eac3to - audio conversion tool

[The_Keymaker]

@bertox,

1. "Conversion options 2" tab is currently blank but will be filled shortly with other conversion options.

2. The options enabled for MLP files are the same as enabled for THd and PCM files and -big and -24 were the default options for PCM files. Just uncheck them if you desire. If other users feel these defaults are a nuisance, I will remove them as default.


The_Keymaker

[The_Keymaker]

@madshi,

I am updating EAC3toGUI an have some questions about some of the new options I am implementing:

1. For which file types are the PAL -speedup and -slowdown options applicable?
2. For which file types are the -resampleToXXXXX and -quality=X options applicable?
3. For the -skip2 option, is the option fixed at -skip2 or is it -skipX where X = the number of sequence headers?

Thanks

[nautilus7]

1. All audio conversions (all audio inputs - all audio outputs).

2. 44100 and 48000 same as above. 96000 can't combined with ac3 as output (there aren't 96khz ac3's, right?).
Note that -quality=x infects resample and slowdown/speedup options.

3. skipX where x is 1..99 sequence headers.

[rickardk]

Anyone remuxed Lucky Number Slevin?

eac3to hangs at the end (when whole window is full of "----").

H264 stream reported as 59.94fps when demuxing video with EVOdemux.

[umaximus]

Anyone remuxed Lucky Number Slevin?

eac3to hangs at the end (when whole window is full of "----").

H264 stream reported as 59.94fps when demuxing video with EVOdemux.

My copy is crashing every single version of eac3to that i tried with (2.13 to 2.20) at the same point every time (somewhere in the middle of muxing).
As madshi wrote, its crashing because of haali spliter. I think we should wait for internal AVC spliter for h264 hddvds.

[Thunderbolt8]

try only to demux the video and remux it then manually with mkvmerge and timecodes of 59.94 fps, maybe this helps.

I remember when I remuxed the iblard jikan blu-ray the video stream was present 2 times, 1 time as mpeg2 1080i with 29,97 fps and another time as AVC (only accessible when the commentary was played). but strangely I had to mux the AVC stream with 59.94 fps and not with 29.97 to have it in sync with the audio stream which as allocated in the other mpeg2 .m2ts file.

[rickardk]

try only to demux the video and remux it then manually with mkvmerge and timecodes of 59.94 fps, maybe this helps.

I remember when I remuxed the iblard jikan blu-ray the video stream was present 2 times, 1 time as mpeg2 1080i with 29,97 fps and another time as AVC (only accessible when the commentary was played). but strangely I had to mux the AVC stream with 59.94 fps and not with 29.97 to have it in sync with the audio stream which as allocated in the other mpeg2 .m2ts file.

I really think it is 23.976fps. Can't see why it would be 59.94fps. mkvmerge hangs btw. But evo plays perfect in PowerDVD.

[Snowknight26]

Seems flac only support one kind of 5 channel: Lf, Rf, C, Ls, Rs

Then only 3/2 files can be encoded. Don't work with 3/1.1 or 2/2.1 like your MLP 4.1 is.

Couldn't an extra channel (C) be added to make it 5.1, but leave the C channel without any data?

[tebasuna51]

Couldn't an extra channel (C) be added to make it 5.1, but leave the C channel without any data?

Yes, of course, but I think eac3to can't solve all particular cases.
Actually all the options are enough complicated, and there are others methods to do special cases.

[Snowknight26]

Alright, but theres still the problem of converting a 4.1 MLP to WAVs resulting in 5 channels, one of which is C, and none of which are LFE.

[bertox]

Snowknight26:
In 4.1 sources LFE channel stays the same place that C channel do.

I can do decoding 4.1 MLP to WAVs...resulting in 5 channels wavs.

What you're doing wrong??

You probe de channels with SoundForge what i say???

Do you see my older posts that i wrote to you??

[Sephiroth0000]

Madshi great work on EAC3TO and with the many new features added. Could EAC3T0 support 5.1 aac at all and if so is it anytime soon :)

:thanks:

[nautilus7]

You can use external aac/vorbis/mp3 encoders by setting eac3to to output data to stdout.

