"Warning: Normalisation gain would have resulted in clipping..."

[projeKct]

I love the ReplayGain DSP feature!

"EBU R128 LUFS level -18" is my target.

But I don't understand the "Warning: Normalisation gain would have resulted in clipping, gain has been reduced to prevent this..." message I get sometimes (with softer tracks). I don't want this. I think this "feature/correction" should be optional.

Most mp3 players have a pre-amp option and also a prevent clipping option, so dBpoweramp shouldn't care about this! In fact, Foobar2000 doesn't care, and I'm happy with that! But I would be even happier if dBpoweramp acted similar.

What do you think?

[projeKct]

Re: "Warning: Normalisation gain would have resulted in clipping..."

Question *2 about the RG DSP:

When I compare the values obtained by DBPA and by Foobar2000 (both with EBU R128 LUFS level -18), I find that the Track Gains and Album Gains and Track Peaks are (almost) the same, so I'm very fine with that!

But the Album Peaks are sometimes very different! For example, Tori Amos - Under the Pink album has an Album Peak of 1.065111 with FB2K and 0.977692 with DBPA.

Maybe there's a connection between my two questions (see above)...

[Spoon]

Re: "Warning: Normalisation gain would have resulted in clipping..."

The difference is 0.08 dB I would not say it is that large, I not think you could even hear that difference, different programs with different algorithms perhaps.

[projeKct]

Re: "Warning: Normalisation gain would have resulted in clipping..."

Thanks for your reply, Spoon! You're right, it's not that large!

I did some more investigation... I've understood something: the 0.977692 album peak is from the lossless file (or CD) and the 1.065111 album peak is from the MP3 / VBR0 file (once converted).

Also, the ReplayGain DSP is calculating values before converting to MP3, so it explains the difference I was noticing.

Now I have only requests left:

1) An option to calculate the replaygain values before or after converting (or ripping) the file. Personally, I would choose after, but I suppose it is faster and more "effective" to do it before (on-the-fly), so it could be before by default.

2) Another option to choose if (yes or no) we want that the ReplayGain DSP has to prevent clipping by reducing the gain (leading to the "Warning: Normalisation gain would have resulted in clipping, gain has been reduced to prevent this..." message). Personally, I would choose no (because all my mp3 players or softwares have this "no clipping" feature), but it could be yes by default.

Then, dBpoweramp would be PERFECT for me!

Do you think it would be nice? Is it to much work to be done?

Thanks again...

[garym]

Re: "Warning: Normalisation gain would have resulted in clipping..."

You can already get "1" above by simply converting the album files to the dsp [REPLAYGAIN] after you've ripped and/or converted to mp3. This is another step, but will definitely create RG tag values on the final files. You can do this in a batch method too.

[projeKct]

Re: "Warning: Normalisation gain would have resulted in clipping..."

Thanks garym for your reply, that's what I'm doing with FB2K until DBPA has this feature... because for a strange reason, the ReplayGain DSP (within the Music Converter) isn't able to write anything to the files. Like in this thread.

And what about "2", my second request? (see above)

[projeKct]

Re: "Warning: Normalisation gain would have resulted in clipping..."

Thanks garym for your reply!

Yes, I'm doing this but not with the DBPA Music Converter, because for a strange reason, the ReplayGain DSP is not able to write new tags in the files, see this thread. This DSP works well in the CD Ripper, but not in the Music Converter...

So, until DBPA has these things sorted out, I'm using Foobar2000 to calculate and add ReplainGain values to the extracted files... Like you said, it's an extra step, but no big deal here.

And what about "2", my second request?

[Spoon]

Re: "Warning: Normalisation gain would have resulted in clipping..."

The designer of Replaygain commented that the differences would be inaudible (if RG was calculated before or after mp3), and thus by the same measure the adjustment to anti clip any mp3 addition to the volume would also be inaudible.