eac3to - audio conversion tool [Archive] - Page 5

The_Keymaker

20th June 2007, 16:40

@honai,

I have been successful at converting TrueHD files to 6 mono waves using using eac3to and my EAC3toGUI program.

I use eac3to/EAC3toGUI to convert the TrueHD track (in an EVO) to wav file. I then use tranzcode to convert the multichannel wav file to 6 mono waves and the use surcode DTS to convert the mono waves to DTS.

Once I finish fixing bugs and adding features to EAC3toGUI, I will tackle a one click utility (using madshi's eac3to, EAC3toGUI, tranzcode and surcode DTS) to do this.

By the way on my system - Krell processor, Pass Labs amps and B&W Nautilus 802 speakers - I can easily tell the difference between high rate Dolby digital (640kbps) and DTS (1536kbps), with DTS sounding better.

since I have eschewed analog out for the time being, I can not comment on the sound quaility of TrueHD vs other similar formats.

honai

20th June 2007, 16:49

@The_Keymaker

Sounds really great, thanks for your efforts!

BTW nice setup there! I'm running standard consumer HW here, but over SPDIF I can also hear the difference between full-bitrate DTS and AC3@640kbps, so TruHD->DTS seems like the only way to go.

crapz0rs

20th June 2007, 16:56

Hello, so i try to encode a dd+, Pitch.Black.HD-DVD.1080pEnglish 5.1.ec3 to test.ac3 with eac3toGui and i have that :

C:\DVD Outils HD\EAC3toGUI 0.90>eac3to.exe "C:\DVD Outils HD\EAC3toGUI 0.90\Pitc
h.Black.HD-DVD.1080pEnglish 5.1.ec3" "G:\test.ac3" -640 -nero
E-AC3, 5.1 channels, 1:51:54, 1536kbit/s, 48khz
Muxing eac3 file to raw. Please wait...
The file size of the raw file doesn't seem to fit.
The expected file size for 16 bit is 1,66 MB.
The expected file size for 24 bit is 2,49 MB.
The real file size is 0 Bytes.

C:\DVD Outils HD\EAC3toGUI 0.90>

Did i miss something, i force nero and i have Nero burning Premium in my hard disk. Did i install the bluray/Hddvd plugin in plus or not necessary !!! is there something wrong.

Thank you for your help
BinnugsaiCorrect, the bluray/hddvd is installed by default, but you need to license it separately.
If you start "Nero ProductSetup" and goto license, you'll you probably do not have a serial installed for bluray/hddvd plugin.
It's a seperate purchase from the nero website.

homerpez

20th June 2007, 23:29

HEAVILY UPDATED TO REFLECT MY PROGRESS:

@homerpez, currently eac3to only supports TrueHD inside of EVO containers. But if you can send me a little chunk of your Blu-Ray M2TS file (20MB should be enough), I may be able to add support for Blu-Ray TrueHD, too. Much depends on how the Nero filters behave. For sure you'll need the Nero HD-DVD/Blu-Ray plugin. You can try right now if things work by renaming GraphEdit.exe to "Recode.exe" and then using "File Source Async -> Nero Splitter -> Nero Audio Decoder -> Dump". Does that work?

I discovered an obscure way to decode this track, and it was similar to what you'd posted here, madishi. However, I still couldn't do anything with the True HD track or LCM streams demuxed... I found that running either the TrueHD or LPCM streams through EAC3toGUI, or tranzcode, or dtscore, would result in garbled output or no data at all.

But... just for kicks, I loaded up the .m2ts stream (instead of any demuxed stream) and connected to the Nero filters that way.

Filesource(Async) "0066.m2ts" --> Nero Splitter --> Nero Audio Decoder 2 (this must be what they call it in the Vista version of Nero Ultra)

I then connected the Nero Decoder to "Directshow device" audio renderer. I could then listen, and select which audio stream I wanted to decode using the Nero Splitter (it was stream "0" which I think is the LPCM).

Once found, I disconnected the audio renderer, saved the graph file, referenced it in an AVS file, and loaded it into BeHappy.

I encoded it into 5.1 ac3 format, and it sounds great. I think even the channels are mapped correctly. :)

Unfortunately, this method didn't use EAC3to... didn't need it... but a working solution is a working solution.

