TMPGEnc 2Pass VBR Confusion

[BOXINGJUNKIE]

This post raises alot of good questions for me. I have an ATI All in Wonder Radeon 8500. Im using MMC 7.7 that came with the card to capture 720x480 MPEG2 DVD Quality. I dont do anything to the video once it is captured. Just burn to DVD. My question is: Would I achieve a better picture if i used something like Virtualdub to capture as AVI, then use a program like CCE or TMPGE to convert to MPEG2. Would I achieve a better picture if I used 2 pass VBR? I know it would take alot of time to do this but If I can get the best picture possible, then it is worth it.

Thanks Boxingjunkie

[jolo]

This post raises alot of good questions for me. I have an ATI All in Wonder Radeon 8500. Im using MMC 7.7 that came with the card to capture 720x480 MPEG2 DVD Quality. I dont do anything to the video once it is captured. Just burn to DVD. My question is: Would I achieve a better picture if i used something like Virtualdub to capture as AVI, then use a program like CCE or TMPGE to convert to MPEG2. Would I achieve a better picture if I used 2 pass VBR? I know it would take alot of time to do this but If I can get the best picture possible, then it is worth it.

Thanks Boxingjunkie

Well, I'm a big boxing fan myself.

I have a Radeon 7500 all-in-blunder.

What do you use to burn you DVD ? What are you capturing ?

Sounds like you have been doing great.
The question about the picture quality is one you have to answer really. 2-pass is about less media needed with same quality, it is not about improving quality. If you use CQ, it will give the highest quality, but with the most disk space.
Virtualdub for me has given me too many dropped frames, so I don't use it anymore. I tried it with huffyuv, divx and others.
If you convert your file to compliant SVCD using TMPG, then you probably will have to go to 480 x 480 is you are using NTSC.

These things are so flipping individualized. I don't know what you are capturing from, but I would capture a small sample, maybe 2-10 minutes where there is some dialogue going on and some movement. Maybe one time using virtualdub and outputting a AVI file using huffyuv. Then encode using TMPG or CCE.
Use your current method as well, output a AVI using MMC as well. Try to encode the small sample and see what you get. Look at some of the guides and see which might be best for you.

What I am saying is, the best way is the best way that works for you, it is your time, your PC configuration, your DVD player. I really think that there is no "best" way for all, when it comes to capturing. There are too many factors going on. There also are other products as well you can use. Just depends how much time you want to put in messing around with it and what is acceptable quality versus time spent.

There are many guides here, you might try some and see how they turn out with the small captured file. But what is so frustrating about capturing is that you can read here about vdub is the best thing in the world, or that it gives too many dropped frames, or the ATI drivers stink, use new ones, or the ATI drivers work for someone else.

If you are getting a good solid DVD picture and are capturing without dropped frames, then you are in darn good shape.

I don't know if I confused you or helped you. But please keep in touch, I'd like to find out what you do. I have yet to come with the "perfect" way myself. I've been trying to capture video tapes, with some great success and some not so great success.

Jon

[BOXINGJUNKIE]

I am capturing fights off my DTV satellite. I have an S cable from the satellite reciever running straight to my ATI card. Im using MMC 7.7 to capture 720x480 MPEG2. Then I use SpruceUp to author, and Prassi to burn. I think the pic looks pretty good, but Im wondering if I captured the same video using Virtualdub as an AVI, and then TMPGE to convert, I can then run it through TMPGE filters, would I end up with a better pic. MMC 7.7 doesnt have any filters. So when I watch the fights that Ive made I get some blocks when the punches really start flying. I have also wondered if it is my tv that is just making it look that way. I have a Sony 36 inch XBR. Just wondering.

Thanks Boxingjunkie

[DJRumpy]

Think of MPEG-2 as MPEG-1 with alot of the restrictions removed. The actual streams are use the same basic structure.

http://viswiz.gmd.de/DVP/Public/deliv/deliv.211/mpeg/an&dif01.htm

Support for VBR, and interlaced video are key differences, as well as a higher bitrate maximum. An MPEG-2, encoded at 352x240, and a MPEG-1, encoded at 352x240, at the same bitrate with the same encoder, should look about the same.