[Snowknight26]

Snowknight26:
I can do decoding 4.1 MLP to WAVs...resulting in 5 channels wavs.
Yes, I can do that too. However, as you can see from my log, eac3to outputs a C channel instead of LFE when going from MLP -> WAV.


You probe de channels with SoundForge what i say???
Not sure how that would help.

[G_M_C]

[...]
Not sure how that would help.

SF 9 is multichannel, so you can load all channels at once, so this way you can see is your supposition is right. Also it enables you to see if there is an LFE in your wav's somewhere, and what channel it is. And offcourse you can see the channel mapping in one glance, and remap using the mixer if needed.

Easy :)

But i might see why you hesitate to use it ... it's Sony software, and well ... Sony havent had a good rep around here have they ;) :D

[madshi]

Can you implement this?? Take longest time to compress in this level....and the players "stressed" to much unnecessarily.

Maybe two (or three) options only is fine...4 and 8.

4 is sufficient to me.
Same problem as always: I'd like to avoid adding too many options. Until today everybody seems to have been happy with FLAC encoding at level 8. You're the only one who wants a different level. If I add an option for every wish that is ever posted, eac3to will soon have a list of 500 options and nobody knows how to use it, anymore.

Are there more people who want to encode FLAC with a lower compression setting?

[madshi]

Does this eac3to program allow you to get around the watermark found on the majority of MLP files?
Nope.

[madshi]

Yes, of course, but I think eac3to can't solve all particular cases.
Actually all the options are enough complicated, and there are others methods to do special cases.
Thanks, exactly my thoughts.

[madshi]

I am updating EAC3toGUI an have some questions about some of the new options I am implementing [...]
nautilus7 has already answered all your questions correctly, so there's nothing more I can say except thanks for updating the GUI! :)

[Tegeril]

Having trouble with The Shining. Whether I rename the files and tell EVODemux to rebuild and take that file or I use eac3to directly from the source. I get an ac3 file (from the TrueHD track) that is 2 hours and 53 minutes (The Shining is 2 hours and 24 minutes). The added time is all silence at the beginning of the file.

Eac3to log looks like:

V:\Encoder\HD-DVD Tools\EAC3toGUI>eac3to.exe "U:\HDDVD\The Shining\SHINING_NA_V1
_0_ADVMAIN\HVDVD_TS\feature_01.EVO"+"U:\HDDVD\The Shining\SHINING_NA_V1_0_ADVMAI
N\HVDVD_TS\feature_02.EVO" 4: feature.ac3 -640
EVO, 1 video track, 5 audio tracks, 2:23:46
1: Joined EVO file
2: VC-1, 1080p24 /1.001
3: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
4: TrueHD, 5.1 channels, 48khz, dialnorm: -27dB
5: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
6: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
7: E-AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
Extracting audio track number 4...
Removing dialog normalization...
Encoding AC3...
Creating/writing file "feature.ac3"...
Video track 2 contains 206820 frames.
eac3to processing took 57 minutes, 40 seconds.
Done.

Any thoughts? :(

[bertox]

@Snowknight26

eac3to outputs a C channel instead of LFE when going from MLP -> WAV.

i don't understand you...

As long as i see the LFE channel stays in same place than Center does in a 5 channel file: FL,FR,C,SR,SL, or FL,FR,LFE,SR,SL. Both are exactly the same. Whats difference??
How eac3to change to C instead of LFE???

If the orders of channels are the same, despite of the file name, how can you listen the difference if there are no difference at all really????? I don't understand you, please be more specific.

Another thing is, perhaps, what i long saying, that the order of channels of the source file maybe are incorrect for commonly multichannel readers.

Maybe: FL,FR,SR,SL,LFE what is wrong. In this case, any player reads the SR channel like the LFE, and LFE like the SL.
The very right order is: FL,FR,LFE,SR,SL with the LFE channel in the middle. With eac3to you can re-map the wrong case to a good channels ordered file.