Also, I didn't need to rename Graphedit to anything else... it just simply worked as described. If it's repeatable, this seems a fairly simple way of accomplishing things.

The_Keymaker

21st June 2007, 06:48

New version (0.96) of EAC3toGUI is now available. It can be obtained from this link:

http://www.sendspace.com/file/btfl0x

Made several changes and improvements. Change log:

0.96 - Reconfigured destination file input scheme.
- Added quote support for executable path. Should fix the
program launching problem reported by Beastie Boy.
- Implemented Source file validation by extension.
- Implemented destination file validation by extension.
- Implemented option enabling by input file type.
- Added Vertical scroll to command line output.

Please report any bugs you may find.

Regards,
The_Keymaker

Binnugsai

21st June 2007, 07:10

@The_Keymaker

Thank you for all, we can't download the new version of EactiGui 0.96. Wrong link !!!!!!!

Binnugsai

21st June 2007, 07:12

Correct, the bluray/hddvd is installed by default, but you need to license it separately.
If you start "Nero ProductSetup" and goto license, you'll you probably do not have a serial installed for bluray/hddvd plugin.
It's a seperate purchase from the nero website.

Thank you crapz0rs, very nice. everything is ok now.

Best regards.

Beastie Boy

21st June 2007, 11:21

The_Keymaker, many thanks for the updated GUI. I have tested it and still ge a file not found error, but it isn't caused by the pathname. By default, EVOdemux outputs the audio file with .DD+. in the file name. When this is passed to eAC3to it appears in the console as .DD>., hence the file isn't found, nor is the output file if this name also contains DD+.
I changed the text to .DDp and all worked OK.
Hope this helps.

Cheers, Beastie.

The_Keymaker

21st June 2007, 13:30

Is anyone besides Binnugsai having problems downloading EAC3toGUI v0.96 from the link?

I tested the link and was able to download it with no problem.

The_Keymaker

The_Keymaker

21st June 2007, 13:34

@BeastieBoy

Yes I believe the "+" sign is reserved by Windows/DOS for another purpose so one should avoid using this symbol in filenames. I'm not sure I can workaround this in EAC3toGUI.

I always rename the files spit out by EVOdemux to something more conventional.

Thanks for the feedback!
The_Keymaker.

Beastie Boy

21st June 2007, 14:26

I always rename the files spit out by EVOdemux to something more conventional.

So do I, so it's not a problem :)
The link worked fine for me.

Cheers, Beastie.

Binnugsai

21st June 2007, 14:45

@The_Keymaker

Ok i the link is ok and worked fine. Thanks

So, i want to know how do we use DTSCore to convert a DTS-HD audio, in command line with Dos !!!!, witch one. Don't you think that we can perhaps using Betsweet to demux into 6 chanel wav and remux with Surcode dvd.......

Thanks

ACrowley

21st June 2007, 14:53

Are you sure this is also true of HD-DVD DTS-HD tracks? I was under the impression that the compatibility core of DTS-HD is only mandatory for Blu-ray titles.

yes im sure on HDDVD too, DTSHD is DTSHD ,its the same

I use a lot of extracted dtscores from HDDVD ( Perfume ...)

@The_Keymaker

Ok i the link is ok and worked fine. Thanks

So, i want to know how do we use DTSCore to convert a DTS-HD audio, in command line with Dos !!!!, witch one. Don't you think that we can perhaps using Betsweet to demux into 6 chanel wav and remux with Surcode dvd.......

Thanks

Why ? Makes 100% no Sense .You will lost a lot of Quality with Surcode reencode

As i say you can extract the dtscore lossless from DTS HD ! So you have the encoded 1536kbps DTS core without Quality loss from a senseless reencode

Its simple. Start your Windows command Promt and type in

dtscore.exe dtshd.dts dtsnew.dts

Thats all to get the DTS Core from DTS HD

Binnugsai

22nd June 2007, 07:09

yes im sure on HDDVD too, DTSHD is DTSHD ,its the same

I use a lot of extracted dtscores from HDDVD ( Perfume ...)

