Re: [LAU] Batch normalising wav files

On 12/24/2010 06:07 PM, Q wrote:
> Jörn Nettingsmeier wrote:
>>> I'm wanting to batch process a bunch of wav files, normalising them all
>>> to -0.3 dBFS.
>>
>> that's high! you will almost certainly get distortion in the analog
>> stages of cheaper playback equipment with a signal that hot. think
>> inter-sample peaks.
>
> I got the impression most masters were peak normalised to 0 dBFS these
> days and less frequently to slightly under to avoid playback issues.

eat shit, millions of flies can't be wrong :)
mastering that hot is a major technical blunder proposed by idiots with
either bad hearing or a profound hate for music.

> Out of interest, how high can intersample peaks get above the highest
> peaks in a file?

imagine the positive half of a sine wave, so that two consecutive
samples representing the very top of the wave are at 0dBFS. it is clear
that between those samples, the real peak value of the sound must be
higher than 0dBFS, and the reconstruction filter will see this higher
value. that means unless your analog stage has headroom for this, it
will clip. the higher the frequency, the higher above full scale those
inter-sample peaks can be.

>>> I thought I could do this with normalize/normalize-audio, but it doesn't
>>> look to be possible. The man page suggests that simply analysing the
>>> file and raising the level so the loudest peak is 0 dBFS isn't
>>> normalising, which is news to me:
>>
>> depends. that's peak normalisation. what you probably want is loudness
>> normalisation, which adjusts the levels so that the perceived loudness
>> is constant across tracks.
>
> No, peak normalisation is what I'm after; this isn't a CD collection I'm
> processing.

ok. though i have a hard time seeing any use for peak normalisation
unless you want to deliver your material at low bit depths, such as a cd.

>>> " --peak
>>> Adjust using peak levels instead of RMS levels. Each file will be
>>> adjusted so that its maximum sample is at full scale. This just gives a
>>> file the maximum volume possible without clipping; no normalization is
>>> done."
>>
>> that is a funny definition of normalisation.
>
> I get the impression that the term normalisation has taken on another
> meaning among non-music producers to mean replay gain normalisation for
> whole albums or music collections. Doesn't replay gain either peak
> normalise the loudest track and raise all the others by the same amount
> (to preserve inter-song dynamics), or in some case do loudness
> normalisation? Personally, I don't bother with such things, volume knobs
> are what we evolved fingers for!

correct replay gain should strive for constant perceived loudness, so
that you can use those wonderful evolved fingers for more important
things, such as doing the dishes or ironing, without having to rush to
the volume knob after each track.

>> like julien suggested, use a standard peak normalizer and then attenuate
>> by 0.3dB )(if you are sure that's what you want to do).
>>
> When you put it as simply as that it's blindingly obvious! I can't
> believe it didn't occur to me and one extra step when batch processing
> is not a big deal at all. Blame the cold weather and a brain freeze!

i hope that come spring, your freshly thawed brain will even realize
that you don't want peak normalisation at all :)

ok, down from the soapbox, and season's greetings,

jörn

_______________________________________________
Linux-audio-user mailing list
Linux-audio-user@email-addr-hidden
http://lists.linuxaudio.org/listinfo/linux-audio-user
Received on Fri Dec 24 20:15:05 2010