Which output codec to choose for TMPGEnc procesing??

[An10Bill]

I'm using Studio9 to edit my DV-tape, and from Studio9 I use DV-encoder to output an AVI-file which I intend to process further into a MPEG/DVD with TMPGEnc.

Can TMPGEnc handle DV-compressed AVI-files, or do I need some kind of reader codec ?? - Or is there some other codec I should use for AVI-file output from Studio9 ??

Thanks in advance!

[gavo]

Good job by not using the build in exporting of any editing program it always second class compared to seperate encodeing programs because a editing program can't do everthing good. Yes it can but it does a very poor job. All dv codec do a poor job when it comes to decodeing this is because dv is a really hardcore enyciptioned encrypted with lots of bullshit and it is quick the old codec. I would suggest that you encode it through vitualdub at fast recomprass settings to huffy and then do filtering in Avisynth, then you can send that script over to tempgenc. For settings in tempgenc follow this guide settings here http://www.afterdawn.com/guides/archive/dvd2svcd_with_tmpgenc.cfm. Just ingrone the dvd ripping junk. This will allow you to pure onto cd's and have the best quaitly possible with the settings calcenter. If you want full support then charge the resolution to 720x480 and use a dvd authoring program like tempgenc dvd author located here http://www.pegasys-inc.com/en/product/tda.html or use the dvd template in tempgenc. Hope this helped.

[Video Dude]

TMPGEnc can handle DV just fine. Use a bitrate of at least 6,000 to encode the DV @ 720x480 (or if you live in PAL land 720x576)

Install a good DV codec like Canopus or MainConcept. They seem to give better results then the built in Microsft DV codec.



Read bb's excellent guide for DV to DVD conversion.
http://forum.doom9.org/showthread.php?s=&threadid=60392

[bb]

Originally posted by gavo
[...]All dv codec do a poor job when it comes to decodeing this is because dv is a really hardcore enyciptioned encrypted with lots of bullshit and it is quick the old codec.
DV codecs are still among the best you can get in terms of image quality and generational loss.
Originally posted by gavo
I would suggest that you encode it through vitualdub at fast recomprass settings to huffy and then do filtering in Avisynth, then you can send that script over to tempgenc.
I fail to see why an intermediate Huffyuv step could make sense.

The "DV to DVD" guide is available at Doom9's guides now, too, including screenshots:
http://www.doom9.org/dv/guide.html

bb

[An10Bill]

I'll try this.... thanks!

-
An10bill

[gavo]

The huffy idea is because when you add filters to a dv avi it adds evermore time wait. If you go to huffy then do filtering you will finsih the job in a faster time.

[bb]

Originally posted by gavo
The huffy idea is because when you add filters to a dv avi it adds evermore time wait. If you go to huffy then do filtering you will finsih the job in a faster time.
You may save time this way if you're doing a multiple pass encode (at least two passes) after the filtering. But it still depends on the filters you use, because reading the large Huffyuv files eats time, too; so there's a break-even when the Huffyuv method is faster compared to on-the-fly filtering in each pass.

bb

[Video Dude]

Originally posted by gavo
The huffy idea is because when you add filters to a dv avi it adds evermore time wait. If you go to huffy then do filtering you will finsih the job in a faster time.
Converting to Huffyuv and then filtering would take much longer then filtering the DV directly. In addition, Huffyuv is huge compared to DV. You might be having software/computer problems if you find it that DV is much slower. What software are you using to edit/filter? Are you using the Microsoft DV codec or a VFW DV codec?

If you are going from DV to MPEG or something like that, you can frameserve with AVISynth to add filters. That way you avoid the temp file and some of the best filters are found with AviSynth.

If you are editing DV for output back to your camcorder, leave it in DV and don't convert.

[bb]

The procedure where you can save time is as follows:

1. Edit the video in your NLE and export it to DV. If you prefer to avoid all possible generational loss, you may export to Huffyuv instead. This will probably take much longer, though, because there's no "smart rendering" due to the conversion DV -> Huffyuv.

2. Open the DV (or Huffyuv) file in AviSynth via AviSource or DirectShowSource. Apply all the filtering you desire, and save to a Huffyuv AVI, e.g. using VirtualDub.

3. Perform your multipass MPEG-2 encode.

The time saving happens in step 3, because you encode the already filtered file. Would you feed the AviSynth script into the encoder instead of the Huffyuv file, you'd apply the filters again and again in each encoding pass, which can be very time consuming, depending on the filter chain used. You have to take into account the extra time for saving the large Huffyuv file, though. And then there's the extra time to read in the Huffyuv file from the hdd during encoding. This alone may take longer than the AviSynth frameserving, if the filters aren't CPU intensive. If you count it all together you'll find that the Huffyuv method is not always faster.

Note that step 1 and 2 above may be combined by bridging directly from your NLE to AviSynth without an intermediate file. I believe this can be done using ffvfw or ffdshow (the versions with AviSynth support). Even if you go this way you can still decide whether you prefer to create an intermediate Huffyuv file before encoding or to directly feed into the encoder. In the latter case not only the AviSynth filtering would be applied multiple times, but the NLE edits, too. I must confess that I never tried if this works with multipass encoding at all :)

bb