Why ? Makes 100% no Sense .You will lost a lot of Quality with Surcode reencode

As i say you can extract the dtscore lossless from DTS HD ! So you have the encoded 1536kbps DTS core without Quality loss from a senseless reencode

Its simple. Start your Windows command Promt and type in

dtscore.exe dtshd.dts dtsnew.dts

Thats all to get the DTS Core from DTS HD

@ACrowley

Thanks a lot. I saw at the end that is not necessary to use DTScore. I try with Rambo First Blood with DTSHD and after use Evodemux with mpa extension for the audio, i have just change the name in xxxxxx.dts and i can read it and use it to make my mkv movie. I read too with Powerdvd without problem. So

xkodi

23rd June 2007, 16:51

@madshi

finally, i've managed to mux correctly standart MLP files (homemade, so with known format) into EVOB container.

i've made small program, that reads the EVOB blocks from the 'V for Vendetta' TrueHD EVO file and then leaves the EVOB headers, but replaces the TrueHD data with the MLP data from my homemade MLP file and it works - 'Nero Audio Decoder 2' plays the newly created EVOB with the MLP track into it.

so i made very extended test with MLP streams with known format and it shows, that:

1) 'Nero Audio Decoder 2' always uses 24bit output, even when the source MLP stream is 16bit

2) when the source MLP stream is 16bit, 'Nero Audio Decoder 2' doesn't always decode least significant byte to zero, so it doesn't decode correctly (losslessly) when the input stream is 16bit

the source code and binary of the test mux program, plus all files using during the test and detailed descriptions are in the following archive (1.8MB):

http://rapidshare.com/files/38903695/nero_mlp_test.rar.html

so everyone of you can re-done the test and review it.

i'm awaiting to hear what you guys think about the test, but seems after all, 'Nero Audio Decoder 2' has a bug.

maybe someone of you will even extend the test with more conclusive samples.

madshi

23rd June 2007, 16:57

@xkodi, thanks very much for your test!

In the meanwhile I've finally got a reply from Nero support, asking for more details. I've send them detailed instructions on how to reproduce the problem. Hopefully they'll fix it!

madshi

23rd June 2007, 16:58

P.S: @xkodi, are the 24bit MLP files decoded correctly? I mean are they 100% identical to the source?

xkodi

23rd June 2007, 17:03

P.S: @xkodi, are the 24bit MLP files decoded correctly? I mean are they 100% identical to the source?

yes, the 24bit MLP files are decoded byte by byte identical to the source WAV file, but decoded 16bit MLP files are not identical to the source WAV file, they are outputed as 24bit, but not always the least significant byte is zero - exactly the same behaviour we noticed with TrueHD files.

p.s.fixed 'info.txt' from the archive is here:

http://rapidshare.com/files/38906969/info.txt.html

madshi

23rd June 2007, 17:10

yes, the 24bit MLP files are decoded byte by byte identical to the source WAV file
That is good news! So at least the Nero decoder is perfect for 24bit MLP and most probably also for 24bit TrueHD files!

but decoded 16bit MLP files are not identical to the source WAV file, they are outputed as 24bit, but not always the least significant byte is zero - exactly the same behaviour we noticed with TrueHD files.
If you compare the original 16bit WAV source to the 24bit MLP decoder output, do you see a pattern? Are the 2 most significant bytes always 100% correct and the least significant byte contains "random" data? Or are the 2 most significant bytes not fully correct, either? If only the least significant byte is wrong and the 2 most significant bytes are correct, I could add code to eac3to to simply zero out the least significant byte and we'd be ready to go!

Thanks!

xkodi

23rd June 2007, 17:18

If you compare the original 16bit WAV source to the 24bit MLP decoder output, do you see a pattern? Are the 2 most significant bytes always 100% correct and the least significant byte contains "random" data? Or are the 2 most significant bytes not fully correct, either? If only the least significant byte is wrong and the 2 most significant bytes are correct, I could add code to eac3to to simply zero out the least significant byte and we'd be ready to go!

