eac3to - audio conversion tool

[Joniii]

Damn, it doesn't work properly. Sound is perfect exept sounds keep jumping from one speaker to another. Like I hear dialog from center speaker then it jumps to right speaker. Any ideas of whats causing this?

[nautilus7]

Is this with 300 or Dracula? Do the channels swap position from time to time or all the time? How many m2ts files is the man movie? How did you merge them to one (if more than one)?

[Joniii]

Originally Posted by nautilus7

Is this with 300 or Dracula? Do the channels swap position from time to time or all the time? How many m2ts files is the man movie? How did you merge them to one (if more than one)?

Both. They swap all the time. I merged those with "copy /b 0001.m2ts+0002.m2ts out.m2ts". I'll do everything from the start again and see how it goes.

[nautilus7]

It's really confusing. Are you sure it's not anything wrong in your setup/playback application? Would you like to upload a sample? If you do, make sure it demonstrates the problem.

[sparknburn]

This latest version takes care of any probs I had!

[nautilus7]

madshi, i noticed this...

eac3to shinobi.eac3 shinobi.ac3

E-AC3, 5.1 channels, 1:42:01, 2046kbit/s, 48khz, dialnorm: -27dB
Removing dialog normalization...
Decoding with DirectShow (Nero Audio Decoder 2)...
Disabling DRC for Nero (E-)AC3 decoding...
DirectShow reports 5.1 channels, 24 bits, 48khz
Encoding AC3...
Creating/writing file "shinobi.ac3"...
eac3to processing took 4 minutes, 13 seconds.
Done.

eac3to shinobi.eac3 shinobi.ac3 -libav

E-AC3, 5.1 channels, 1:42:01, 2046kbit/s, 48khz, dialnorm: -27dB
Removing dialog normalization...
Remapping channels...
Loading white noise (needed for dithering)...
Encoding AC3...
Creating/writing file "shinobi.ac3"...
eac3to processing took 2 minutes, 27 seconds.
Done.

1. Is libav faster than nero (i did the encoding twice to be sure)? Aften is latest build (SSE3) and i have a Core 2 Duo.

2. What is the dithering needed?

3. Why it says remapping channels where it actually no remap is talking place? See the images for that:

Original E-AC-3 track--Nero decoded AC-3 track--Libav decoded AC-3 track

[Richum]

eac3to reported;

eac3to F:\FEATURE_1.EVO+F:\FEATURE_2.EVO
EVO, 1 video track, 2 audio tracks, 2:05:42
1: Joined EVO file
2: VC-1, 1080p24 /1.001
3: RAW/PCM, 2.0 channels, 16 bits, 48khz
4: DTS Master Audio, 5.1 channels, 24 bits, 48khz, dialnorm: -4dB



I ran the following;

eac3to F:\FEATURE_1.EVO+F:\FEATURE_2.EVO 2: D:\MKV\cream1.mkv 4: D:\MKV\cream1.dts

Video is good, Audio plays in GraphEdit, however when I try run MKVmerge I get the message "No DTS headers file may not play" or words to that effect. Once muxed to a MKV, there is no sound.

DTS track is listed as DTS Master Audio by eac3to. Is there some option I need to add? Is it even possible to do it this way?

[2themax]

madshi,

The VC1 stipPulldown function doesn't seem to be working with Scenarist BD. Below is the error thrown during import.

Error Content :A Value of FRAMERATENR and FRAMERATEDR or INTERLACE is not supported in the BD standard

[bendurdunthat]

First of all, madshi-you are incredible-I have been following this post since late August-page 30-along with about 8 other threads on this site and several others. Eac3to is an awesome program and your continuing dedication and patience is certainly commendable.

I keep an updated word file as to the various issues-unfortunately it is almost as long as this thread-I've started to index it-will share it when I get it done-Frequently asked Questions-might help you avoid replying to questions that have been asked before?

Anyway, I now demux with eac3to rather than Evodemux and don't have problems running the program either with the commands or whenever the GUI is updated. I have all programs installed but have not purchased Surcode as of yet.

I have completed many encodes using various encoders-Windows Media Encoder, Windows Media Stream Editor, Zambelli's program, Expression Encoder without 5.1 audio, TmpEnc Xpress 4 etc. the video is generally quite good-the audio is the problem-but the audio is fine until its encoded-then its usually out of sync.

I'm like Richum-want the best quality video & audio to stream but I don't have the HD library that he has-at least not yet-have over 400 standard DVD's & 20 HD DVD's & 2 BluRay.

