Problems encoding AC3 2 AAC

[tomos]

i cant get belight to work at all at the moment, it just craps out straight away.

had some luck with behappy but it doesnt like some AC3 and just fails.

at the moment i am trying to encode an ac3 -> AAC. all starts ok in behappy but after a few seconds, i get this with directshowsource

Starting job loong.m4a
Found Audio Stream
Channels=6, BitsPerSample=16, SampleRate=48000Hz
neroAacEnc.exe -q 0.2 -ignorelength -if - -of "F:\looong.m4a"
Writing RIFF header to encoder's StdIn
Writing PCM data to encoder's StdIn
Error: System.ArgumentOutOfRangeException: Non-negative number required.

tried with nicac3source as well as directshowsource

get this error for nic ac3

Error: System.ApplicationException: Can't find audio stream!
at BeHappy.Encoder.encode()

with Belight (i tried again just now, downloaded the latest beta - BeLight 0.22 beta 9) i run the prog, drag and drop the ac3 into the input window and choose my settings. i click to encode and it goes green like its finished after 2 seconds.

the log just goes:

Using Shibatch.dll v0.25 by Naoki Shibata & DSPguru (shibatch.sourceforge.net).
Using bse_Nero7WA.dll by dimzon, Build Nov 10 2005, 14:35:28
Manual Dynamic-Compression algorithm by LigH (author of WaveBooster).

Logging start : 10/04/06 , 04:04:31.

BeSweet.exe -core( -input F:\looong.ac3 -output F:\looong.m4a -logfile F:\looong.log ) -azid( -c normal ) -dimzon( -dllname bse_Nero7WA.dll -6chnew -vbr_radio -codecquality_high -aacprofile_he ) -ota( -d 0 -g max ) -boost( /b=3 /l=0.95 )

[00:00:00:000] +------- BeSweet -----
[00:00:00:000] | Input : F:\looong.ac3
[00:00:00:000] | Output: F:\looong.m4a
[00:00:00:000] | Floating-Point Process: No
Error 84: error configuring bsn!
Quiting...
[00:00:00:000] Conversion Completed !

Logging ends : 10/04/06 , 04:04:31.


so i searched around, found out others had the same 'bsn' err. i grabbed the dll off the corecodec site but i still get the same error. i then tried the one from 'gotdotnet' (behappys site) but still get the same thing. :(

i followed advice of another thread - try the latest daily build - same :(

[tebasuna51]

had some luck with behappy but it doesnt like some AC3 and just fails.
If you have problems with only some ac3, try with this nicaudio.dll (http://forum.doom9.org/showthread.php?p=870339#post870339) version, more permissive with errors and samplerate. Replace the dll at AviSynth\Plugins and/or BeHappy folders.
with Belight (i tried again just now, downloaded the latest beta - BeLight 0.22 beta 9) i run the prog, drag and drop the ac3 into the input window and choose my settings. i click to encode and it goes green like its finished after 2 seconds.
...
so i searched around, found out others had the same 'bsn' err. i grabbed the dll off the corecodec site but i still get the same error.
With BeLight use the last bsn (http://forum.doom9.org/showthread.php?p=881619#post881619) version and Nero Digital (CLI) instead Nero7 to encode (neroAacEnc.exe at Belight folder).

[tomos]

weird, changing to the CLI version works. prob is you lose the profiles.

any idea what the numbers correspond to in profiles? .5 equals normal?

[tebasuna51]

See
Approx. average bitrate <-> Quality table (http://www.hydrogenaudio.org/forums/index.php?showtopic=44310)
For transcode ac3 (192 Kb/s stereo, 448 Kb/s 5.1) you can use between 0.3-0.4 (in my opinion)

[shon3i]

See
Approx. average bitrate <-> Quality table (http://www.hydrogenaudio.org/forums/index.php?showtopic=44310)
For transcode ac3 (192 Kb/s stereo, 448 Kb/s 5.1) you can use between 0.3-0.4 (in my opinion)
for 5.1 is enought 96kbs for solid quality without artifacts, 96-200, is better to use HE-AAC, and for other bitrates >200 is LC-AAC. Using 448kb/s is no sense here because than is better keep original AC3 track, if mp4 container not used. And of course my opinion is CT AAC, because defintly have better 5.1

[Rockaria]

@shon3i, glad you are enjoying the new CT AAC. guess there's going to be another listening test.

The linked table is a very rough(no-realistic) estimation of the average AAC bitrate that can correspond to their quality based VBR -q value.
The actual psychoauchostic model can be much different(than the simple table) and the bitrate should vary by the actual complexity of the signal.
And as discussed in the own thread, it(-q bitrate) absolutely has no relation to their CBR mode. And their CBR mode only can be resonably comparable to the CT version which is CBR. But if they suggested the table as a guideline, the applied psymodel can actually be(based on) the simple conversion table.

BTW, isn't the AAC CBR max bit rate around 320(effectively)? while the AC3 can have upto 640, of course all lossy, having artefacts by nature, more or less.

[magicclue]

i cant get belight to work at all at the moment, it just craps out straight away.

had some luck with behappy but it doesnt like some AC3 and just fails.


if your source AC3 is somehow defect try to fix bad frames with AC3Fix first:

AC3Fix (http://www.videohelp.com/tools?tool=91)

you may also try BeSliced for that purpose! It can fix defect AC3 streams as well!!!

BeSliced -> see more audio tools and voila (http://www.doom9.org/software.htm)

[tomos]

thanks all :)