BOXINGJUNKIE, you should definately capture to AVI if you want to improve quality. Your capture will usually suffer from noise, and motion detection errors, when encoded directly to MPEG. The good MPEG hardware capture cards usually start around $1000.00 US and go up from there .

If you capture to AVI, you can pre-filter your AVI before burning. Use a temporal smoother to remove noise. I do my caps using VirtualDub, and the VirtualDub start timer to auto start/stop my recordings.

You should capture using the huffman codec, as it's lossless, but it requires a VERY large hard drive(s). Optionally, use one of the newer MPEG-4 codecs, like DivX, or XviD ( mpeg4 requires a fast cpu [P4] or equivelent to capture at higher resolutions). You can capture hours of video in a GB or two. Look into getting CCE for your encoding. You should also learn how to use AVISynth. It makes removing commercials a snap. You simply TRIM them out with AVISynth. It takes about 10 mintues to parse through your source video, to find the commercial points, and to trim them out with AVISynth.

[jolo]

DJrumpy,

I believe that, as some user pointed out in another post, how some people use opinions as facts.

First of all, it is debated on whether huffyuv is lossless. Second, you might need 30-60gb of open work area and it better be defragmented before hand.

As I mentioned before, I found vdub to give me too many dropped frames. This happenes to some people, but not to others. The ATI cards seem to be particularly vulnerable to this.

It is true that converting to MPEG-2 is lossy, but how lossy, meaning would you notice the difference by looking at it and is it worth your while to spend 8 more hours of encoding times, having a large huge amount of free disk apce to possibly get a less lossy result. The answer is, that it is up to you !!

Trust me boxingjunkie, look at some of the guides and opinions and go with the way which gives you the results that you are satisfied with. All of us express their opinions here, but the best result is the one the suits your configuration, life style and movie viewing habits. Try different techniques and pick which is correct for you personally.

Jon

[DJRumpy]

I had no idea there was a debate about the huffy codecs. Can you supply me with more info? The author of the codec warns that selecting conversion to YUY2 will result in some loss in quality, but you are not required to use those options. The standard options should be lossless, as documented.

You'll have to forgive me on the MPEG opinion. I use CCE or encoding. The encodes take about an 1 1/2 hours for my captures. BOXINGJUNKIE indicated he wasn't satisfied, and asked about alternatives.

If he captues in MPEG, he cannot do 2-Pass. That being the case, he should capture to a lossless AVI, so he doesn't lose quality encoding twice, once during capture, and then re-encoding for filtering, or two-pass.
jolo wrote:

Sounds like you have been doing great.
The question about the picture quality is one you have to answer really. 2-pass is about less media needed with same quality, it is not about improving quality. If you use CQ, it will give the highest quality, but with the most disk space.
Virtualdub for me has given me too many dropped frames, so I don't use it anymore. I tried it with huffyuv, divx and others.

2-Pass will certainly improve quality, especially on hard to encode scenes (high action), scene changes, and fades. This is well documented.
BOXINGJUNKIE wrote:

So when I watch the fights that Ive made I get some blocks when the punches really start flying.

This is a classsic description of macroblocking due to bit shortage. It's also common when capturing direct to MPEG. Capturing to AVI, will allow better encoding options to MPEG-2 later on.

Jolo, if I've offended you in some way, I apologize, but your post is very defensive. No one is holding a gun to BOXINGJUNKIES head you must do it this way. This is a forum. He asked for opinions.

[jolo]

DJRumpy,

You obviously are a class act.

I have gotten my head spun around more times than Linda Blair in the Exorcist (I hope you understand the reference) by statements on the forum, which were valid opinions and techniques which work for some, but entered as facts, then contridicted later, etc. Spend countless hours trying to figure them all out. I have been agonizing over the capturing issue as well.