Maybe your case is different. I try to say that you can probe it in SoundForge to see the real order too. I saying that you may want to forget the file names, and probe the real order in SF. The file naming maybe are wrong..., but...they are only this..., only names; nothing real.

Eac3to don't change the order of channels of your soucers as their own at all. Eac3to only change the order of channels if you ask for.

.:)

[nautilus7]

Having trouble with The Shining. Whether I rename the files and tell EVODemux to rebuild and take that file or I use eac3to directly from the source. I get an ac3 file (from the TrueHD track) that is 2 hours and 53 minutes (The Shining is 2 hours and 24 minutes). The added time is all silence at the beginning of the file.

Eac3to log looks like:

V:\Encoder\HD-DVD Tools\EAC3toGUI>eac3to.exe "U:\HDDVD\The Shining\SHINING_NA_V1
_0_ADVMAIN\HVDVD_TS\feature_01.EVO"+"U:\HDDVD\The Shining\SHINING_NA_V1_0_ADVMAI
N\HVDVD_TS\feature_02.EVO" 4: feature.ac3 -640
EVO, 1 video track, 5 audio tracks, 2:23:46
1: Joined EVO file
2: VC-1, 1080p24 /1.001
3: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
4: TrueHD, 5.1 channels, 48khz, dialnorm: -27dB
5: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
6: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
7: E-AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
Extracting audio track number 4...
Removing dialog normalization...
Encoding AC3...
Creating/writing file "feature.ac3"...
Video track 2 contains 206820 frames.
eac3to processing took 57 minutes, 40 seconds.
Done.

Any thoughts? :(There must be some data in the beginning of the track that confuse eac3to. Does this happen with the thd track only or with others too? Perhaps uploading a sample will help madshi finding the problem.

As far the ac3 track is ok, you can just cut the first part (silence) by using delaycut.

[Tegeril]

I'll have to examine the sound file more closely, but if I recall correctly it seemed like it was just the entire movie pushed back. I'm encoding the video right now so we'll see if that does something odd too.

Will try using the AC3 source. Also, what is the best method for submitting a sample? (How do you typically suggest splitting it down to a functional file that is <100mb)

[nautilus7]

I forgot to ask you before... If you demux the track with evodemux and then use eac3to to make an ac3 file, does the problem also occur?

A ~50MB sample is big enough. You can cut it by using evodemux to demux the track and canceling the process or by using a hex editor or by using drmpeg's tools (http://forum.doom9.org/showthread.php?p=1078871#post1078871).

[Tegeril]

I forgot to ask you before... If you demux the track with evodemux and then use eac3to to make an ac3 file, does the problem also occur?

A ~50MB sample is big enough. You can cut it by using evodemux to demux the track and canceling the process or by using a hex editor or by using drmpeg's tools (http://forum.doom9.org/showthread.php?p=1078871#post1078871).

Well I first tried selecting only the MLP track and rebuilding an EVO only containing the MLP. It resulted in the same 2 hour 54 minute track. That's when I went to the original source files and ended with the same result. Haven't gotten a chance to try the other options yet as for AC3, I have Nero 8 installed and have to do it on another computer when I run into discs without TrueHD :(.

I will also try the demux to ddp method instead of rebuilding into an evo.

If all this fails, I'll do the evodemux and stop the process method and see if I can replicate the silence before uploading a sample file.

Thanks for all the help.

[Edit] It doesn't help that I have to rename the EVOs to get EVODemux to span them as they don't seem to conform to its naming scheme.

[Tegeril]

OH man. I fail.

I re-encoded the audio with the source files...and then tested my original ac3 file instead of the new one. The eac3to encode from both source evos worked fine.

Sorry for wasting your time :(

[nautilus7]

Never mind. I'm glad it works.

[Snowknight26]

@Snowknight26
i don't understand you...


Sorry, might be my fault. I'll try again.

