in the DivX, THEN you will see a difference. There are barely any motion artifacts in the DivX sample. Very little. So if this source is 100% digital, then why are there more artifacts with the CCE sample? Simple. Because the encoder is not so good in motion estimation as TMPEG. You're encoding a 100% DIGITAL source, without any type of ANALOG noise. And THAT is a good source for comparison.

-kwag

[kinneera]

God this is sad...

A high quality DVD source has been digitized just the same - it is no longer analog. So any validity to your argument regarding Divx would be far superceded by using a DVD as the source. Of course, a studio master would be the best source. Furthermore...

The DivX DOESN'T contain any noise, because it's an MPEG file. So both encoders will see a 100% digital image. There's NO analog noise.

This is absolutely ridiculous. Of course there's analog noise - it's just been digitized, and at an extremely low bitrate no less. But its still noise. Just because it's digital doesn't mean it isn't still more garbage for the MPEG1/MPEG2 decoder to deal with. Thus, it is a simple logical conclusion that the encoder that faithfully reproduces the most of that garbage is the one that is actually performing better, because the encoder doesn't know that a macroblock produced by the Divx encoder isn't supposed to be there - it's all image data to the encoder. If then, you can take a crappy Divx and make it look better - in the sense of reducing the artifacts that were introduced by the Divx encoding - with the second encoder, you have proven that the second encoder is poorer than one that manages to reproduce the Divx artifacts. Can you follow this concept? I doubt it... but this is precisely why no one ever uses Divx as a test case for encoder comparisons if they want it to be useful.

Also interesting...I'm kind of wondering if anyone else has noticed that it takes a GOP that is 4 times as long and a highly tweaked q-matrix for TMPGEnc to match or surpass the efficiency of CCE's encoding. Because that is really all that Kwag's tests have proven...

I would also suggest you investigate the concepts behind compression and data entropy before you start pulling yet more stories out of your ass to cover for when your logic fails, such as Divx being a better source for comparison than DVD...

[adam]

Oh give me a break Kwag. This is the perfect example of how you just make things up as you go along. You need to learn the definition of artifacts as applicable to mpeg encoding. There are a number of different artifacts which can be introduced during ANY DIGITAL encoding just from the inherant flaws in all mpeg encoders. These include gibbs, ringing, aliasing, blockiness, posterization, checkerboards, and color bleeding. Taking a near near flawless source such as dvd and encoding it to mpeg4 will almost certainly introduce a number of these artifacts. Once the file is encoded these artifacts are physically a part of the picture. With any further encoding the encoder has no way of knowing which parts of the picture are supposed to be there or not.

Most artifacts actually appear to the encoder as requiring a good amount bitrate because of the sharp contrasts between them and the background.
The encoder not only preserves these artifacts but it also wastes bitrate on them that could be used somewhere else. The BETTER encoder will preserve MORE of the artifacts.

Well this thread has been frustrating but I guess it has also been entertaining as well. I guess I lied before when I said I was done with this thread, but then again so did you. In any case I mean it now. Say what you want, make up any kind of ridiculous and random tests you want, I don't care anymore. If people are gullible enough to believe completely biased and self serving quality comparisons than they deserve to be mislead.

"The DivX DOESN'T contain any noise, because it's an MPEG file"

That's classic. I may not get the last word in but at least I get in the last laugh.

[kwag]

A high quality DVD source has been digitized just the same - it is no longer analog. So any validity to your argument regarding Divx would be far superceded by using a DVD as the source. Of course, a studio master would be the best source. Furthermore...

Isn't that what I said before?

The DivX DOESN'T contain any noise, because it's an MPEG file. So both encoders will see a 100% digital image. There's NO analog noise.

This is absolutely ridiculous. Of course there's analog noise - it's just been digitized, and at an extremely low bitrate no less. But its still noise. Just because it's digital doesn't mean it isn't still more garbage for the MPEG1/MPEG2 decoder to deal with. Thus, it is a simple logical conclusion that the encoder that faithfully reproduces the most of that garbage is the one that is actually performing better, because the encoder doesn't know that a macroblock produced by the Divx encoder isn't supposed to be there - it's all image data to the encoder. If then, you can take a crappy Divx and make it look better - in the sense of reducing the artifacts that were introduced by the Divx encoding - with the second encoder, you have proven that the second encoder is poorer than one that manages to reproduce the Divx artifacts. Can you follow this concept? I doubt it... but this is precisely why no one ever uses Divx as a test case for encoder comparisons if they want it to be useful.