To date I have not kept 1 final encode-it is kind of like searching for the holy grail-I have learned alot thanks to you, nautilus7, ACrowley & too many others to mention. Why bother? Because I can!!

It seems obvious to me that the majority of the problems Richum, myself & others are having relate to the encoding process-audio & video. I have followed at least 6 different tutorials-most of them use your program for audio with varied success.
The point I'm trying to make is that if we are all using different programs, filters, codecs etc. to encode-is it no wonder that we are experiencing problems? Even to list problem titles is not fruitful if we are all using different methods/programs etc.

I know that this thread is for eac3to audio development only-but I for one feel I'm spinning my wheels trying to find that one encoding method that takes your eac3to audio and syncs it perfectly with the video stream (from eac3to).

If we aren't using similar programs/methods to complete the final encode of audio and video won't we just continue to run into the problems like Richum? If others reading this thread don't even know or use Richum's final encoding methods-we probably would never be able to help him out ie.

Stuttering problems likely have nothing to do with eac3to. Also, for example, madshi, at one point earlier, I think that you even indicated that you were not encoding audio and video together at all? Are you still doing that-if you are what precisely are you doing-would solve my final encoding problems if it is feasible at all to keep the audio & video streams separate.

Even if different groups of us were to use certain methods/programs to encode we would be able to determine success and failures and at the same time develop this program to its fullest potential-an all-in-one encoding process & develop a list of problem titles etc. using a particular final encoding method.

Perhaps a separate thread should be started that outlines the most effective programs after using eac3to demux evo (video & audio streams), audio for BluRay streams using eac3to, programs to demux BluRay. If enough of us or groups of us used similar encoding methods then at least you would know if its a problem with the eac3to or a filter or something else.

Sorry for the rant-just a little frustrated-but not with eac3to-just everything after it-brought another computer home tonight-just to do encoding-my 5th one-including my laptop.

bendurdunthat

[K-rnivoro]

Originally Posted by bendurdunthat

Sorry for the rant-just a little frustrated-but not with eac3to-just everything after it-brought another computer home tonight-just to do encoding-my 5th one-including my laptop.

bendurdunthat

...and I thought I was crazy just because I planned to do overnight encoding with my laptop as third encoding computer at home.

Off topic: My 3GHz Core 2 Duo encodes at less than 2 FPS in x264 2nd. pass at "HQ-Lowest" MeGUI Video Profile...

[Joniii]

Originally Posted by nautilus7

It's really confusing. Are you sure it's not anything wrong in your setup/playback application? Would you like to upload a sample? If you do, make sure it demonstrates the problem.

Yeah I'm sure that the problem isn't on my setup, other DTS files play ok. This has to do something with how I demuxed that LPCM, not eac3to problem.

[Joniii]

I think I finally got it to work alternative way.

1)Open .m2ts in gdsmux.exe (haali media splitter dir)
2) Select PCM track and save, it will save it as .mkv
3) Open .mkv with mkvextractgui.exe, extract audio, it extracts it as 5.1 wav (dunno if gdsmux.exe converts it to wav or mkvextractgui?), anyway I end up with 5.1 16bit 48kHz .wav.
4) eac3to audio.wav audio.pcm
5) eac3to audio.pcm audio.dts (eac3to detects it as blu-ray 5.1/16bit/48kHz, swaps endian, remaps channels.)

Now only thing i'm wondering is:

1) Is WAV <-> PCM conversion lossless?
2) Do I need to convert WAV to PCM or just straight from WAV to DTS (I do step 4 so that eac3to would swap endian and remap channels)

[Geleodor]

3) Open .mkv with mkvextractgui.exe, extract audio, it extracts it as 5.1 wav (dunno if gdsmux.exe converts it to wav or mkvextractgui?), anyway I end up with 5.1 16bit 48kHz .wav.

It should be PCM track and you dont need to do this -- 4) eac3to audio.wav audio.pcm

[Joniii]

Originally Posted by Geleodor

It should be PCM track and you dont need to do this -- 4) eac3to audio.wav audio.pcm

eac3to reports it as wav, even if I change file extension to pcm.

[madshi]

Originally Posted by Chumbo

I may be wrong, but I believe, with PCM, you have to specify the resolution and correct Endian. If either is wrong, you'll get the screeching result.

Run "eac3to dracula.lpcm" to get the info back on it and see what it says is the bits and Endian setting. Then feed those manually.

eac3to automatically tries to find out the correct parameters and also automatically uses those parameters. You only need to manually enter the parameters if eac3to guesses wrong.