Thanks!

yes, the most significant bytes of the decoded 16bit MLP are always identical to the source, so if the least significant bytes are stripped off from the decoded file, then it is identical to the source. the wrong least significant bytes usually are small numbers like 0x01, 0x03, etc.

so, yes, code to eac3to to remove the least significant byte, when the source TrueHD/MLP is 16bit, seems to be the most proper solution based on the info we now have and we can easily judge, according to the previous test with the compression ratios, which TrueHD track is 24bit and which is 16bit, but still the thing that most least significant bytes with 16bit TrueHD stream are not zero bugs me, because with 16bit MLP most least significant bytes are zero and very few of them are not zero, but my MLP test samples are small, around 100Kb, so maybe then it is normal.

madshi

23rd June 2007, 17:27

xkodi

23rd June 2007, 19:06

I've just checked your files. "164851_nero_decode" and "164851_merge.wav" are 100% identical - if I strip the least significant byte from the "164851_nero_decode" file. And I've done another check: The least significant byte is *ALWAYS* zero in "164851_nero_decode".

Now I'm confused... :confused:

Where did you find a non-zero least significant byte?

you are right, i've done so many tests and now, after i've double checked the files, i'm also confused where, when and if i really saw that or made a mistake in the hurry. that is why i made the archive, so someone else to verify my tests and conclusions, because maybe i've made mistakes, for the moment seems that the correct conclusion is that Nero always uses 24bit output and always correctly decodes MLP, so we are almost back to the very beginning with the TrueHD tracks ...

madshi

23rd June 2007, 19:12

you are right, i've done so many tests and now, after i've double checked the files, i'm also confused where, when and if i really saw that or made a mistake in the hurry. that is why i made the archive, so someone else to verify my tests and conclusions, because maybe i've made mistakes, for the moment seems that the correct conclusion is that Nero always uses 24bit output and always correctly decodes MLP, so we are almost back to the very beginning with the TrueHD tracks ...
Agreed. Seems that MLP decoding is perfect, but TrueHD decoding probably not. Too bad that you don't have a TrueHD encoder... :)

xkodi

23rd June 2007, 20:21

Agreed. Seems that MLP decoding is perfect, but TrueHD decoding probably not. Too bad that you don't have a TrueHD encoder... :)

maybe i should look if there are Blu-ray discs, that have both PCM track and TrueHD track, then i can put part of the TrueHD track into EVOB, the same way i put MLP into EVOB, decode and compare it with the PCM track, if i'm lucky the TrueHD should be created from the same PCM.

p.s. Ghost Rider Blu-ray disc seems to have both PCM and TrueHD tracks, so maybe in several weeks, it will be possible for me to do such test, sounds like good test.

madshi

23rd June 2007, 20:40

maybe i should look if there are Blu-ray discs, that have both PCM track and TrueHD track, then i can put part of the TrueHD track into EVOB, the same way i put MLP into EVOB, decode and compare it with the PCM track, if i'm lucky the TrueHD should be created from the same PCM.

p.s. Ghost Rider Blu-ray disc seems to have both PCM and TrueHD tracks, so maybe in several weeks, it will be possible for me to do such test, sounds like good test.
I think those Blu-Rays with both PCM and TrueHD have 16bit PCM and 20bit TrueHD.

Beastie Boy

24th June 2007, 18:46

I'm currently trying to go from eAC3 to multichannel WMA whilst preserving as much quality as possible, but something isn't right.
I used eAC3to to go from my ec3 file to wav. The resulting wave file would not play using MPC, it simply crashed. Maybe I need another filter?
Anyway, undeterred I used Tranzcode to produce 6 mono wave files. These files will play but the sound is garbled, just a series of squarks. I know eAC3to works OK on my machine, because the AC3 files it produces are OK.

Any ideas where I am going wrong?

Cheers, Beastie.

PS. If this problem is not related to eAC3to I'll start a new thread so as not to go off-topic.

The_Keymaker

25th June 2007, 08:08

Latest version of EAC3toGUI (version 0.97)

http://www.sendspace.com/file/zdfq3x

Changelog:
- Fixed bug that would not allow *.wav files as input.

Apologize for any inconvenience this may have caused.

The_Keymaker

ACrowley

25th June 2007, 10:06

Thx keymaker

Great work!