What kind of a dingy are you? ( for lack of a better word ). Because I refuse to sink to your level. We are on a technical level here, at least some of us. I do think I'm loosing my time here trying to make you and Adam understand. But here it is, so that you and Adam can finally shut up, and stop misleading people with your ridiculous word twisting nonse phrases. Analog noise lives in a dynamic space domain. What you have in a DVD is "embedded" of what used to be noise ONCE. Analog noise is NOT static. It's DYNAMIC, meaning that at different time domains, there's different noise at different quantities, at different levels, and different "noise floors". On a DVD, if you go for example to say frame 1000, and you analyze the bit pattern "IT WILL ALWAYS BE THE SAME", because it's static data. There's no noise there.
Now you take an analog recording, say a VHS tape, and you "freeze" say frame 1000. Analyze the noise. "IT'S NOT STATIC". It changes by MANY factors. Media, environment, EMF, Eddie Currents, radio waves, want me to go on! I don't hold a First Class FCC license, Radar endorsement, and an EXTRA class ham radio license for reading comic books.

Also interesting...I'm kind of wondering if anyone else has noticed that it takes a GOP that is 4 times as long and a highly tweaked q-matrix for TMPGEnc to match or surpass the efficiency of CCE's encoding. Because that is really all that Kwag's tests have proven...

You're right there! You couldn't have made my point better for me
"surpass the efficiency of CCE's encoding"

I would also suggest you investigate the concepts behind compression and data entropy before you start pulling yet more stories out of your ass to cover for when your logic fails, such as Divx being a better source for comparison than DVD...

I would suggest you go and study the theory behing digital data systems ( PCM, TDM, etc. ) and then come back and make credible arguments.
The only ones pulling things out of their "behinds" are you and Adam.
My logic has been put into practice with all the samples and explanations I have provided, and your words and Adams are only whispers of incoherence and dead end stories which don't make sense to anyone.

[kwag]

"The DivX DOESN'T contain any noise, because it's an MPEG file"

That's classic. I may not get the last word in but at least I get in the last laugh.

Actually the laugh is mine, because you can't tell the difference between real analog noise and a digital picture
A DivX is a digital stream. There is no noise in a syncronous digital stream.
Go back to school!

-kwag

[Martyn1980]

Has this descended into squabbling and name calling now.

Kwag

Encode the AVI sample with CCE with the best parameters and variables you have. But the final file can't be larger than the sample I provided

Maybe im a bit too stupid to be joining this discussion but im pretty sure, but increasing the Max bitrate is using the best parameters and variable i have, you mention only final filesize, why wouldn't anyone want to increase the final filesize, you cant using CQ because your final size would be hugh, it would look great but it wouldn't be usable to get films on a certain amount of media, VBR is as you all say, restricted by the average bitrate, but theres nothing to stop it increasing, and decreasing when its needed, i changed from 4000max to 2520max to keep it within SVCD standards and see if its was still better quality, using the same Min Average and Maximum bitrates as CQ the results are so close, thats its personal taste, that chooses your favourite,but increasing the max makes a world of difference to the quality without increaing the filesize, and i know i cannot back up my tests, but will anyone back me up, i know im out of my league here, so is there anyone willing to encode that trailer using Kwags template exactly and then take the average bitrate and use that in CCE VBR with a Minimum300 Maximum2520, final filesize is the important thing here,or has everyone lost sight of that amongst the arguing, is anyone willing to do these tests and upload the samples for everyone to see, there is to much talking, and not enough proof, and Kwag is winning if there is anything like a winner, because he is the only one with concrete proof.

[Martyn1980]

All this has turned into, is a personal fight, and a "i Know more than you" battle, when none of this is the point, and one is going to give in , no matter what happens or what results are shown.

[offline]

Somebody cross post this to doom9's forum. Some genuine
experts hang out there. All this debate has shown is that software
mpeg encoding is a religion, not a science; - and that
everyguy wants to be head priest and defrock the opposition.

Having said this I can see why the moderators have stepped in.
Kwag has made some statements that need to be countered as
some might believe what he is saying is somehow scientifically
true. Yes, Kwag as spent time doing tests (similar to a test I
read on geocities a year back btw) but introduces so many
value judgements as to near anhiliate any of his claims - which
may or may not be true. At best Kwag's statements may be
persuasive.

[injunpana]

No its not silly. What's silly is running a test on mpeg2 encoding by taking an mpeg2 source then transcoding it to mpeg4, then transcoding it back to mpeg2. The point it not necessarily to have the highest possible source to test with, though this is ideal. The point is to NOT intentionally use a bad, noisy source and I already explained the reason for this. Don't blame me if you didn't understand it.

injunpana wrote:

Maybe you forget that the the test is to see what mode reproduces closest to the original source. .


I swear I could just should just let my opposition do the arguing for me, they can back up my points better than I can. Don't you see that you have just reiterated my EXACT point? The source is very noisy and contains lots of artifacts. You are judging the encoder by how well it preserves the quality of the original source. The BETTER encoder will preserve MORE of the artifacts and noise!