[madshi]

Originally Posted by umaximus

madshi, is it safe to say that v2.20 is 'as good as it gets' for VC-1 muxing, so i can re-do my VC-1 titles, and finally get rid of the sources from HDD? Or do you have some new ideas for improving of VC-1 muxing?

At this point in time I think that eac3to's VC-1 EVO -> MKV muxing is as good as it can possible be. I currently don't see any way to further improve it. I may continue on improving the gap/overlap handling. But as long as eac3to doesn't post any "found a gap/overlap of xxx frames" warnings, your MKV should be pretty perfect for VC-1 EVOs.

[madshi]

Originally Posted by Joniii

Damn, it doesn't work properly. Sound is perfect exept sounds keep jumping from one speaker to another. Like I hear dialog from center speaker then it jumps to right speaker. Any ideas of whats causing this?

This is probably caused by the process of joining the m2ts files and then demuxing sound. Make sure you are using the latest xport version. Maybe that solves the problem. IIRC Ron has improved xport so that it makes sure that this channel changing doesn't happen if you demux the m2ts files separately. You can afterwards use "eac3to demux1.pcm+demux2.pcm+..." to join the various demuxed PCM parts.

Originally Posted by Joniii

I think I finally got it to work alternative way.

1)Open .m2ts in gdsmux.exe (haali media splitter dir)
2) Select PCM track and save, it will save it as .mkv
3) Open .mkv with mkvextractgui.exe, extract audio, it extracts it as 5.1 wav (dunno if gdsmux.exe converts it to wav or mkvextractgui?), anyway I end up with 5.1 16bit 48kHz .wav.
4) eac3to audio.wav audio.pcm
5) eac3to audio.pcm audio.dts (eac3to detects it as blu-ray 5.1/16bit/48kHz, swaps endian, remaps channels.)

Using gdsmux is similar to using h264tsto. h264tsto will demux directly to PCM. However, you then need to use the eac3to "-blu-ray" switch, or else the channel order will be incorrect.

Originally Posted by Joniii

1) Is WAV <-> PCM conversion lossless?
2) Do I need to convert WAV to PCM or just straight from WAV to DTS (I do step 4 so that eac3to would swap endian and remap channels)

Instead of steps 4) and 5) you can use "eac3to audio.wav audio.dts -blu-ray". WAV is the same as little endian PCM, just with a header in front of the PCM data.

[madshi]

Originally Posted by nautilus7

1. Is libav faster than nero (i did the encoding twice to be sure)? Aften is latest build (SSE3) and i have a Core 2 Duo.

Don't know, haven't tested it myself, but it's quite possible.

Originally Posted by nautilus7

2. What is the dithering needed?

I've hacked libav to always output floating point samples. eac3to takes these and converts them to 24bit PCM by using TPDF dithering. That's the proper way to convert floating point samples to PCM. Unfortunately libav is internally not doing it correctly. That's why I hacked libav to output floating point so that eac3to can do the right thing.

Originally Posted by nautilus7

3. Why it says remapping channels where it actually no remap is talking place?

Remapping is taking place. libav outputs the channels in a different order than you'd expect.

[madshi]

Originally Posted by Richum

4: DTS Master Audio, 5.1 channels, 24 bits, 48khz, dialnorm: -4dB

[...]

when I try run MKVmerge I get the message "No DTS headers file may not play" or words to that effect. Once muxed to a MKV, there is no sound.

mkvmerge doesn't support DTS-HD yet. You can ask eac3to to extract the DTS core for you. That should work. Or alternatively you can convert DTS-HD Master Audio to FLAC (you'll need the Sonic audio decoder for that, though).

[madshi]

Originally Posted by 2themax

The VC1 stipPulldown function doesn't seem to be working with Scenarist BD. Below is the error thrown during import.

Error Content :A Value of FRAMERATENR and FRAMERATEDR or INTERLACE is not supported in the BD standard

Blu-Ray doesn't support HD DVD style VC-1 streams. It also doesn't support the streams created by the "-stripPulldown" option because this option only removes some parts of the pulldown, but not everything. It would be theoretically possible to remove the full pulldown stuff from the VC-1 stream but that would be really difficult to do. Microsoft has written a tool for that purpose, but I don't have access to it. The "-stripPulldown" option is made for people who want to play HD DVD VC-1 movies through their Xbox 360. It's not meant to be used for Blu-Ray authoring.