Whats the BEST way to create 720x480 XSVCDs ??

[Zaun]

Im having some trouble getting DVD2SVCD to let me create some high resolution SVCDs from DVD source. I want to keep the DVD resolution, and pump the bitrate to 4500 since my player can handle it. Whats the best way to do this? any help is MUCH appriciated!

[dc91gt]

Just do it manually in tmpgenc.

Load up the unlock.mcf and set what parameters you want (res/bit rate).

-d

[markums2k]

Unless you prefer 10 minutes of video per CD, you may also want to check out www.kvcd.net. There are some templates for TMPGEnc there that produce video that is darn near close to looking like DVD. I think they have a few templates into the 704x480 range that still allow over 30 minutes of footage per CD.

[crahak]

KWAG's templates is probably your best bet right now. But keeping full resolution, and high bitrates will REALLY limit on how much you can fit on a disc. 15 minutes per disc is not what I'm aiming for, y'know?

[kwag]

@Zaun,

Download the KVCDx2 ( 704x480 ) PLUS template and change the MAX bit rate to 3,500Kbps and motion estimation to "Fast". No need to go to a higher bit rate. With this changes, your video will look like original.
Use the file prediction formula to target your movie for two CD's, if it's longer than 60 minutes: http://www.kvcd.net/forum/viewtopic.php?t=1294

-kwag

[Zaun]

Ok, Thanx, I'll give that a shot. I wish there were a way to do it with DVD2SVCD, ive used that all along, and I love 5pass quality.

I will try this, we will see.

Thanx

[snowmoon]

If you are using CCE ( the 5 pass comment make me think you do ) you might be better off using one pass VBR.

I have found setting the image complexity to ~5 and then setting the Q=factor I can adjust the bitrate fairly easly and produce excellent results all the way down to 1.5mbps for 720x480 ( digital calbe full screen IVTC Q=30 not that detailed and low action ). For the most part my cap's are running 2.5-3.5 for 23.976 progressive materal at full resolution. I started using One Pass VBR when I was perfecting SVCD encoding, but now have moved on to DVD's with a higher bitrate. One pass vbr is also much more friendly for long GOP's like 15 ( I can get CCE up to 21 ).

[kwag]

Ok, Thanx, I'll give that a shot. I wish there were a way to do it with DVD2SVCD, ive used that all along, and I love 5pass quality.

I will try this, we will see.

Thanx

You can use DVD2SVCD. Select the option to edit the .avs script before the video starts to encode, and set your resolution to 704x480. Just be sure to get the latest version of DVD2SVCD "1.1.0 Build 1". It already integrates the file prediction formula and the KVCD Quantization Matrix. Be sure to select that too under the "Matrices" tab. Not sure as to how it will work with CCE, because it was heavily tested with TMPEG. Not with CCE. Give it a try with TMPEG and make sure you use CQ_VBR. Not CQ. Also, change the GOP to 1, 18, 2,1, 18.

Regards,
-kwag

[wulf109]

Rip the DVD as one large file. Encode the vob directly in Tmpeg. Use 2-passVBR mode. Set the resolution to 720x480,don't use 704x480 it just slows down the encodeing. Change the advanced and video tab in settings to non-interlaced. I find this method produces near DVD quality. How good depends on the average bitrate you set.

[kwag]

Encode the vob directly in Tmpeg.

And how do you plan to select your audio language with this method? You're lucky if track one is english. Unless you plan to encode audio separately. But then you can't use any advanced filters. Like AviSynth resizing, TemporalSmoother, Blockbuster, etc.

Set the resolution to 720x480,don't use 704x480 it just slows down the encodeing.

The opposite. Lower resolutions, faster encoding.

-kwag

[snowmoon]

Set the resolution to 720x480,don't use 704x480 it just slows down the encodeing.

The opposite. Lower resolutions, faster encoding.

-kwag

The resize speed slows it down more than the small reduction in image size.

Also resizing from 720->704 is a bad idea since it does nasty things to the nice high fequency data.

[kwag]

Set the resolution to 720x480,don't use 704x480 it just slows down the encodeing.