You may have to save your smug self-congratulatory remark. Of the encoder capable of reproducing the closest copy to the original being the best encoder, that we agree completely. We never had a difference of opinion in that, but that's not really the point of contention here, you know.

The contention is it shouldn't matter whether we intentionally (or not) use a source of high quality (or not) to make it a fair trial. Just the fact that the same source was used by both encoding modes makes it already fair. If supposing we used a good to mediocre source, why wouldn't it be fair? Would one be disadvantaged, or have the advantage?If one is indeed disadvantaged right away, right there and then, just from choosing a source, then I would say that that encoding method earns minus points from me right away.

What I said was for this particular test it may be better to use an ordinary source because it is easier to tell the difference of a poor quality video from another poorer quality video than a [b]high quality video from a higher quality video[b]. I'm afraid that's the simplest I could put it. If you still don't get it, I'm stumped.

[energy80s]

Probably the best thing ot do is to try out Kwag's templates in TMPG for yourselves. If you like the quality, then use them ... if you don't, then don't use them!

I personally have tried all 3 modes and wouldn't recommend VBR to anyone as I have never seen it produce great quality. If I want the best quality in MPEG 2 coding I use CBR. If I want to reduce the file size a bit to get a file onto fewer discs I use CQ. It's that simple.

[Martyn1980]

I agree with you on a few parts energy, if you like the templates use then if you dont then dont, and that CBR is the highest quality encoding mode, but its also a very wasteful mode, and CQ is only going to be a constant quality if the bitrate is allowed to go high enough, in Kwags templates i dont think it is, thats part of the reason im not totally happy with them, they underuse my DVD player, and if your not getting great quality from VBR then i imagine your only using Tmpegs VBR which is only equivelant to CCE's 1 pass, and the results i have gotten using Tmpegs 2 pass arent great, my personaly recommedations for the best quality are encode a 1 minute clip at 2,3,4,5,6,7,8 and 9000kbps a second, burn all the samples to a disc and stick it in your DVD player, you should then be able to work out your player maximum bitrate, thats the maximum you use in VBR from now on, with a minimum of 300, keep it low if you want to go high.

This is what ive been doing the last few days and its the best results ive ever gotten, if you want constant quality how can having a maximum bitrate of 2300kbps give you constant quality, if your player supports high bitrates why not use them, trust me you will see the difference, im still ranting i know that, but when im the newbie and theres aparent experts telling me i cant use high bitrates even though im matching and even beating the filesize and others are saying i cant use a higher maximum because it increases the filesize i start getting annoyed, the average determines the filesize, nothing else, some scenes look no better at 5000k a sec than they do at 500, thats why CBR is a waste, and limiting maximum bitrates is never going to ger high action stuff looking good enough, no matter how much you mess with other settings.

[mpb]

Look, folks, before you rip into Kwag, be very sure that you can produce better work than he has. I have used his templates and they have not only helped me produce some nice video files, but have expanded greatly my understanding of how quantization can be tweaked, and which refinements of the encoding settings deliver which consistent results. Until you're delivering that kind of useful material to the forum, it would be appropriate for you to "put up or shut up".

When people share their knowledge, everybody wins. When people thump the chest based on hypotheticals, it just wastes time, space, and effort. Cheers to Kwag for sharing some genuinely useful tools and material.

[Martyn1980]

Im not ripping into Kwag i use his templates for one of my DVD players, and i am impressed by his dedication and his knowledge i just dont like how he expects everyone to think as he does, and i get better results using my way than i do with his templates.

I have 2 DVD players as most people do these days, A Toshiba one and a Yukai one, the toshiba has brilliant picture and sound on DVD's and VCD's, It has brilliant video on SVCD's but the sound is out of sync, there are a couple of fixes but they mostly dont work, so Kwags templates are perfect for that.
The yukai one plays everything pretty average, Video and audio cant compare to the toshiba bt it plays Mpeg2 up to 6mbps as far as i know, might go higher, but i dont know and wouldn't go higher anyway, any resolution too and no problems at all with audio sync, but unfortunately it doesn't play Mpeg1 over 352x288 resolution, all i get is a green screen, so Mpeg2 VBR is perfect for that player.

I will continue to use both ways and wont try to brainwash everyone into what i think is best, i need high bitrates on the yukai as 1.5mb a sec looks better on the toshiba, than 2mb a sec on the yukai.

Night folks

[vitualis]

@ Martyn: no I wasn't referring to you with the "nonsense" statement -- it was a rhetorically play on kwag's post somewhere up this thread...

