AVI2SVCD - Why does it separate audio & video?

[sync]

I know there must be a good reason that the audio and video are separated. But I haven't found an explanation and would like to understand this.

[hoozdapimp]

Hi,
Not really sure if I understand your question, but to convert an AVI to an SVCD, this is a necessary step. An AVI consists of a an audio stream and a video stream muxed together. So AVI2SVCD seperates the streams, converts them both individually and muxes them back together in MPEG2 format.
Scott

[sync]

I don't understand why it's a necessary step. I've seen guides that discuss avi to svcd and don't separate the audio and video.

[hoozdapimp]

then maybe you'd like to point us in that direction...i don't see how this is possible.

[sync]

I took a look at the guides again. The CCE guides seem to always separate the audio and video. The ones for Procoder and TMPGEnc often don't. I guess if AVI2SVCD has to do it for one encoder then it makes sense to use the same method for all of them.

[DDogg]

I guess if AVI2SVCD has to do it for one encoder then it makes sense to use the same method for all of them. Exactly, but it is more involved than just that.

DVD2SVCD is designed to mimic the methods used by the most demanding hand encoders for the creation of SVCD. The highest standard for down conversion of ac3 to mp2 is realized by using BeSweet and its component parts. Historically CCE was deemed very inferior and in fact, the earlier versions would just crash when attempting audio conversion. This was well before TMPG, also considered poor in audio, and much later ProCorder were supported.

Also, SVCDs in general are done in multiple disk sets of at least 2 disks (NTSC also requires that a pulldown operation be done upon the video only stream). It so happens that the best multiplexer of the "pulldowned" video and audio (BBMplex)is also the best way to cut the large video files in to disk sized chunks while doing the mux.

Avi2svcd was an addition bolted on to the original program and, btw, was originally intended for DV conversion and not this silly devil's brew of "I rolled my own brand of avi to make me look cool" of the week that we see now and it followed the same method of conversion used for DVDs. Also, do remember that svcd allows multiple audio tracks as well as subtitles. All of these require the muxing of the many separate streams as one final cohesive operation while at the same time also doing the cutting needed.

Hope this helps you better understand. I will add that your questions indicate you do not really have much of a grasp of all the operations involved. It takes a while. Just give it time to soak in.

[sync]

Thanks DDogg,

I'm the type of guy who isn't satisfied with a step-by-step guide. I need to know why things are done the way they are. Your post was very helpful.

And I'll try to take your advice to "give it time to soak in". :)

[DDogg]

It is funny you said that. I was just thinking that one of the best things a person could do who actually wanted to understand all the bits and pieces would be to go back and trace one of the "old" ways. Here is a shorthand for one of the simplest (NTSC):

1> rip - DVDDecrypter
2> create d2v file and wav - DVD2AVI 1.76
3> serve video to encoder of choice - Avisynth 2.51 using mpeg2dec3.dll (also may need VFAPI/Link2 depends if encoder loads AVS or not)
4> Add pulldown - Pulldown.exe (NTSC)
5> encode wav to mp2 - toolame.exe
6> Mux and cut - BBMplex or Mplex.exe
7> Create image and burn - VCDEasy

If you do all that, you will have a very good feel for the job. Additionally you would then want to add subtiles, multiple audio tracks, Chapter extraction from original DVD and placement in svcd image, etc.

[sync]

My source is DV AVI that is captured from Tivo. The sample I'm using is telecined and needs IVTC, so everything from step 3 on is relevant (except for mpeg2dec3.dll).

I tried serving the video to CCE using an Avisynth script that does IVTC, resize, noise reduction and sharpening. The noise reduction is really slow and I was doing several passes on CCE. Apparently each CCE pass reads the AVI file, so the noise reduction, etc. is done several times which takes a really long time (26 times longer than the clip).

Someone suggested filtering the file and saving it to a temporary file with a lossless codec and then feeding that file to CCE. That makes sense but some of my source files will be 20GB and filtering that to a temporary file will require another 60GB. Even saving the temp file to DV isn't feasible right now.

So I'm stuck on trying to figure out how to do this without using up so much disk space or time. Would it work to serve the DV AVI to VirtualDub and then serve it's output to CCE?

[edit] I just tried it and it seems to be working. But, although I think this is very cool, I just realized that all I've done is create a more complicated and even slower way of generating the same problem. :)

[DDogg]

I just realized that all I've done is create a more complicated and even slower way of generating the same problem. Nah, you just learned a boatload of more stuff is all :) Hey, on the cce thing, just do it and go to bed. Also, if you set the piority to low, it will happily run in the background (if you have plenty of ram). 26x? I think you are running way too many filters! (or have a P133? :) )

[sync]

It's a 1Ghz P3. It's the pixiedust filter that is so slow. I figure a 20GB file will take 52 hours to encode. I guess I need to find another noise reduction filter.

[sync]

Originally posted by DDogg
It is funny you said that. I was just thinking that one of the best things a person could do who actually wanted to understand all the bits and pieces would be to go back and trace one of the "old" ways. Here is a shorthand for one of the simplest (NTSC):

1> rip - DVDDecrypter
2> create d2v file and wav - DVD2AVI 1.76
3> serve video to encoder of choice - Avisynth 2.51 using mpeg2dec3.dll (also may need VFAPI/Link2 depends if encoder loads AVS or not)
4> Add pulldown - Pulldown.exe (NTSC)
5> encode wav to mp2 - toolame.exe
6> Mux and cut - BBMplex or Mplex.exe
7> Create image and burn - VCDEasy

If you do all that, you will have a very good feel for the job. Additionally you would then want to add subtiles, multiple audio tracks, Chapter extraction from original DVD and placement in svcd image, etc.
I'm on step 6 and wondering if you meant bbMpeg instead of BBMplex?