Maybe its better to rename "Orbitlee" to "Sonic" ?
I mean Orbitlee DTS/AC3/DD+ Source is only the Source Filter ,the Decoder is SonicCinevisionAudioDecoder4.2

So you can select Nero=NeroAudioDecoder2 or Sonic=SonicCinevisionAudioDecoder4.2

And a .raw input extension would be good

madshi

25th June 2007, 12:51

Beastie Boy

25th June 2007, 15:19

Maybe you just need to wait until I've added mono wave output support to eac3to.

I was hoping you would say that :p :thanks:

Cheers, Beastie.

ACrowley

25th June 2007, 15:40

Damn. i tried everything
I cant get anything working from a 24bit lpcm from Bluray

I tried to demux with SonicHDDemuxer, also with xport.

I use this Sox commadline and tried it as little endian or big endian, with -2 and -3
sox.exe -l -r48000 -t .raw -c 6 -3 -s e:\test.raw e:\test.wav

nothing! When i use eac3to do it autom. its the same Problem

The wave File makes only a heavy loud noise, 6 mono channels ( created with wavewizard or tranzcode) to.

The full lpcm 24bit is 7Gb big.
The encoded flac from eac3to is "6Gb" and playback fails.
Its wrong cause the flac must be around 3Gb.
And AC3/ wave output only the loud noise too.
A 24-16bit conversion create the same Result

When i try it on a 100Mb sample its the same

It happens with "all" 24bit lpcm Tracks from Bluray.
I really tried everything, checked it it again and again , with no succes

LPCM 16 bit works without Problems.

I really dont know where the Problem ? Any advice ?

I could only image its a Problem caused by demuxing.

Heres the same Problem, but no Solution so far i can see ?
http://forum.doom9.org/showthread.php?t=124080&page=2

infernix

25th June 2007, 15:42

I have a feature request.

Currently I'm running eac3to on a windows virtual machine in which it stores everything (source + raw + wav + ac3/flac) to a network share. For multiple reasons, I'd like to proceed on linux after the raw file has been dumped.

Seeing that most the tools used (all but the directsource stuff?) are available there, I'd like a simple parameter (like -n for make) that spits out the commands it would run after the directshow dump to raw, so that I can run sox and aften in linux. Thats basically everything after:
E-AC3, 5.1 channels, 1:59:02, 1536kbit/s, 48khz
Muxing eac3 file to raw. Please wait...
Audio Decoder output was 24 bit this time.

Yes, I could just dump to raw myself, then do strings eac3to.exe and figure out the right commands, but i'm lazy, and it's probably really simple to add :)

madshi

25th June 2007, 17:01

Damn. i tried everything
I cant get anything working from a 24bit lpcm from Bluray
With the SonicHDDemuxer you'll probably get little endian instead of big endian. You may need to adjust the parameters of the tools you're using. E.g. when using eac3to, add "-little" as a test.

madshi

25th June 2007, 17:03

I'd like a simple parameter (like -n for make) that spits out the commands it would run after the directshow dump to raw
Ok, you're the 2nd person asking me about this. So I guess I should implement it.

ACrowley

26th June 2007, 09:10

With the SonicHDDemuxer you'll probably get little endian instead of big endian. You may need to adjust the parameters of the tools you're using. E.g. when using eac3to, add "-little" as a test.

EDIT :
I have it! Very strange cause i tried all theye combinations before...mhh

Problem : even the 24bit lpcm was demxued with graphedit, it works only with Big Endian Switch in Sox

1 .sox.exe -B -r48000 -t .raw -c 6 -3 -s e:\test.raw e:\test.wav
2. wavewizard 0 1 2 5 3 4
3. encode to whatever

@Madshi
Ofcourse i tried all Combinations mayn Times before.

I can say :eac3to fails (for me)on these 24bit lpcm , But it works manually .
Even with Big or Little Endian switch, nothing! No working wave, no flac,no ac3. outputs is a loud noise.

eac3to e:\test.pcm e:\lpcm.wav -blu-ray -24 ( i tried with -B / -L all combinations)

Heres the sample. Its LPCM 24Bit 5.1 from BluRay "Apocalypto"

http://rapidshare.com/files/39409043/LPCM_24Bit.rar.html

kweek

26th June 2007, 16:15

First thanks for this great prog :)

I don't know if this is a bug or if I'm doing something wrong but I discovered this because some one said that you can create smaller flacs with the -8 option (default is -5 see http://flac.sourceforge.net/documentation_tools_flac.html#encoding_options).