Obviously, I'm not going to change some people's mind here as the are dead set in their dogma. However, please consider the following:

• making "tests" are all good and well, but you can't interpret it if you do it badly. Furthermore, your observations have to make sense. I'm sorry, but kwag has wrapped up his test results with so many layers of crap it simply doesn't mean anything.
• there is obviously something wrong with your mentality when you start critcising people for demanding good methodology on your tests
• making a good comparison isn't that hard. If you want to compare two modes of ONE encoder, then just make sure all the other variables are the same. If you want to compare two encoders, make sure that all other variables are the same. Pretty logical stuff.
• You should always test by using a HIGH quality source. We are testing the quality of the encoder (that is the LACK of generation of encoding artifacts). We are NOT testing the ability of it to filter noise or the effect of filtering noise on the encoder, though this may be a worthwhile test. This is another variable that needs to be eliminated.
• name calling doesn't improve your argument

It has been mentioned up there somewhere that CQ is somehow different from other forms of encoding. It is as if belief in this magical premise bestows on it amazing abilities of improving quality. Obviously, this is not true.

CQ like all other modes of encoding is STILL ENCODING ONLY MPEG FRAMES. It is just another method of bitrate distribution.

Regards.

[Martyn1980]

Dont know if anyone is still watching this topic but i thought i'd add a question, I recently acquired Blade 2 Divx, beautiful Video quality best i seen, and even better sound 448k AC3 file, problem is doing it justice.

Ive tried both encoders both ways and the best results come from CCE VBR set at a high maximum, the reason for this is, theres a lot of CGI people exploding and even at 4000k CBR theres still a fair amount of blocks, but this isnt a complaint on Tmpeg its a complaint on CCE, when i compared the two files from both encoders the VBR CCE one looked a lot less blocky but a tad blurry, so i re encoded the first 10mins of the film on both encoders at CBR 6000kbps, Tmpeg seems to have a lot more artifacts, but any sort of text like subtitles is sharper and i mean a lot sharper, and everything looks a lot more real,on the CCE one the subtitles are blurred and the overall picture looks clearer until it moves faster than 2 people talking, is there something in CCE making this blur, i have anti noise off, and image quality set at 17, could it be that or is there another option ive missed, or is this just the encoders fault, id love the sharpness of the Tmpeg one with the smoothness of the CCE one is this possible or am i after the best of both worlds.

[adam]

Well I said I wouldnt post again but since this is unrelated to the actual argument...

Martyn I'd raise the image quality some, I personally think the default of 17 is too low. At this level the encoder doesnt allocate enough bitrate during low motion/quality scenes so text like subtitles isn't as sharp or detailed. Try using between 22-24. Also you should look into using a sharpen filter, this will definitely offset any blurriness in your image, though I'm not sure what would be causing that. A sharpen filter will definitely make subtitles much clearer. Try using "unfilter." It works as well as other sharpen filters but is very fast. Unfortunately any sharpen filter you use is going to slow down the encoding speed quite a bit but with cce it will still be much faster than TMPGenc. Oh and by the way, these are avisynth filters I'm talking about.

[Martyn1980]

Thanks Adam, i will try increasing the Image Priority, and give some of the filters a go.

As for the blurriness it not that bad, i only noticed it when i started making comparisons between CCE and Tmpeg, and these test were done using DVD and Divx as the source both at a CBR of 6000kbps.

thanks again

a reminder............

Musicnyman started with these ?`s

1) as far as i know, it's not worth using VBR if you going to use the max bitrate in VBR to 2520 for a SVCD anyway right? many times i have a 3 cd SVCD backup that I set to 2520kbs since 2 cds would cause a very low bitrate. in cases where the max bitrate is the avg bitrate in VBR, use CBR?

2) is CQ even worth experimenting with in SVCD's cds? does anyone use this for DVD->SVCD conversion? i personally don't like it since target file size is very unpredictable and 2 pass VBR looks the same.

1)The best results for minimum effort - use Tmpegenc 2passVBR(new) i.e. Easy to calculate the average bitrate to make optimum use of the desired media -CDR/DVD-X

2) If you have a little time on your hands give CQ a try, needs a lot of experimenting to get the resulting mpg to eficiently use the resulting media, but can give impressive results


Cheers

[banjazzer]

I would add that it shouldn't really take much experimenting to predict the filesize with CQ. So long as a reasonable-sized representative sample is chosen, it just takes a few test encodes, a bit of basic maths and the ability to use a pocket calculator. 8) Even with 3 or 4 test encodes, the time taken is half that for 2-pass VBR.

[bilbogod]

Well this stuff about predicting CQ filesizes is old hat. Anyone who uses DVD2SVCD latest versions will notice that it runs 4 test encodes automatically from differing areas of the movie. For the eight films that I have done so far through this method only 1 went over size and the others filled the 2 cds to capacity.

[banjazzer]

I shall certainly have a look at DVD2SVCD.