I made a general statemet that I saw someone else put enter on a posting and I thought it was really cool and think we could all (including me) maybe state when something is their opinion, say things like "this works for me" and of course remove statements like "it sucks", or "it stinks", etc. OF course weighing out quality versus effort and time is always something I juggle with.

Now that my generalization is over with, it sounds like we are totally in synch with each other. I don't think there usually is a single best way and that taking a small sample and trying several techniques, from others experience and from guides and see what works best for each of us is the only way I know to do it.

The great thing is that BOXINGJUNKIE is starting from a point where he has a technique that is working !! Which is a better one than from me using my ATI 7500 all-in-blunder card.

DJ, you are not only astute technically, but had the humility to communicate to me ending possible misunderstandings and escalations of our blood pressure. It is very easy to miscommunicate intent on BBs and e-mail.

I will send you some information from guides, and postings about huffyuv.
The more we learn the more we have to share with each other.

Thanks so much for your thoughtful reply.


Jon

[DJRumpy]

If you find them, let me know. I'm currently using huffy, but I'm always striving for better captures. If it turns out it is lossless, then I'll want to move to something else.

::Edit::
Just an afterthough..to anyone using this forum, should consider the moderators above general users such as myself. Even the moderators make mistakes, but chances are the information they put out here is more accurate than most.
::Edit::

[groyal]

I believe that, as some user pointed out in another post, how some people use opinions as facts.

I believe that people are sometimes confused about the difference between an opinion and a fact, causing them to alter the facts to fit their opinion rather than the other way around.

First of all, it is debated on whether huffyuv is lossless.

I debate that. :)

HuffYUV is a lossless YUY2 codec. That is a fact.
HuffYUV is the best lossless YUY2 codec. That is an opinion.

All Huffman-based encoding schemes are 'lossless' because the original data can be recovered from its compressed representation intact. This does not mean that if you use a Huffman-based codec you won't drop any frames -- that's not the kind of loss the description represents.

'Lossy' compression methods like JPEG and MPEG work by throwing parts of the original data set away. The higher the compression level, the greater the loss. If the compression is moderate you can actually trick the human perceptual system into not noticing the difference. But if the compression is severe, the data is reproduced with so many errors (artifacts) that the quality of the recording is ruined.

'Lossless' compression methods like Huffman (used in the ZIP file compressor and the HuffYUV codec among others) use a subsitution table to map sequences with a long bit length to sequences with a short bit length. The subsitution table can be constructed on-the-fly (as with ZIP) or a pre-computed table can be used (as with HuffYUV). In either case, when the data is decompressed the substitution table is used in reverse: short sequences get mapped back to long sequences, restoring the original file.

I'm currently using huffy, but I'm always striving for better captures. If it turns out it is lossless, then I'll want to move to something else.

Here is an opinion: I've had very good experiences with the PicVideo Motion JPEG codec. Lossless and lossy versions are provided in the package. I don't use the lossless version because being, well, lossless, it doesn't compress the data to any significant extent. The lossy version on the other hand gives you fine control over the compression ratio, allowing you to fit lengthy captures into the disc space you have on hand.

It's very fast, plays back in real time, and outputs native YUY2 so the MPEG encoder doesn't waste time on colorspace conversion from RGB. It's also a good "VCR codec" for shows you want to capture for later viewing but not necessarily burn for posterity.

One caveat though: when a dropped frame does occur, it's a bad thing because the codec doesn't insert a repeat-frame unless your capturing program supports it (AVI_IO does). Without it, other programs will choke when they attempt to decode a frame that isn't there.

If you haven't tried MJPEG in awhile it's a nice alternative, particularly for captures from VHS tape.

[BOXINGJUNKIE]

Hey DJRumpy you said to use a Huffy codec. I havent use VDub in about a year. But I did use a Huffy codec back then. Is there newer codecs out that would work better? Thank you all for your help.