When I use the following command
"eac3to.exe input.ddp output.flac -nero"
and I want to compress the flac some more with the command
"flac output.flac -o output.reenc.flac -f -8".

I get the following error
"output.flac: ERROR: FLAC input has STREAMINFO with unknown total samples which is not supported"

With "metaflac.exe output.flac --list" I get this.
METADATA block #0
type: 0 (STREAMINFO)
is last: false
length: 34
minimum blocksize: 4096 samples
maximum blocksize: 4096 samples
minimum framesize: 0 bytes
maximum framesize: 0 bytes
sample_rate: 48000 Hz
channels: 6
bits-per-sample: 24
total samples: 0
MD5 signature: 00000000000000000000000000000000
METADATA block #1
type: 4 (VORBIS_COMMENT)
is last: false
length: 40
vendor string: reference libFLAC 1.1.4 20070213
comments: 0
METADATA block #2
type: 1 (PADDING)
is last: true
length: 8192

Notice that streaminfo is corrupt. No minimum/maximum framesize, no total samples and no MD5 signature.

Now that someone who told me about the -8 option also gave me some pointers to fix this. Remember the sample_rate, channels and bits-per-sample.

Create a raw file from the corrupt flac.
"flac -d output.flac -o temp.raw --force-raw-format --endian=little --sign=signed"

Next create the flac from the raw with the information from metaflac (with the -8 option if you want)
"flac -8 --force-raw-format --endian=little --channels=6 --bps=24 --sample-rate=48000 --sign=signed -o "output.fixed.flac" - < temp.raw"

Now you can see that when you use metaflac on the new file it has the correct streaminfo, but I think eac3to should set the correct streaminfo to begin with :).
METADATA block #0
type: 0 (STREAMINFO)
is last: false
length: 34
minimum blocksize: 4096 samples
maximum blocksize: 4096 samples
minimum framesize: 417 bytes
maximum framesize: 42568 bytes
sample_rate: 48000 Hz
channels: 6
bits-per-sample: 24
total samples: 310497536
MD5 signature: c5fd63920d96db81ec780bacb00ca4eb
METADATA block #1
type: 4 (VORBIS_COMMENT)
is last: false
length: 40
vendor string: reference libFLAC 1.1.4 20070213
comments: 0
METADATA block #2
type: 1 (PADDING)
is last: true
length: 8192

ACrowley

26th June 2007, 17:02

kweek

26th June 2007, 19:35

-8 is maximum quality with maximum compression ,but slower encoding.
-5 should output a liottle bit bigger File with less compression and lower Quality

eac3to use -8 too, i would only use 8 for "HD Audio"

Hi, thanks for you response. But the post was more about that eac3to doesn't set the streaminfo right.

Do you mean that eac3to uses the -8 option for flac? I'm using eac3to mainly for highdef audio.

ACrowley

26th June 2007, 20:30

Hi, thanks for you response. But the post was more about that eac3to doesn't set the streaminfo right.

Do you mean that eac3to uses the -8 option for flac? I'm using eac3to mainly for highdef audio.

yes eac3to should use -8
The flac output has the same Size compared to a media coder -8 output.

################

ABout Sonic Cinevision 2 beta :

There a a few new Decoder Versions. Located in
C:\xxSonic/Cinevision/Filters

1. Video Decoder (I think based on Elecard/Moonlight) all are .ax Files and you can reg/use them all
-Sonic AVC/H264 Decoder
-Sonic VC1 Decoder
-Sonic MPEG Demuxer
-no AudioDecoder

2. The Cinemaster HD 4.3 Decoder ,all are .dll Files
-SonicHDDemxuer 4.3
-SonicCinemsterVideoDecoder 4.3
-SonicCinemsterÁudioDecoder 4.3

The .ax Filter are working very good on BluRay H264/VC1 Video.
No Artefacts on H264 PAFF etc.

