[MLA] Touch base with Buddy...

[Bob Katz]

Congratulations in advance, Eelco! I have yet to read your last
revision.
I have Buddy's ear and he and I have conversed several times in the
past year, but all about MFIT stuff and not about  normalization. 
I'd like us to touch base with Buddy. Should we compose a group email
summarizing our thoughts and send it to Buddy?

Let's summarize our points:
a) loudness normalization on by default. Ideally BS.1770. 
b) album norm on all the time.
c) No upward normalization unless the peak level would not clip.
d) No peak limiting
e) Point out that singles (which have no album) will automatically
sound as loud as the loudest song in any album even if the single is a
ballad or soft song. Well, this behaves no worse than track
normalization ever did. I guess it will work itself out. The big
discrepancies will be between singles that are soft songs against that
same song in the album. But it's no worse than the old days when
ballads were often released as singles, peak normalized, louder than
how they appeared on the album. 
f) If they choose a higher target, like -14 LUFS, it works out fine and
does not cause a loudness race:
--- ONLY if album normalization is on all the time and
---ONLY if album normalization is based on the loudest track in the
album and
---ONLY if there is no upward normalization beyond the clipping point. 
g) Album norm would require a new field in the metadata that Apple is
apparently not using. One which calculates the loudest song in the
album. Personally I think the field should contain the absolute LUFS of
the loudest song. That allows playback systems to easily update to new
targets.
h) Currently itunes only turns on album norm in a playlist. It also
only turns it on with contiguous tracks....     My studies and your
study have shown that contiguous is not necessary to enjoy the benefits
of album norm. 
iTunes currently keeps on calculating a new "average" and recalculates
the album normalization in a playlist depending on how many tracks from
the album it discovers. It's actually far more complicated and fraught
with bugs than the simple "album norm all the time" principle so the
apple algorithms will become simpler, not complicated. 


Best wishes,


Bob

On Wed, 2017-05-24 at 08:09 +0200, Grimm | Eelco wrote:
> Hi Thomas,
> 
> > Well done if Apple has been convinced by your study to turn on
> > loudness normalization by default. Still, I encourage you to submit
> > it for JAES or a conference so the scientific scrutinization
> > process can get started, and we can get out of this stalled state.
> 
> Yes I will. It is planned for the New York conference. I will
> definitely keep you all involved in the writing of that paper.
> 
> Cheers,
> Eelco
> 
> 
> _______________________________________________
> MLA mailing list
> MLA at grimmaudio.nl
> http://grimmaudio.nl/mailman/listinfo/mla_grimmaudio.nl
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://grimmaudio.nl/pipermail/mla_grimmaudio.nl/attachments/20170526/223b889f/attachment.html>