Boxingjunkie [/quote]

[DJRumpy]

If I'm cramped for space, I'll use DivX, 1-Pass @6000Kbs. The quality is excellent.

[BOXINGJUNKIE]

I have plenty of space. I have two 80 gig hd. Space is not a concern. Im not sure I understand these DIVX codecs. I can use a DIVX codec to capture an AVI at 720x480. Then use TMPGE to convert from AVI to MPEG2? Isnt DIVX a different format than say MPEG or AVI? Hope Im not confusing you.

Thanks Boxingjunkie

[adam]

Divx is just a codec, actually there are several different versions, and you can capture to any codec provided your capturing software supports it. Its actually an mpeg4 codec but is most commonly used in the avi container, basically you would be capturing to an avi file.

Divx really is not meant for capturing but, like DJRumpy said, if you are really short on hard drive space it is a good alternative since it has such a good compression ratio. If you have the space than definitely use huffyuv or mjpeg.

[BOXINGJUNKIE]

I downloaded Huffy 2.1.1. Unzipped it to my VDub folder. But when I open VDub and go to compression, there is nothing there. It just says (No recompression: YUY2). It wont let me adjust the settings or nothing. When I go to video format it gives me the option to choose the size, 720x480, then it gives me the option to choose YUY2 or UYVY. I want to choose the YUY2 right? But it isnt giving me anywhere to adjust the settings. Any help aprreciated.

Thanks Boxingjunkie

[oceanlvr01]

OK, While we are on the subject of Divx, I have a question..
I have the Canopus 100..I love it..
As you know however, it captures DV.
Here is my problem..
I want to convert this DV into Divx..to burn on a CD.
Now I know that VirtualDub does this. I can not however , to save my soul, get VirtualDub to work on my PC.
It is a fast enough PC.
It is an AMD1700, 512MB PC2700 ram, 80GB hard drive...etc...
It is also running the XP platform.
I have the Canopus codec loaded on my PC.
I have the Nimo 5.8 codec pack on my PC.
Heck, I have the K-lite codec Pack on my PC.
I have added codecs one-by-one and, by now, I should be able to decipher anything....INCLUDING alien text and the Egyptian hieroglyphs..
Yet.....VirtualDub thinks that I do not have a codec.
I have played with it...pleaded with it...heck, I have considered throwing chicken bones at it and doing dances in front of it....
Anyways..I am getting off track...
Just venting....
As I have given up on VirtualDub..I am find it hard to get a program that will allow me to convert DV to Divx.
Anybody got any ideas?
Thanks

[DJRumpy]

BOXINGJUNKIE, you have to install the Huffman codec, before it will be available in the VirtualDub compression menu. It should come with instructions on how to do this. Once you have installed it, from the VIDEO menu, select VIDEO | COMPRESSION. You should get a list of all of your installed video codecs. Find Huffman 2.1.1 in the list. Once you highlight it, select the OPTIONS button. On the options page, for "Compression Method", select "Predict Left (fastest)". For the "RGB Compression Method" setting, select "<--Convert to YUY2". NOTE: This last option is 'slightly' lossy, because it converts your output from raw RGB, to YUY2, but I can't tell a difference. For options at the button, you can select "Enable RGBA (RGB with Alpha) compression". Click OK

To answer your other question, it's unlikely, that you could capture at full 720x480 using DivX, unless you have a top of the line processor/memory config (say 2.5Ghz or faster). I get dropped frames at that resoution with a 2.4GH, so I have to drop to 640x480 if I capture using that codec.

Huffy doesn't compress that much (about 2.1:1), so the CPU overhead is very minor. You should be able to capture at the full 720x480 with no problem. If you look under the VIDEO | SET CUSTOM FORMAT, you'll see "UYVY YUV 4:2:2 interleaved" selected by default. This option is fine.

[BOXINGJUNKIE]

Thanks for your help DJRumpy. I did get it working. Im going to give it a try and see how it works. Thanks again for your help.