More interesting for this Thread is ,that the 4.3 CinemasterAudioDecoder connects to TrueHD!
The TrueHD is playing ,but you can hear a noise in the backround. But basicly it should work now.
I Tried it on Matrix HJDDVD TrueHD 24bit

The other Thing is the Cinevision 2 has no installation fo the Cinemaster 4.2 Decoder. When i try to reg the manually it fails with a Error Message.
You can overwrite the 4.2 dlls with the new 4.3. But Avisynth wont connect anymore.
So, it would be nice if somebody can find out a workaround how to properly reg. the Cinemaster HD 4.3 Decoder

The_Keymaker

26th June 2007, 20:50

New version of EAC3toGUI (version 1.01). EDIT: Version 1.00 had a channel mapping bug! The following link is for the latest bug free version

http://www.sendspace.com/file/pkz6ew

There are extensive changes and and improvements for this latest release. Some that can be noticed are:

- Changed to TABBED GUI interface.
- Added custom channel order capability w/error checking.
- Added a channel order preview window.
- Added additional (32kbit) data rate.
- Renamed "orbitlee" filter to "Sonic-Orbitlee".
- Fixed a bug that disabled destination file input when source file changed.

Remember, this is just a GUI. Fot it to work, you must download and install Madshi's eac3to.exe program as well as the others listed in the first post of this thread. Just copy EAC3toGUI to the same directory as eac3to.exe and run (if you decide to place EAC3toGUI in a different directory, please set the location of eac3to.exe using the "Settings" menu item in EAC3toGUI).

Please feel free to post bugs, improvement suggestions or usage experience.

Regards,

xkodi

27th June 2007, 01:24

@madshi
new Sonic decoders version 4.3 are capable to decode raw MLP/TrueHD stream, so they connect to both EVOB or raw MLP/TrueHD stream.

i made the same test i made before with Nero:

24fake4851_nero_decode and 24fake4851_sonic_decode are identical

244851_nero_decode and 244851_sonic_decode are indentical

but 164851_nero_decode and 164851_sonic_decode are not, Nero output is 24bit with least significant byte zero, but Sonic properly detects, that the stream is 16bit and outputs only 16bit:

http://img72.imageshack.us/img72/4364/sonic16bitju1.png (http://imageshack.us)

and the result from the test with TrueHD:

Vendetta TrueHD stream, Sonic output:
http://img165.imageshack.us/img165/6774/sonicvfvthdgc4.png (http://imageshack.us)

Vendetta TrueHD stream, Nero output:
http://img72.imageshack.us/img72/5306/nerovfvthddl4.png (http://imageshack.us)

so Sonic decode of Vendetta TrueHD stream has least significant byte of every sample set to zero, but when listen to the decoded stream it has some kind of noise, so seems it's faulty.

Nine Inch Nails: BYIT' bonus track TrueHD, Sonic output:
http://img72.imageshack.us/img72/5986/ninsonickb4.png (http://imageshack.us)

Nine Inch Nails: BYIT' bonus track TrueHD, Nero output:
http://img72.imageshack.us/img72/4136/ninnerooo3.png (http://imageshack.us)

with NIN TrueHD track there are some different bytes, but most of the bytes are identical, from previous tests we know, that NIN TrueHD is probably 24bit, so seems that with 24 Sonic performs almost perfect, again there is some noise in the sound.

so both Nero and Sonic are perfect with MLP, Sonic is better, because has 16bit output, when MLP stream is 16bit, plus can decode raw MLP.

when it comes to TrueHD, like Nero, Sonic always reports the TrueHD stream as 24bit, but when TrueHD stream is 16bit (Vendetta) seems Sonic has faulty decode, because there is noise, when TrueHD stream is 24bit (NIN bonus track) again there is noise, but as bytes the output is almost identical to Nero.

ACrowley

27th June 2007, 08:41

xkodi

27th June 2007, 10:32

@xkodi have you properly reg the 4.3 Decoder dll , or only overwritten the 4.2 Decoder ?
Cause i cvant reg the 4.3 dll.

