HydrogenAudio

Lossless Audio Compression => WavPack => Topic started by: Didjeridoo on 2010-06-08 15:10:45

Title: wavpack.exe VBR Settings
Post by: Didjeridoo on 2010-06-08 15:10:45
Hello))
I want to ask a question about the settings wavpack.exe
Why can raise only a nominal bitrate mode hybrid, why is not quality mode?

I need to encode the set of wav files with different bit depth and sampling rate. I need to set settings for each sample?

Thank you
Title: wavpack.exe VBR Settings
Post by: shadowking on 2010-06-08 16:16:44
You can still use BPS instead of bitrate:  -b  2.0 to 23.9 bits/sample . The bitrate will then adjust automatically and even work for mono. It wont measure the audibility of noise though as VBR isn't yet implemented . For 44.1khz of more i think you can use the same BPS rate for cd audio. For low rez, you have to be more conservative as there is less masking .

Example :

-b4x = 350k for 44.1khz 16bit or 760k for 96khz 24bit
Title: wavpack.exe VBR Settings
Post by: Didjeridoo on 2010-06-08 17:30:09
Thank shadowking, you helped me a lot)))
Title: wavpack.exe VBR Settings
Post by: Didjeridoo on 2010-06-09 16:24:04
-miqhhb4.5j1 --use-dns
This parameters gives me normal quality?
Title: wavpack.exe VBR Settings
Post by: shadowking on 2010-06-10 14:22:10
-miqhhb4.5j1 --use-dns
This parameters gives me normal quality?



It will give high transparent quality for almost any situation. You can drop -j1 as it is already default. It may be good idea to include at least -x(1) for high-rez audio. For something much lighter to decode or transcode consider the normal or fast mode in conjunction with -x :

-b384hh >> -b415fx3




Title: wavpack.exe VBR Settings
Post by: bryant on 2010-06-10 17:43:22
-miqhhb4.5j1 --use-dns
This parameters gives me normal quality?

I would agree with shadowking that -hh is probably overkill at that bitrate, and is slower to decode. I would drop it to -h and add at least -x. I generally recommend the highest -x argument that you are willing to wait for (up to -x6) because it does not affect decode speed.

Note that --use-dns is also enabled by default at normal sampling rates (under 64 KHz). At high sampling rates the default is for the noise to be shaped up into inaudible frequencies (-s1) which I think is better than the dynamic shaping (which tries to put the noise under the music).

David