When going from a 4.1 channel MLP that I know for a fact is ordered FL,FR,LFE,SL,SR to WAVs, eac3to produces FL, FR, C, SL, and SR .wav files (order is unimportant for multiple WAV files.. only the channels outputted is important). The C channel wav file could well be the LFE channel, but apparently eac3to renames it incorrectly.

[bertox]

@Snowknight26:

The C channel wav file could well be the LFE channel, but apparently eac3to renames it incorrectly.
OK. Sorry, if there are problems with file naming only, seems that i can't help you.
But you don't suffer to much if the problem is file naming only...;)

Greetngs.

[Chumbo]

@madshi,
I finally got around to doing this again with Fantastic 4 Silver Surfer with 2.20. Just wanted to post the results for you if you're interested:
eac3to "hvdvd_ts\HF_0.EVO"+"hvdvd_ts\HF_1.EVO" 2: ff4-2.mkv
EVO, 1 video track, 5 audio tracks, 1:31:49
1: Joined EVO file
2: VC-1, 1080p24 /1.001
3: DTS Hi-Res, 5.1 channels, 16 bits, 2082kbit/s, 48khz, dialnorm: -4dB
4: DTS Hi-Res, 5.1 channels, 16 bits, 2082kbit/s, 48khz, dialnorm: -4dB
5: AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
6: AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
7: AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
Extracting primary video track...
Muxing video to Matroska...
Video has a gap of 4576725 frames at playtime 0:06:12.
Video has a gap of 1144178 frames at playtime 53:07:40.
Video has a gap of 1144181 frames at playtime 66:23:24.
Video has a gap of 1144180 frames at playtime 79:43:28.
Video has a gap of 1144181 frames at playtime 93:02:33.
Video has a gap of 1144180 frames at playtime 106:24:15.
Video has a gap of 1144180 frames at playtime 119:39:37.
Video has a gap of 1144181 frames at playtime 132:55:00.
Video has a gap of 1144180 frames at playtime 146:10:23.
Video has a gap of 1144181 frames at playtime 159:28:07.
Video has a gap of 1144180 frames at playtime 172:43:30.
Video has a gap of 1144180 frames at playtime 185:58:52.
Video has a gap of 1144181 frames at playtime 199:17:39.
Video has a gap of 2288362 frames at playtime 212:33:29.
Video has a gap of 1144179 frames at playtime 239:05:40.
Video has a gap of 1144181 frames at playtime 252:27:12.
Video has a gap of 1144180 frames at playtime 266:01:18.
Video has a gap of 1144180 frames at playtime 279:20:43.
Video has a gap of 1144181 frames at playtime 292:36:27.
Video has a gap of 1144180 frames at playtime 305:53:22.
Video has a gap of 1144181 frames at playtime 319:08:44.
Video has a gap of 1144180 frames at playtime 332:32:28.
Video has a gap of 1144181 frames at playtime 346:00:04.
Video has a gap of 1144180 frames at playtime 359:15:56.
Video has a gap of 2288362 frames at playtime 372:32:08.
Video has a gap of 1144179 frames at playtime 399:05:26.
Added fps value to MKV header.
Video track 2 contains 132086 frames.
eac3to processing took 3 hours, 42 minutes.

[EDIT] I meant to add this too, when I played the stream in the MKV via WMVideo Decoder DMO, it played the beginning perfectly. All the green garbage and rainbow stuff is gone. Amazing stuff man. I haven't seen the whole thing, but I have no doubt that it'll be fine. I'm not sure if those vide gaps are something to worry about or not.

[phineasbg]

I apologize in advance if this question has been asked:

I have a unique situation. I'm converting a commercial PAL DVD 5.1 ac3 to NTSC 5.1 ac3. Eac3to does this perfectly, without having to strip the channels to WAVs, etc.

The problem I have with this DVD, is that the center channel needs to have its volume cranked down about 10db. The rest of the channels are fine. Is there a way I can do this with eac3to?

[nautilus7]

No, you can't change volume with eac3to. Why do you want to do that?

