Convert m2t to avi using AviSynth

[manono]

Why can't you make one D2V for the entire thing by opening all the M2T files in DGIndex at once?

[Soopafresh]

How about joining the M2T files with TSjoiner http://www.softpedia.com/progDownload/TSJoiner-Download-38806.html

Then convert the joined file and you'll have 1 AVI

[PeterK]

Hello,

ok I use TSjoiner to merge m2t files.
I have still problems with the aspect resolution from the recorded 1440:1080 m2t file to the smallest pal 16:9 resolution that I can use.
Is 1440:1080 to 1024:576 the best way to go?

Peter.BE

[Soopafresh]

1024x576 for the conversion to MJPEG 1:1 A/R, then 720x576 16:9 A/R for MPEG2/DVD encoding

[PeterK]

Ok Soopafresh, thanks.

I will use 1024:576 for editing, 1280:720 resolution avi files are to big for my pc.

Peter.BE

[grkiwi]

Sorry for undigging this topic but I realy need some help here.

What I want to do is batch convert very many m2t files I have captured from my Sony s270 HDV camera.
I want to convert all these the files to DV avi files with an Avisynth script I have made but I definetly need to preserve the file name of the m2t file in the corresponding avi file. I found the fantastic bat script from Soopafresh but I have no sound in the final avi. The sound is in a seperate wav file.
I have tried manipulating the script but unfortunately my lack of knowledge in scripting doesn't let me register the wav files in the script.
I am asking for some help in registering in the avs script the corresponding wav file by batch creating a script like the one below

setmemorymax(512) 
v=mpeg2source("\test\m2t\00_0062_2008-05-31_133120.d2v") 
a=wavsource("\test\m2t\00_0062_2008-05-31_133120 MPA PID 814 DELAY -264ms.wav") 
audiodub(a,v)
converttoyv12() 
AssumeFPS(25) 
AssumeBFF() 
separatefields() 
lanczos4resize(720,288) 
converttorgb() 
Sharpen(0.2) 
weave()

I have figured everything else out except the part of how to get the wavsource with correct filename line in the script

Can anyone help me out here?

[Soopafresh]

Give this a shot:

m2t_preprocessor_grkiwi.rar



It uses Xport.exe to do the demuxing, which should take care of the A/V sync issues you normally must deal with in DGindex. The PC I'm doing this on doesn't have speakers, so I don't know for sure.

Also included a test method of using ffmpeg to convert the AVS file to DV . You'll have to play around with the multitude of ffmpeg parameters if you want to go in that direction. The example I included needs work. If you don't want to do that, just delete the bottom 3 lines from the script.

A couple of things about your script:

- I think all HDV is TFF
- No need to convert to RGB if you're converting to DV .

Good job doing those slow camera pans. I really like what your footage looks like deinterlaced. I also wish I was there!

test.avi

[grkiwi]

Good job doing those slow camera pans. I really like what your footage looks like deinterlaced. I also wish I was there!

As usual Soopafresh is AWESOME.

At this stage, I wish I was there too!!!!

I will give the method a try and get back with results.

As for the script, I came out with this setup by trial and error and found out that this can get into Procoder and then into Edius with no problems at all. Possibly it is not the most optimal script but it worked. As for the color conversions I will try and optimise them as soon as I get to a final script.

Thnx again!!!

[grkiwi]

Well I am in the right track thanks to Soopafresh.
The script worked, the synch is fine (better than the other one I think!). The only part left to sort out is the dv avi compression. ffmpeg does the job but I think is a bit slow. The aspect and field ordr in the final video need changing too.
Will work on it a bit and probably get back with a working (100%) script.

[grkiwi]

One more question...

Do I need to convert from 709 to 601 colorspace when going from HDV to DV? If yes how would I put a colormatrix() line in the script?

Cheers

[Soopafresh]

Using your favorite text editor change this line:

for %%a in ("*.d2v") do @echo v=mpeg2source("%fpath%%%a") >>"%%~na.avs"

to this:

for %%a in ("*.d2v") do @echo v=mpeg2source("%fpath%%%a").ColorMatrix(interlaced =true) >>"%%~na.avs"


Make sure you put the Colormatrix dll in your C:\Program Files\AviSynth 2.5\plugins\ folder

By default, Colormatrix converts 709 to 601 so the above line should be correct. However, I don't know if DV is Rec 601. I don't think it is, as the colormatrix docs only refer to Divx/Xvid.

Finally, I don't know which DV converter app you're using, but make sure you save to 16x9 aspect ratio.

[FranklinQuinten]

If you need to maintain separate audio tracks, what would be the workflow? If I capture via HDV Split and sync and then convert to AVI for import into a NLE, the tracks will be joined, no?

[poisondeathray]

If you need to maintain separate audio tracks, what would be the workflow? If I capture via HDV Split and sync and then convert to AVI for import into a NLE, the tracks will be joined, no?

I'm not sure what you mean by audio track(s)? Do you have multiple tracks from different sources?

You can use dgmpgdec to demux the audio, it will automatically give you the delay (if any), and mpeg2source() with the .d2v file for the video

If you are using the DirectShowSource() method you can specify audio=false, and use lagarith or uncompressed avi for video only (you can still demux the audio with dgmpgdec or tsmuxer)

(I'm also assuming your camera is using MPEG2; if it is using AVC/h.264 use dgavcdec instead of dgmpgdec)

[FranklinQuinten]

Multiple tracks meaning two tracks of audio from 2 different mics - one wireless and one on the camera.

When I converted .m2t file using TMPGEnc, the two tracks of audio were joined together from the two different sources so I couldn't adjust the levels of each one individually in my NLE.

Will the above workflow correct that?

Thanks for your response!

[Soopafresh]

I can only imagine you mean the audio is on one stereo track - Left channel from the wireless and Right from the built in mic. (or vice-versa)

[poisondeathray]

Yes, you can keep audio & video separate to feed into your NLE with either method

I would ditch TMPGEnc. You can use either the DirectShowSource() method or dgmpgdec method to have a video without audio. Using this script you feed into an encoder; you can use vdub for example. I would use either lossless compression (like lagarith or huffyuv) or uncompressed .avi for best quality results which then you would feed into your NLE

To separate the audio from the camera recorded source, use dgmpgdec or tsmuxer. I assume you have the mic recorded source already separate. You can feed these audio files separately into your NLE storyboard (assuming it is able to handle multiple audio tracks). Another option would be to edit/adjust the audio before plugging it into your NLE with an audio editor (not sure how comprehensive your NLE is)

If it's what Soopa said, L+R mono tracks meant to be as channels, you can use a free audio editor like audacity to edit/adjust and export as a single .wav to plug into your NLE

Cheers

[FranklinQuinten]

I used dgmpgdec to demux audio but my NLE (Avid) does not recognize it. What format does this convert to? I can't find any conversion settings for audio format. Why not use TMPGEnc?

[poisondeathray]

Most MPEG2 transport streams will have AC3 audio. Use mediainfo (view=>text) on both the .m2ts and the demuxed track to confirm the identity. I'm still assuming it's MPEG2 video stream that you have. dgmpegdec just demuxes the audio track (same as tsmuxer in demux mode), it does no conversion.

Why are you using TMPGEnc for? To convert to uncompressed? I don't see it's purpose in this workflow? Am I missing something?

I'm not familiar with Avid's editor. If it doesn't accept the AC3 (I'm assuming that's what it is until you confirm it with mediainfo) just convert it to .wav with belight (or dozens of other free audio converters)