Boxingjunkie

[SaSi]

I have missed the thread for some time and did some catching up after doing a new capture from VHS. This time I used the hufyuv 2.1.1 codec (as pointed out in this and another thread).

I used VirtualDUB 1.5.1 optimized for P4 and the cappture was done through an ASUS 9180 VGA.

Lossless or not, (not a challenge here) the hufyuv codec is excelent both in terms of quality and in terms of editability. The last part is important. DivX can produce similar (at least better than what VHS needs) quality for a fraction of the disk space. However, hufyuv doesn't do any temporal compression (all frames are complete pictures) and so one can cut and paste with absolute precision on the original capture. I didn't realize this aspect until editing became a snap.

On the downside, disk requirements are huge (and as always, defrag before use is recommended). It takes 30Gb for an hour or so at 720x576 but performance was good and no lost frames (only the usual 1-2 at the beginning of the capture).

[BOXINGJUNKIE]

DJRumpy you said to use a temporal smoother. Whenever I add the smoother it will only let me use it at 320x240. Is there anyway to change this to 720x480?

Thanks Boxingjunkie [/quote]

[DJRumpy]

I used the temporal smoother that comes with MPEG2DEC, with the AVISynth frameserver. If your talking about the temporal smoother in VirtualDub, reporting the 352x240 resolution, don't panick. It doesn't seeem to affect output size.

[dphirschler]

I think I finally understand what the 2-pass settings do. And I also understand why a high minimum setting can actually harm your quality. The bitrate can vary from scene to scene to any value between the minimum and maximum bitrates... but it must maintain an average bitrate that you set. So, your final file will have an average bitrate much lower than the constant bitrate in CBR mode. This gives you a much smaller file. Setting the minimum as low as possible gives the encoder more bits to use in the high action scenes. Setting the mimimum high is stealing from the parts of the movie that need the bits. If you find you need to raise the quality of a VBR encode, then raise instead your average bitrate. Raise the minimum bitrate only if the file has problems playing on your DVD/SVCD player.

Can somebody post the minimum, average, and maximum settings that are within SVCD standards?

FWIW, I once encoded 2 hours (from a DVD rip) using SVCD VBR (using the usually piss poor default settings), and it looked pretty damn good. But I was amazed at 120 minutes on an 80 min CD!


Darryl

[adam]

You got it, dphirschler.

SVCD standard supports any video bitrate between 0-2600kbits, of course you could never actually allocate 0 bits though. Basically you can set the minimum as low as you like. As long as you don't run into specific playback problems (player chokes or low motion scenes don't get enough bitrate) then you should always use 0-500.

The total bitrate, however, cannot exceed 2778kbits and this includes video + audio + subs. So factoring in a reasonable audio bitrate, you are looking at a max video bitrate of around 2500kbits.

These numbers really only exist on paper though. Most standalone dvd players can support higher, and some even require lower despite what the standard says.

[DJRumpy]

Just use a bitrate calculator, to determine the highest average bitrate, according to the length of your movie. There's a couple of good ones in the TOOLS section. They'll get you close to your target size. Using multiplass will shrink the output fill a bit more for each pass, so it's not exact, but what in life is?

[adam]

DJRumpy the number of passes should have no effect on the filesize. Filesize is solely determined by the average bitrate, and as long as you do at least 2 passes then you can set this to whatever you like.

I know you say your experiences show otherwise, but I have never heard of this happening for anyone else. At the same avg setting, it doesnt matter if you do 2 passes or 200, the filesize should still be the same.

[DJRumpy]

You know, I never did get around to re-testing this. I'm wondering if I used a CQ VAF file, for a second multipass run. When in doubt, listen to Adam, as I haven't verified my results as I should. My pile of download Divx/AVi/SVCD "To Be Converted" has been growing too large, and I haven't been spending the time I should on following up with some of my posts.

[gf]

Hello
I was wondering if someone could help me on this one. How or could you get a MPEG2 clip into CCE?