[phineasbg]

No, you can't change volume with eac3to. Why do you want to do that?

The center channel was cranked too high on the dvd itself. It was poorly mixed.

I guess I'll do my old method of exporting them to 6 WAVs, then cranking the center channel down with Audition, then recombining them to ac3 with Besweet. The only problem with this, is that they never seem to realign just right.

Thanks anyway!

[Thunderbolt8]

a question regarding the stereo .pcm track from the queen live blu-ray: eac3to couldt detect the specs and told me to enter depths, channels and endian manually. so my command line was
"eac3to G:\input.pcm G:\input.flac -blu-ray -24 -2 -big"

so was it correct to assume that this track has a big endian? I dont know how to find this out. and is the -blu-ray switch still needed here when I only have 2 channels and already specified the rest manually?

[nautilus7]

I can only say that a wrong endian is easily identified because you will get only noise to the output file.

Can't load the tracks (input/output) to an audio editor to figure whether the channel mapping is correct?

[Thunderbolt8]

i played the track with winamp and it sounds good, theres no garbage noise. and left & right channel are also working well, just wanted to confirm if that -blu-ray switch is still needed.

[Thunderbolt8]

hm that tracks works when I play it in winamp for example, but when I remux it into .mkv as secondary audio track together with the flac track from the 5.1 dtshd, then theres only this garbage sound :S whys that? Ive also had that with other movies, in which a 2.0 pcm track + other lossless audio was included. the framerate of the concert here goes up to ~30-60 fps when I switch over to the pcm flac track.

the problem still exists when I put the track into .mka outside the .mkv and then try to select it in mpc
but it doesnt exist when this flac track from pcm is the only audio muxed together with the video, when the other flac track is left out :S

as soon as there are 2 tracks inside the mkv the 2nd track causes trouble, when selected :/

[madshi]

eac3to "hvdvd_ts\HF_0.EVO"+"hvdvd_ts\HF_1.EVO" 2: ff4-2.mkv
EVO, 1 video track, 5 audio tracks, 1:31:49
1: Joined EVO file
2: VC-1, 1080p24 /1.001
3: DTS Hi-Res, 5.1 channels, 16 bits, 2082kbit/s, 48khz, dialnorm: -4dB
4: DTS Hi-Res, 5.1 channels, 16 bits, 2082kbit/s, 48khz, dialnorm: -4dB
5: AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
6: AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
7: AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
Extracting primary video track...
Muxing video to Matroska...
Video has a gap of 4576725 frames at playtime 0:06:12.
Video has a gap of 1144178 frames at playtime 53:07:40.
[...]
Video has a gap of 1144179 frames at playtime 399:05:26.
Added fps value to MKV header.
Video track 2 contains 132086 frames.
eac3to processing took 3 hours, 42 minutes.
[...] I'm not sure if those vide gaps are something to worry about or not.
That gaps are going to be problematic. Can you make a sample for me which covers one of those big gaps? Thanks...

[madshi]

hm that tracks works when I play it in winamp for example, but when I remux it into .mkv as secondary audio track together with the flac track from the 5.1 dtshd, then theres only this garbage sound :S whys that?
Which FLAC decoder are you using? Try both ffdshow and madFlac. Does the problem occur with both decoders?

The "blu-ray" switch only shows effect with multi channel tracks. So it's not needed for stereo tracks.

[Joniii]

I'm trying to rip The Hunt for Red October HD DVD to MKV. Strange thing about this disc is that video is MPEG-4 AVC and being HD DVD it has framerate of 29.97. How would I change the pulldown flag of a AVC stream?

[rickardk]

I'm trying to rip The Hunt for Red October HD DVD to MKV. Strange thing about this disc is that video is MPEG-4 AVC and being HD DVD it has framerate of 29.97. How would I change the pulldown flag of a AVC stream?

All AVC HD DVDs are seen as 29.97fps.
I've been looking for a way to do this myself. Not found one yet...

