All my questions about audio (BeSweet, AAC, boost, DRC etc).

[har-vas]

Hello all. I have some questions regarding audio in general, and I thought to assemble them in a single thread. First of all, I do a lot of TV captures with VirtualDub, using my Cinergy TV 600 PCI card. Then I use the same program to extract the audio as wav (PCM, 16 bit, stereo, 32 KHz).

1. When I am trying to open the .wav file in HeadAC3he (0.24 a13), I am receiving this error: "Could not find data chunk". What does it mean? Is there any workaround?

2. I also tried to encode the .wav file with aoTuV encoder (venc.exe) and I received the following message: "This wav file is 0bit". What is happening with my file? I have to mention that the file is perfectly playable and I can encode it normally to Vorbis with GordianKnot (BeSweet 1.5 b31).

3. Another problem I have is that I can't increase the volume level of such .wav files. I am checking the boost checkbox in GordianKnot audio tab, but I can't notice any difference in the volume level of the generated vorbis. I also tried to change the DRC from normal to heavy, but again I didn't "hear" any difference.
To be honest I don't understand exactly those settings, but I think that are volume-related. (Can someone explain me their functionality with few simple words?) So how can I increase the volume level in my .wav files? I suspect that these settings (boost, DRC) are applied only to .ac3 files or that my .wav file is not spec-compliant... Thanks.

[SallyDog]

I don't have much to offer, but since nobody else is jumping in, let me ask you one question. What are the audio settings you're using in Virtualdub (when you are capturing)?

[tebasuna51]

1-2)Something in your wav file is wrong. If you can send the first 100 bytes maybe we can know the problem.

3) The DRC is a parameter to use decoding ac3, without effect with wav. Boost is used to amplify the low volume without modify high volume but can produce distort. There are other parameters, maybe Normalize or Maximize (I don't know Gordian), to produce the max gain without distort.

[har-vas]

Hello all and thank you for your answers.

@SallyDog: When I capture with VirtualDub I select "No compression" for the audio. The resulting specs are PCM, 16 bit, stereo, 32 KHz.
@tebasuna51: Well, I believe that indeed something may be wrong (I prefer the words "not spec-compliant"). I cannot explain differently the error messages of two transcoding programs. But don't forget that I am not talking about a single .wav file but for every .wav file which comes from TV captures. Keep in mind that these files are perfectly playable and also that BeSweet can handle them.

My opinion is that: VirtualDub may extract ("Save WAV...") the audio stream in an incompatible form. How can I send you such a .wav file?

As for the volume increase, I understand that I should use boost in order to produce a stronger vorbis file from a .wav source, correct? And what is going to happen if I set PostGain-->Normalize to 150%? Thanks.

[tebasuna51]

My opinion is that: VirtualDub may extract ("Save WAV...") the audio stream in an incompatible form. How can I send you such a .wav file?
VirtualDub -> Tools -> Hex Editor -> File -> Open... your wav file, Edit -> Extract Segment... -> (let defaults 0-1000) Ok and save like segment.wav.
Use http://www.mytempdir.com/ to host this 1 Kb file.
As for the volume increase, I understand that I should use boost in order to produce a stronger vorbis file from a .wav source, correct? And what is going to happen if I set PostGain-->Normalize to 150%? Thanks.
The limit for normalize is 100%, any value over 100% produce clipping.

[har-vas]

Hello tebasuna51 and thank you for your answer. I followed your instructions and I extracted the segment (4 KB). Here is the link: http://www.mytempdir.com/1278363 I hope you can find the source of the problem.

As for normalization, do you mean that every value above 100% is the same as 100% or is even worse? What do you mean by "clipping"? You mean distortion, maybe?

[tebasuna51]

Hello tebasuna51 and thank you for your answer. I followed your instructions and I extracted the segment (4 KB). Here is the link: http://www.mytempdir.com/1278363 I hope you can find the source of the problem.
The problem is a header with a non-standard subchunk 'JUNK', then some strict software don't accept this wav header with wrong messages (now are wrong because your wav have 'data' subchunk with a data length of 568371200 bytes and not '0bit')

You have a "spec-compliant" wav with a length of 568371312 bytes, 2 channels, 16 bit sample, 32000 KHz and 1h 14m 0.4s duration.

If you need use this 'strict' software you can fix the wav with WaveWizard or in command line with WavFix (http://www.mytempdir.com/1278461).

As for normalization, do you mean that every value above 100% is the same as 100% or is even worse? What do you mean by "clipping"? You mean distortion, maybe?
You have a wav with 16 bit signed sample then you have sample values between -32767 and 32767.
If your original wav have a max peak with a value 16383 (don't exist a upper value) you can amplify all the wav by 2 without problem. This is normalize 100% or maximize.

If you normalize 150% mean amplify by 3 this wav, all original values between 10923 and 16383 reach the max value 32767, this is clipping, a kind of distortion.

Of course the rms value of this wav (the volume) is greater but is distorted. The right way is normalize 100% and turn up the audio amplifier volume if you need more.

[har-vas]

Hello tebasuna51 and thank you for your help. I have used your tool to fix my .wav file and indeed I finally managed to open the fixed .wav with HeadAC3he! If I understood correctly, WavFix just makes changes to the header of a .wav file to make it more compatible, correct? I used it only with "-s 32000" switch to fit my needs.

The problem is a header with a non-standard subchunk 'JUNK'
Who is responsible for that subchunk? The Cinergy tv-tuner card and the way it captures the sound or the VirtualDub program and the way it extracts the audio stream from the .avi?

Finally, do you propose me to use the HybridGain in BeSweet or the "PostGain-->Normalize 100%"? I think that the latter is "better", but the former is a newer and faster method. Thanks.

[tebasuna51]

Hello tebasuna51 and thank you for your help. I have used your tool to fix my .wav file and indeed I finally managed to open the fixed .wav with HeadAC3he! If I understood correctly, WavFix just makes changes to the header of a .wav file to make it more compatible, correct? I used it only with "-s 32000" switch to fit my needs.
Nope, in help say:
If any of the next parameters are present the input is considered RAW:
[-i #] BitsPerSample Integer. Default 16. Valid 8, 16, 24, 32
[-f #] BitsPerSample Float. Valid 32 or 64.
[-c #] NumChannels. Default 2. Valid only 1 to 8.
[-s #] SampleRate. Default 48000 Hz. Any value is allowed.
You don't need put nothing only:
wavfix your.wav

Then the header is modified to the standard format, with:
wavfix your.wav -s 32000
you obtain a wav 16 bit (def), stereo (def) and 32 KHz but all the file (also the haeder) is treated as audio data (RAW mode), then you have a click at begining.
Who is responsible for that subchunk? The Cinergy tv-tuner card and the way it captures the sound or the VirtualDub program and the way it extracts the audio stream from the .avi?
I'm not sure but I never see this with Virtualdub.
Finally, do you propose me to use the HybridGain in BeSweet or the "PostGain-->Normalize 100%"? I think that the latter is "better", but the former is a newer and faster method. Thanks.
PreGain: Normalize before encode.
PostGain: Normalize after encode.
HybridGain: a mix between both methods.

Maybe the best is HybridGain but this method, and PostGain, are only available for some output formats (mp3 for instance).