Thanks Steve

[adam]

gf you have to frameserve. The best way is to use avisynth. Create a d2v project file in dvd2avi and, with avisynth installed of course, write an AVS script. If you are unsure how FitCD is a wonderful bitrate calcualtor that can also create AVS scripts for you. You can then load the AVS script directly into most versions of CCE. Avisynth is the fastest possible way to frameserve, and there are dozens of excellent filters that you can implement in your AVS script. This guide might help, https://www.videohelp.com/forum/userguides/87270.php

The other, slower, method of frameserving is with Virtudub. You can use VFAPI converter to create a psudo (fake) avi file from your d2v file, or you can use the new modded version of virtualdub which can directly accept mpeg2 input. From there you can frameserve from Vdub to CCE using this guide. https://www.videohelp.com/virtualdubframeserve.htm

[DJRumpy]

Adam, just ran a few tests. I wasn't able to reproduce the shorter MPEG. Dunno what I was doing to get it. I'm also using a different version of CCE now. Came out about the same every time.

[yg1968]

jolo said:

First of all, it is debated on whether huffyuv is lossless.

I may have said that Huffyuv wasn't truly lossless in other posts. I was wrong. I just assumed that compression implied some loss of information (whether it's called lossless or not). But I guess that HuffYuv only throws out duplicate information. (I got tired of hearing that DV is not as good as Huffyuv because DV is "lossy".)

Anyways here is a good explanation of various codecs:

http://www.animemusicvideos.org/guides/avtech/video3.htm
http://www.animemusicvideos.org/guides/avtech/video4.htm

2) Lossless Compression
Lossless compression, as the name implies, means that after compressing the video, and then decompressing it, you wind up with the exact same data as you put in. This is comparable to something like ZIP or RAR (Infact, the best lossless codec out there, Huffyuv, basically uses the same compression algorithm as ZIP on each frame of video to achieve its compression). Lossless has the advantage that no matter how many times you compress it, you still haven't lost any video data. The bad part is that most often you don't save nearly as much space as you would with other lossy compression algorithms. Modern adn well-known lossless video codecs include Huffyuv (mentioned and linked earlier in this paragraph), Lossless MJPEG (different than plain old MJPEG), and zAVI (not recommended, huffyuv is vastly superior).

3) Huffyuv - Lossless Video Compression, for a price.

Origins: A guy named Ben Rudiak-Gould wrote this wonderful little codec, located here, which is a lossless compression codec for both YUV and RGB video data.

How it Works: Huffyuv is a lossless codec. The name stands for Huffman-compressed YUV. Huffman was a guy who came up with something called "Huffman entropy encoding" which is basically what all lossless compression is originally based off of (Huffman coding is used in things like ZIP, RAR, etc). So you could say that basically Huffyuv simply ZIPs every frame for its compression.

Benefits: Obviously, the biggest benefit is that you have a perfect recreation of the original video data (unless you do colorspace conversions). That means no matter how many times you recompress the video in Huffyuv, you'll still have the same video data that you had in the beginning. This is a very nice thing.

Disadvantages: Unfortunately, while compressing video in Huffyuv is pretty fast, decompressing it is not. Also, due to the very large space requirements (Huffyuv takes a LOT of disk space, sometimes 4x more than even DV), disk throughput becomes a very large factor. Playback at full resolution and full framerate is a daunting task for even the fastest computers.

Recommendations: I only use Huffyuv when I want to store something, only to open it in another application like VirtualDub so I can do some filtering/etc on it. If you have a capture card that allows you to choose your codec while capturing and Huffyuv is an option, by all means use it. But it does take quite a lot of disk space and processing power.

[Anonymoususer]

It's amazing how far my one little question has gone...well good news people, I got a monster upgrade and will begin testing soon, on my 2.4GHz Processor, 512 MB DDRam, 80 GIG Hardrive custom built for
700$ computer...it's amazing at how fast it runs...just amazing...now I can ACTUALLY start to encode properly...