When i overwrite the old 4.2 ( from Oct.2006 new dll are from march 2007) i can use teh Decoder in graphedit or for playback.
But avisyth cant use them with a incorrect Pin Error Message

i overwrote the old 4.2 DLLs with the new 4.3 ones, because trying to regsvr32 them doesn't work and after that - no problems using them with GraphEdit without any kind of error messages, i don't know for Avisynth, because i haven't tried that yet, but probably it won't work and give the same error message like you said.

IS there DRC on TrueHD from Sonic? So i still prefer Nero cause i know surely no DRC will be appilied on anything

But im not sure if theres any DRC in a MLP decoded Stream

I have a noise in backround on the EVO truehh of Matrix with Sonic

at least with MLP stream the output of Sonic is byte by byte identical to the original PCM source from which the stream is created, so nothing is additionally applied, with TrueHD stream, i don't know, i also have noise and i'm still not sure what is the cause, because in hex editor the output of Sonic filter seems much more proper than Nero - every least significant byte of the Vendetta TrueHD track decode is zero as it should be, because my previous test with MLP compression ratios shows that Vendetta should be made from 16bit PCM source and with the NIN TrueHD track, which should be made from 24bit PCM source, the Sonic output is almost identical to Nero, very few bytes are different, but there is still noise

ACrowley

27th June 2007, 12:18

xkodi

27th June 2007, 12:36

Maybe ist possible to tweak the Registry Settings from Sonic Audio Decoder 4.2, to get clean TrueHD Output c

There are a lot of more Options as in System Control Panel

HKEY_LOCAL_MACHINE\SOFTWARE\Sonic\CommonMPEGDecoders\4.2\AudioDecoder

according to CineVision help:

The Audio track allows you to select an audio file for playback with the encoded video during preview. You can select the following types of audio files:

PCM (.WAV)
DTS (.dts, .cpt)
DTS-HD (all modes)
Dolby Digital (.AC3)
Dolby Digital +
Dolby Tru-HD

but i don't know how to open file for preview using CineVision and need to read more of the help, but if Tru-HD file plays fine in CineVision, there is hope, that it can be made to work correctly outside CineVision, because these filters are beta and maybe simply the TrueHD decode is broken, so first we should learn how to play TrueHD using CineVision and then if the playback is OK search for solution, how to use the filters with GraphEdit.

ACrowley

27th June 2007, 13:32

yes, youre right

With 4.3 and new HDDemuxer i can start/play a single trueHD (.mlp) in zoomplayer.
But also with the backround noise and crazy runtime (few thousand h)

Is there any chance to output SPDif as AC3 from DD+.TrueHD etc ?

I tried some registry switches and i got the Audiooutput Option " DolbyDigital" avaible in the Sonic Audio System Control Panel.
I cant rember which Registry switch it was.
Sonic should be able to output HD Audio as AC3 ,or not ? i m not sure..

However, i use AC3 Filter behind NeroAudioDecoder2 , works perfect

madshi

27th June 2007, 13:36

Hi, thanks for you response. But the post was more about that eac3to doesn't set the streaminfo right.
I'm aware of that problem - and it can't be fixed right now. Why? Because I'm intentionally calling flac.exe in a way that produces the un-initialized header info.

:eek:

I have to do that because flac.exe can't cope with >2GB output files otherwise. Try your "fix" on a flac file which is bigger than 2GB and you'll see that it won't work. Flac.exe just can't handle such big flac files (in Windows only), unless you run it in a special mode. The special mode results in the flac header not being fully initialized, unfortunately.

madshi

27th June 2007, 13:40

@xkodi & ACrowley, that's nice that Sonic's newest beta filters handle MLP and TrueHD! Too bad, though, that they are buggy with TrueHD decoding right now... :( Hopefully they'll fix that in the final version.

@xkodi, it seems that we finally have a final confirmation about that the Vendetta track is really only 16bit and the NIN one really more than 16bit. It's good that we have this information confirmed!! Now we just need a bug free TrueHD decoder, of course.

Have you guys tried "Sonic HD Demuxer -> Nero Audio Decoder" yet? Of course that's no full solution right now, either, but at least it might allow direct playback of TrueHD tracks without background noise right now? (Haven't had a chance yet to check out the beta Sonic filters myself yet).