The opposite. Lower resolutions, faster encoding.

-kwag

The resize speed slows it down more than the small reduction in image size.

Also resizing from 720->704 is a bad idea since it does nasty things to the nice high fequency data.

Really? Try encoding a sample at 740 or 720x480 and try a 352x240. Your 352x240 encode will be about more than three times faster.
As for 704x480, tell that to shh, developer of FitCD, and see what he's going to tell you. That's the second resolution on FitCD's choices. It's also the mpeg still standard resolution.

-kwag

[wulf109]

If you stream rip the DVD you can select the language you want so that's not a problem.
720x480 is DVD's native resolution. I have done many encodes at 720 v/s 704 and 704 add's about 20% to the encodeing time. The other poster was correct 704 requires resize,720 does not. The small difference in pixel count is why 704 takes longer than 720. 352x240 or 352x480 has a much lower pixel count so the comparison is not valid.

[kwag]

If you stream rip the DVD you can select the language you want so that's not a problem.
720x480 is DVD's native resolution. I have done many encodes at 720 v/s 704 and 704 add's about 20% to the encodeing time. The other poster was correct 704 requires resize,720 does not. The small difference in pixel count is why 704 takes longer than 720. 352x240 or 352x480 has a much lower pixel count so the comparison is not valid.

And the time to resize from 720x480 to 704x480 is negligible, compared to the time to encode 720x480=345,600 pixels and 704x480=337,920 pixels. So you're encoding 345,600 - 337,920 = 7,680 pixels. Not that many, but it does make a difference.
What do you mean by "352x240 or 352x480 has a much lower pixel count" That's exactly what I'm saying. The less pixels you encode, the faster you encode. The same thing applies to say resizing from 720x480 to ~352x240 to encode at 352x240. Unless you are doing a weird resize from 720x480 to 352x480 to encode at 352x240 It's not really the resizing what slows you down. It's the amount of pixels you are encoding.
Just try it. In my machine, a 10 second clip from 720x480 resized to 704x480 is 2 seconds faster than encoding with a target of 720x480 without resizing. 32 seconds without resize, 30 seconds with resize.
As for the streaming, that's fine. But still, you don't get the bonus of filtering to remove low frequency DCT blocks, noise, sharpen, etc, by encoding directly from a VOB.
Take a look at this thread, and see all you're missing
http://www.kvcd.net/forum/viewtopic.php?t=1604

-kwag

[wulf109]

Any option in Tmpeg available from a frameserved source is available on a directly encoded source,but none are needed when you encode fron the vob. There is no better source to encode than the original vob file. You will get color saturation and contrast that are indistinguishible from the original DVD source. No matter how what method is used to frameserver it degrades the vob source.

[crahak]

Set the resolution to 720x480,don't use 704x480 it just slows down the encodeing.

The opposite. Lower resolutions, faster encoding.

-kwag

The resize speed slows it down more than the small reduction in image size.

Also resizing from 720->704 is a bad idea since it does nasty things to the nice high fequency data.

Really? Try encoding a sample at 740 or 720x480 and try a 352x240. Your 352x240 encode will be about more than three times faster.
As for 704x480, tell that to shh, developer of FitCD, and see what he's going to tell you. That's the second resolution on FitCD's choices. It's also the mpeg still standard resolution.

-kwag

Well, I must say kwag is right. The smaller the faster. (Oh. can anyone quote this too? LOL)

[Nasp]

Im having some trouble getting DVD2SVCD to let me create some high resolution SVCDs from DVD source. I want to keep the DVD resolution, and pump the bitrate to 4500 since my player can handle it. Whats the best way to do this? any help is MUCH appriciated!

DVD2SVCD + Cinema Craft Encoder (CBR bitrate 2520 ( 700MB-40min ) )
+ 720*480/576 + audio 48HZ 224Kbps.
With this settings the quality are super. 2,3,4 CD per movie and this way are very very very fast.

Cinema Craft encoder is the best for work with high bitrates on mpeg 2 in CBR mode. One more time, it is very very very fast.

