Replay Gain and Clipping/Compression

[David Lutz]

I recently went through my FLAC collection and added Replay Gain tags to
take advantage of the new support for Replay Gain. My collection is
heavily biased toward good classical recordings (i.e. discs with large
dynamic range). I have noticed that the Replay Gain for a good number
of these albums is actually a positive value, which I take to mean the
audio gain will be increased instead of decreased. A quick run through
the numbers shows that for some of the albums the new peak sample value
with gain applied will be greater that 1 (in one extreme case the new
peak sample is 3.4). My question is how does slimserver/squeezebox
handle this? Is it clipped or compressed somehow?

D.L.

[Jim]

http://www.replaygain.org/ sort of explains but then doesn't. Without having access to the CD or the waveform just telling me it's +3.4 doesn't mean it's going to clip, load it up in a sound editor, make a test decode (keep the orignal as this does actually write a changed file) with a --apply-repaly-gain-which-is-not-lossess option and examine that file.

I don't like the fact it's webpage still says a "Propsed Standard" but anyhow if it's ReplayGain is set to a value that clips then the Squeezebox would be a pretty poor implementation of ReplayGain if it didn't clip like it was told too. http://www.hydrogenaudio.org is the place to go for questions like this.

I think the idea of ReplayGain is much more suited to someone who listens more to rock/pop/dance music etc... where some super-compressed modern rock record makes a earlier one sound lower.

If you mostly listenig to Classical and are aware of things such as Dynamic Gain and louder does not mean better then just because it's there it doesn't mean you have to use it.

Anyhow, it's your ears and your music - it's only a switch so try it out for a while, nothing gets changed.

[cliveb]

I don't like the fact it's webpage still says a "Propsed Standard" but anyhow if it's ReplayGain is set to a value that clips then the Squeezebox would be a pretty poor implementation of ReplayGain if it didn't clip like it was told too. http://www.hydrogenaudio.org is the place to go for questions like this.
When you add ReplayGain tags to a file, you should get values for the existing peak level as well as the suggested level adjustment. The playback device can (should?) figure out if applying the full suggested gain will then cause clipping, and back off the gain accordingly. Example: Foobar2000 has an option for this. There is no reason why the Squeezebox could not also have this option. (I don't know if it does; I don't use ReplayGain on my SB2 - haven't even loaded up a recent enough firmware to check it out).

[egd]

My question is how does slimserver/squeezebox
handle this? Is it clipped or compressed somehow?

If I've understood the various threads I've looked into I think the answer is yes, clipping does occur.