minor lossy audio problem

[thundat00th]

http://www.dapreview.net/p/forum/forum_viewtopic.php?77579

just to make the developers aware of this slight problem, i wish i could offer more help, but im only an amateur in digital audio, and know nothing of programming

[Llorean]

This really isn't anything for Rockbox to be concerned with. If it's as described, it's entirely a problem on the encoding and/or replaygain scanning side.

[saratoga]

That guys like 8 years late in discovering replaygain.

[Lear]

The effect noticed is a) well known (at least at places like Hydrogenaudio) and b) inherent in lossy coding. Two things worth mentioning (should be mentioned at DAPreview really, but...):

1) The peak value doesn't say anything about how often it clips or how audible it is.

2) The WMA (and probably "wavepack") decoder used is likely clipping the output before the ReplayGain scanner sees it, so claiming that WMA is more accurate based on the reported peak value is incorrect.

Anyway, Rockbox can't really do anything about it except support ReplayGain, which it does.

Edit: Oops, missed the Rockbox specific bits in the DAPreview thread... The Rockbox WMA decoder does not clip the output (that is done by the core, as it converts the samples to the format required by the hardware, after having done equalizing, ReplayGain processing and the like), so in that way it behaves like MP3.

[thundat00th]

i was 11 8 years ago

and thank you for clarifying, i only know what i can see and/or measure

[Zardoz]

keep on keepin on thundat00th!
The wavpack format seems to have two modes, am i right:? Compressed and non-compressed....

[Lear]

WavPack has two modes, yes, but they're called lossless and lossy. Both are compressed.