[wulf109]

I'll give you an example. I've been encodeing StatTrek TOS episodes,51 minutes long. Encodeing from the vob it takes Tmpeg 3hrs,20min. at 720x480. I tried an episode at 704x480,all other settings the same. It took Tmpeg 4hrs to encode that 51 min.
The problem is that the pixel count is so close at 704 v/s 720 that the resizing at 704 makes the time longer. The number of pixels is amajor factor in encodeing time but so is resizing.

[kwag]

I'll give you an example. I've been encodeing StatTrek TOS episodes,51 minutes long. Encodeing from the vob it takes Tmpeg 3hrs,20min. at 720x480. I tried an episode at 704x480,all other settings the same. It took Tmpeg 4hrs to encode that 51 min.

Well of course wulf109 , you're using TMPEG to resize. Why don't you use an AviSynth script, and then you'll see the difference. This way, TMPEG will not have to resize at all, because the correct size will already be fed to it by the .avs script.

-kwag

[adam]

Any option in Tmpeg available from a frameserved source is available on a directly encoded source,but none are needed when you encode fron the vob. There is no better source to encode than the original vob file. You will get color saturation and contrast that are indistinguishible from the original DVD source. No matter how what method is used to frameserver it degrades the vob source.

Wulf you have posted similar statements all over this board and you have been repeatedly told how wrong they are. Frameserving itself does absolutely nothing to the source. It does not affect the color or contrast and it does not degrade the quality. Skipping the frameserving process is generally a very bad idea since it causes many unnecessary problems and for ntsc users it forces you to use 29.97fps which makes for a huge decrease in quality compared to ntscfilm. If you do not understand what you are talking about them perhaps you shouldn't post at all. Please for the last time, stop spreading this senseless advice because it is only spreading ignorance. If you have any factual basis for your statements by all means let us know, otherwise consider this your last warning.

[wulf109]

adam

I thought I was living in the United States,were free speach and the expression of opion were guarteed by the constitution.
You are a moderator and I suppose you can block my password,so please exercise your dictatorial right and block me.
My proof are the eyes in my head.

[adam]

The constitution protects free speech but not necessarily stupidity. Your statements are completely false and you have been told this numerous times. People will read your fallacies and follow your advice and they will be worse off for it. Just grow up already and learn some accountability. This forum is about helping people, not misleading them. If this had happened once, maybe twice then it wouldn't be a big deal, but I have personally asked you to stop this nonsense over a dozen times.

You need to realize that the differences in quality that your eyes perceive are caused by the obvious flaws in your encoding method NOT by frameserving. If you can't be bothered to learn how to properly use dvd2avi or avisynth then that is fine, but don't assert that there is something inherantly wrong with these programs simply because you don't understand the fundamental concept of frameserving. Stop beating this dead horse, it has been explained countless times how ridiculous your claims are. Don't do it again!

[wulf109]

Are insults your only means of disagreement. I would never insult or impune in any way your statements or advice.

[adam]

I've asked you nicely so many times in the past that at this point, yes I do feel that some amount of insult is warranted to get the point across.

Bottom Line: Bad advice is worse than no advice at all. Frameserving itself has no effect on quality. To suggest otherwise is to give bad advice, so don't do it.

[wulf109]

In point of fact I have frameserved at least 200 SVCd's with DVD2AVI,forced film and not and used the correct Tmpeg template when I use foced film so I don't think I'm ignorant. I just prefer the results with the direct method. You have alot to learn if you think that insults are ever an approiate response.
And yes the constitution protects me even when YOU think it's stupid.

[snowmoon]

wulf109:

You are out of line. Knock it off or I will close this thread.


And back on the whole 720 vs 704. If bitrate is really that important you can acheive a better reduction through letterboxing the top and botton 16 pixels from the fullscreen image. You can also crop 16 from the sides, but that does show up on some TV's. Resizing 720->704 will knock out the fine detail in the image.

[kwag]

Resizing 720->704 will knock out the fine detail in the image.