[hristoff2]

All AVC HD DVDs are seen as 29.97fps.
I've been looking for a way to do this myself. Not found one yet...

Remux with mkvmerge using 24000/1001 timecodes or (when encoding) just add Assumefps(23.976024)

[B4tm4n]

I'm trying to rip The Hunt for Red October HD DVD to MKV. Strange thing about this disc is that video is MPEG-4 AVC and being HD DVD it has framerate of 29.97. How would I change the pulldown flag of a AVC stream?

The latest H264info, I think, can remove or add pulldown to avc streams.

[jimmie]

lo there...

i think i am one of many people who dont know how to get eac3to working with nero... as i know it wont work with nero 8 but....

what version will work with eac3to? i tried many codec packs for nero but none worked... anyone able to help?

tia,

jimmie

[Nikos]

I had no problems with The Hunt for Red October HD DVD muxing into an mkv and extract the audio to 6 mono wavs.

[nautilus7]

Nero 7 (use one of the latest) & Nero Blu-ray/HD DVD plug-in.

[madshi]

eac3to v2.21 released

http://madshi.net/eac3to.zip

* latest libav MLP/TrueHD decoder fixes "lossless check failed" bug
* latest libav MLP/TrueHD decoder fully decodes 7.1 TrueHD tracks now
* Matroska muxing speed dramatically improved
* eac3to now detects and handles E-AC3 7.1 tracks correctly
* option "-core" extracts 5.1 core from E-AC3 7.1 tracks
* added support for small DTS files (< 300kb)
MKV muxing speed test:

GraphEdit (Haali): 810 seconds
eac3to v2.20: 3065 seconds
eac3to v2.21: 740 seconds

[nautilus7]

Thanks!

2 questions:
a) Is there a real core in 7.1 E-AC3 tracks?
b) How did you manage to speed up mkv muxing so much? Does this apply to all video formats? Do you still use mkvtoolnix?

[mommyman]

* latest libav MLP/TrueHD decoder fixes "lossless check failed" bug

Madshi, which one is the latest? http://www.db-instable.org/misc/truehd-decoder-v2.patch (from 16-jan-2008)

[madshi]

a) Is there a real core in 7.1 E-AC3 tracks?
"Normal" AC3 and E-AC3 tracks have one (E-)AC3 frame per audio frame. 7.1 E-AC3 tracks have two E-AC3 frames per audio frame. The first E-AC3 frame is a conventional 5.1 E-AC3 core. The 2nd E-AC3 frame is an E-AC3 frame with 4 channels in it. Two of those channels replace the surround channels of the 5.1 core. The other two channels are the back surround channels. The surround channels of the 5.1 core are totally dropped when a decoder decodes the 7.1 extension. The surround channels of the core contain a downmix of the surround and back surround channels.

The Nero and Sonic E-AC3 decoders can handle 7.1 E-AC3 tracks, but they ignore the 7.1 extension. The current version of the libav/ffmpeg E-AC3 decoder complains and stops decoding if you feed it 7.1 E-AC3. You can feed it the 5.1 core, though, then it decodes the track just fine (but only in 5.1, of course). eac3to v2.21 automatically extracts the core if you use the "-libav" switch for 7.1 E-AC3 decoding.

This "two E-AC3 frames per audio frame" logic means that delaycut 1.3 doesn't handle E-AC3 tracks correctly. delaycut doesn't make any difference between core and extension. As a result the audio tracks may be messed up by delaycut. eac3to v2.21 should be able to delay 7.1 E-AC3 tracks just fine, though.

b) How did you manage to speed up mkv muxing so much? Does this apply to all video formats? Do you still use mkvtoolnix?
You should ask how I managed to slow muxing down so much in the older versions... :o The video parsers had a bad loop design which slowed everything down a lot. I just did a tiny bit of optimization which fixed the problem. VC-1 and h264 handling is about 4x faster, I think. With MPEG2 the difference was smaller in my test. I still use the Haali Media Splitter followed by mkvtoolnix timestamp rewriting for h264 and MPEG2 content.