Why not TMPGEnc audio

[tatty]

Apologies if this is answered elsewhere but, although I frequently see it written - and apparently accepted as Law - that 'no one encodes their audio with TMPGEnc', it's not obvious to a novice like me why this is so.

I have been rattling out my VHS-DVD conversions for some time now, without too many problems. Capture with VirtualDub, code with TMPGEnc and author with TDA. I haven't succeeded in producing a 'coaster' yet (obviously not living close enough to the hairy edge) and results are perfectly adequate for my purpose. However, still haunted by the prospect of making some fundamental audio blunder that will catch up with me later.

At the moment I just roll the avi through TMPGEnc (vid + aud) and take what comes. I convert the wav (always ludicously large) to ac3 with ffmpggui before authoring. How else would I do it, and what is the problem with TMP audio?

[teegee420]

I've never encountered any problems myself. I even have toolame but I don't see(or hear) a reason to use it.

[FOO]

I think it's an Urban Video Legend.
Actually I think it used to be broke , but they fixed it.
You can always tell it to use lame instead

[daamon]

Hi tatty,

I've read that, way back when, TMPGEnc's MP2 encoder wasn't the best. But, I believe that either that's been sorted or using toolame gets around it.

Just something I think I've seen...

[ZippyP.]

My NTSC DVD player doesn't like MP2 audio so I have to encode to AC3 anyway.

[adam]

The main problem with TMPGenc's audio encoding was its frequency rate conversion, ie: 44100kHz->48kHz. Under certain cirumstances when converting frequencies you would get a very noticable tinny effect to the audio. It sounded like your speaker was inside a tin can. This was definitely no urban legend, but the problem might have been solved in later versions. Personally, I find that toolame is just such an exceptional audio encoder, and it can easily be used as an internal plugin from within TMPGEnc, that there would be no reason to use TMPGenc's audio encoder unless it was equal or better in quality. I don't think its significantly worse than toolame, but I don't think its quite as good so I personally always use toolame from within TMPGenc.

As far as converting frequency, again you've got ssrc which is very high quality, is at least slightly faster than TMPGenc, and can easily be used a a plugin from within TMPGenc as well. So unless TMPGenc can be faster or higher quality, I don't see the point in using it.

[duhmez]

I too use toolame wit tmpgenc, just because people say thats what i should do cause "Tmpgenc audio conversion sux"

[vcdforme]

I concur with the CyberJanitor.
It's in the sample rate conversion. 48k to 44k, or vice versa.
Try converting a DV/AVI file into a VCD and you'll see.
Or convert a VCD file into a DVD compliant file.

[daamon]

@ adam,

I'm in PAL land and, because of this supposed problem with TMPGEnc audio failings, I go to the trouble of extracting the WAV from my DV AVI projects and converting it to AC3 using ffmpeggui.

Life would be much simpler if I could just do both audio and video with TMPGEnc - so, how do I get TMPGEnc to use the toolame and ssrc plugins?

Cheers.

[adam]

Well that's another issue entirely. TMPGenc is using direct show file reader to access your compressed audio. This is horribly unstable. Direct show is meant to play sources, not to pass them on to an encoder. It is necessary, or at least a good idea, for you to decompress your audio first so that TMPGenc accesses it via its wave file reader instead. This is more stable and prevents the need to decompress and then recompress all in one step. This prevents the snap, crackle, pops that you often get using compressed audio in TMPGenc. (many have noticed that these problem occur more often when converting frequency as well. If you aren't doing this you may be willing to take your chances, but not me.)

Anyway, I do recommend using both toolame and ssrc as internal plugins. At the very least I think they are both faster, and I do trust the quality more as well. So, on the main screen, click on option/environmental setting/ and then either audio engine or external plug-ins depending on what version you are using. You should see options for mpeg1 audio layer II, Audio layer III, and for sampling frequency converter. Set it to use an external program and point it to the relevent .exe. So for layer II, I point it to toolame.exe. For layer III, I point it to lame.exe., and for frequency conversion I point it to ssrc.exe.

Have fun

[Brent212]

I have a related question -

I had an mpg file that was dvd-compliant (captured with hauppauge 250 in DVD Standard setting), but the sound was too loud. So I demuxed the file with tmpgenc to get a .mpv and a .mp2 file. Now I want to lower the volume of the .mp2, but can't figure out the best way to do it. I can open it up in a program like audacity, but after altering the volumen level, I have to re-encode it with lame, and then I can't re-mux the resulting .mp3 with the original .mpv. So I'm thinking that I should save it as a .wav from audacity and then somehow get it muxed back with the .mpv file.

So my question is this: What would be the best way to get a wav file to a dvd-compliant format, and then muxed with a dvd-compliant .mpv? The main thing I don't want to do is have to re-encode the video before burning it to a dvd. Can TMPGEnc do this, and if so, how?

It'd be nice to be able to change the volume of the .mp2 without re-encoding, but I don't think it's possible, except maybe with besplit, but I can't find any info on using besplit to lower the volume of a .mp2. The closest thing I can find to that is using it to normalize a .mp2 (using "-ota( -G max)" ), which doesn't help.

Thanks for any suggestions.

[daamon]

@ adam,

Cheers for the info - I have (after a bit of trawling to find the executables rather than the source code - Doom9.org came through...) got TMPGEnc set up as you said.

I have to confess that I'm a little mystified by your opening paragraph. I'm not sure if I need to be doing (or not doing) something...

I shoot my own footage (holidays etc.) on a DV cam, transfer to DV AVI using Adobe Premiere 6.0, edit with the same tool, then frameserve the video only to TMPGEnc. Separately, I export the audio to WAV @ 48KHz and convert to AC3 with ffmpeggui. I then bring the video.m2v and the audio.ac3 together in TMPGEnc DVD Author.

I was hoping that I could frameserve both audio and video to TMPGEnc from Premiere now that TMPGEnc is set up with toolame and ssrc (though I won't need the latter). Is this the case? I know I'll get MP2 rather than AC3, but that's OK as I'm in PAL land as long as there's no real discernible difference.

Ta, Daamon.

[adam]

I'm pretty sure the problem lies with direct show file reader, and I'm pretty sure it won't be used when frameserving DV, so I'm pretty sure you can do what you are asking.

For 2 channel audio I'd say no, you aren't going to hear any difference between mp2 and AC3.

[daamon]

Cheers adam. That'll make life a bit easier...

[Poppa_Meth]

I've use the built in encoder for a long time with no problems, but I go primarily from DivX to DVD with it and I always extract audio in Virtual Dub, and thus set my final audio frequencies at this time so TMPGEnc never has to deal with it.