..... that only a microscope can see

But then, the 2.22222% reduction also applies in file size reduction. And that, my CD-R can see as ~2 to ~3 additional playing minutes

-kwag

[Zaun]

Ok... So... best way to create with CCE and DVD2SVCD??? When i edit the avisynth file, i get a field like 480,480,0,0,0,6 so I change it to 720,480,0,0,0,6 ... or...

Anything else I'm missing?

also when I select CBR, it doesnt give me any option to set the actual constant bit rate, I simply set a minimum and maximum in the bitrate folder of DVD2SVCD...

Im such a newbie

sorry

[SansGrip]

Frameserving itself does absolutely nothing to the source. It does not affect the color or contrast and it does not degrade the quality.

Sorry for bumping up an old thread, but I can provide some proof that wrt TMPGEnc frameserving could theoretically cause a drop in quality (albeit an insignificant one).

As far as Avisynth 2.0 goes, the colourspace conversions that usually take place are as follows:

YV12 -> YUY2 in mpeg2dec
YUY2 -> RGB in VFAPI plugin
RGB -> YV12 in TMPGEnc

If one were to read a vob directly with TMPGEnc, I believe the process would instead be:

YV12 -> RGB in VFAPI plugin
RGB -> YV12 in TMPGEnc

It may well be possible to implement the YV12 to YUY2 conversion losslessly (since one is upsampling from 4:2:0 to 4:2:2), but this is complicated by interlace issues and I'd have to examine the procedure in more detail -- or ask someone more knowledgable than myself -- to give you a definitive answer.

I'm not saying it does introduce errors, just that it could . The "problem" could also be averted by using Avisynth 2.5 which supports YV12 natively.

Skipping the frameserving process is generally a very bad idea since it causes many unnecessary problems and for ntsc users it forces you to use 29.97fps which makes for a huge decrease in quality compared to ntscfilm.

Since TMPGEnc would use the DVD2AVI plugin to read the vob(s), and assuming the plugin respects the field settings, wouldn't it still pull in 24fps material even without Avisynth? This of course assumes one doesn't use the DirectShow plugin to read the vob, but let's not go there...

Also, one might be able to obtain the same results when pulling directly from a vob by using TMPGEnc's built-in IVTC, but I'm guessing it's not as sophisticated as something like Decomb.

Just playing devil's advocate .

Regards,
SansGrip

[adam]

Well when I said frameserving I wasn't really talking about avisynth at all, though I wasn't aware that it would make any difference. Wulf's numerous posts on this matter have suggested skipping dvd2avi altogether and loading the vobs directly into TMPGenc. This is the "frameserving" process I was specifically referring to. From my experience if you load the vob directly into TMPGenc you do get the telecined material ie: 29.97fps. Sure just loading the d2v into TMPGenc w/out avisynth can give you the 23.976fps, this is exactly why you should at least use dvd2avi. As long as tv scale is set then the conversion to RGB should still be done by VFAPI plugin in the exact same manner, so no potential for quality loss and %20 less frames to work with plus progressive frames...can't beat that.

As for using TMPGenc's IVTC, if it even works correctly, which from my experience is pretty rare, you are still looking at a significant increase in encoding time. This is the beauty of dvd2avi's forced film and why I have so strongly advocated it over direct input of vobs to TMPGenc. Sure decomb would give even more consisent results but the best way to get 24fps material is to simply not telecine it at all. I see no reason to telecine something just to inverse telecine it immediately after. Forced film is the perfect solution.

On to avisynth...I wasn't aware of the extra conversion. When I use TMPGenc I usually do just load the d2v directly, and when I use CCE I do use avisynth but I am able to skip the conversion to RGB altogether. So perhaps this is an issue to pay attention to when using avisynth, but in any case it is no justification for directly loading the vobs into your encoder, and by that I mean skipping dvd2avi.

Also, as you noted, I kinda doubt the extra conversion in avisynth would even yield a noticable difference. It certainly could never cause enough damage to offset the %20 decrease in quality you get when encoding in ntsc vs ntscfilm.

Basically my point is don't skip dvd2avi unless you actually have a valid reason, and increased quality certainly isn't a valid reason since you actually get the opposite.