20 Dec
2010
20 Dec
'10
10:27 a.m.
Hello
I bought the natural drum samples (http://www.naturaldrum.com/). It contains
WAVs and presets for Kontakt and Halion. Now I'd like to create some gigasampler
files in order to use it with linuxsampler.
The documentation of the natural drum sample library is quite good. The only
thing missing is the "loudness" of each sample in order to map each sample to a
velocity level from 0-127.
What would you recommend in order to calculate the "peek" of each drum sample
automatically? Is there a library which could do this? I would also be happy
with a command line tool like this:
$ peek bla.wav
Peek value: 12345
I could then write a C++-App using libgig.
Any ideas? Libraries? Algorithms?
Thanks!
Oliver
20 Dec
20 Dec
10:36 a.m.
On Mon, Dec 20, 2010 at 4:27 AM, Oliver Jaun <olijaun(a)yahoo.com> wrote:
Hello
I bought the natural drum samples (http://www.naturaldrum.com/). It
contains WAVs and presets for Kontakt and Halion. Now I'd like to create
some gigasampler files in order to use it with linuxsampler.
The documentation of the natural drum sample library is quite good. The
only thing missing is the "loudness" of each sample in order to map each
sample to a velocity level from 0-127.
What would you recommend in order to calculate the "peek" of each drum
sample automatically? Is there a library which could do this? I would also
be happy with a command line tool like this:
$ peek bla.wav
Peek value: 12345
I could then write a C++-App using libgig.
Any ideas? Libraries? Algorithms?
Thanks!
Oliver
_______________________________________________
Linux-audio-dev mailing list
Linux-audio-dev(a)lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-dev
You might try looking at Vamp Plugins:
http://www.vamp-plugins.org/index.html
More specifically, the developer tools seem to allow you to use a vamp
plugin easily from the command line:
http://www.vamp-plugins.org/develop.html I'm sure there's one to do RMS
measurement, or whatever.
Jeremy
10:41 a.m.
2010/12/20 Oliver Jaun <olijaun(a)yahoo.com>om>:
Hello
I bought the natural drum samples (http://www.naturaldrum.com/). It contains
WAVs and presets for Kontakt and Halion. Now I'd like to create some
gigasampler files in order to use it with linuxsampler.
The documentation of the natural drum sample library is quite good. The only
thing missing is the "loudness" of each sample in order to map each sample
to a velocity level from 0-127.
Loudness is a complex beast (i.e., lots of psychoacoustics involved).
What would you recommend in order to calculate the
"peek" of each drum
sample automatically? Is there a library which could do this? I would also
be happy with a command line tool like this:
$ peek bla.wav
Peek value: 12345
I could then write a C++-App using libgig.
Any ideas? Libraries? Algorithms?
Something like this could do the trick for you (pseudocode):
attack = 0.99
level = 0
peek = 0
for each sample
level = (1 - attack) * level + attack * abs(sample)
peek = MAX(peek, level)
(I suppose that samples have 0 mean, i.e. DC component filtered out)
The attack thing should be close but less than 1, you can try different values.
Stefano
3:50 p.m.
On Mon, Dec 20, 2010 at 4:27 AM, Oliver Jaun <olijaun(a)yahoo.com> wrote:
Hello
I bought the natural drum samples (http://www.naturaldrum.com/). It contains
WAVs and presets for Kontakt and Halion. Now I'd like to create some
gigasampler files in order to use it with linuxsampler.
The documentation of the natural drum sample library is quite good. The only
thing missing is the "loudness" of each sample in order to map each sample
to a velocity level from 0-127.
What would you recommend in order to calculate the "peek" of each drum
sample automatically? Is there a library which could do this? I would also
be happy with a command line tool like this:
$ peek bla.wav
Peek value: 12345
sndfile-info /music/misc/onewayjam_organ.ogg
Version : libsndfile-1.0.20
========================================
File : /music/misc/onewayjam_organ.ogg
Length : 3939068
Vorbis library version : Xiph.Org libVorbis 1.3.1
Bitstream is 2 channel, 44100 Hz
Encoded by: Xiph.Org libVorbis I 20020717
End
----------------------------------------
Sample Rate : 44100
Frames : 18924456
Channels : 2
Format : 0x00200060
Sections : 1
Seekable : TRUE
Duration : 00:07:09.126
Signal Max : 0.805912 (-92.18 dB) <<<<<<<<< HEADS UP!
Works for any audio fle that libsndfile can read, which is just about anything.
there is a complication: loudness is no identical to maximum sample
value, but the relationship is good enough for government work, so to
speak.
--p
7:24 p.m.
New subject: peak has nothing to do with loudness (check out EBU R128)
On 12/20/2010 03:49 PM, Paul Davis wrote:
<...>
The
documentation of the natural drum sample library is quite good. The only
thing missing is the "loudness" of each sample in order to map each sample
to a velocity level from 0-127.
What would you recommend in order to calculate the "peek" of each drum
sample automatically? Is there a library which could do this? I would also
be happy with a command line tool like this:
sndfile-info /music/misc/onewayjam_organ.ogg Signal Max : 0.805912 (-92.18 dB)
<<<<<<<<< HEADS UP!
Works for any audio fle that libsndfile can read, which is just about anything.
there is a complication: loudness is no identical to maximum sample
value, but the relationship is good enough for government work, so to
speak.
no, it's not. there is absolutely no meaningful relation whatsoever
between perceived loudness and peak value. that's why the EBU is (after
all those years) moving to k-weighted loudness as a metering standard
rather than QPPM. true, peak measurement used to be the norm, but it was
never good for anything (not even government work) except to avoid
clipping downstream.
florian camerer of orf presented a very enlightening talk at the
tonmeistertagung in leipzig this year, which i had meant to post earlier
but forgot. thanks for this reminder :)
http://tech.ebu.ch/docs/techreview/trev_2010-Q3_loudness_Camerer.pdf
it starts a bit boastful, but rightfully so, since it's in effect _the_
peace treaty to end the loudness war.
bottom line: those who have been using bob katz' k-system (for example
as implemented by fons adriaensen in jkmeter) don't have to change much...
but if you're interacting with radio stations regularly, you might want
to implement the EBU R128 standard, whose loudness measuring method is
slightly different.
7:29 p.m.
New subject: peak has nothing to do with loudness (check out EBU R128)
2010/12/20 Jörn Nettingsmeier <nettings(a)folkwang-hochschule.de>de>:
there is if they are all samples of the same instrument.
that's why i said "good enough for government work" :)
there is a
complication: loudness is no identical to maximum sample
value, but the relationship is good enough for government work, so to
speak.
no, it's not. there is absolutely no meaningful relation whatsoever
between perceived loudness and peak value.
7:44 p.m.
New subject: peak has nothing to do with loudness (check out EBU R128)
On 12/20/2010 07:29 PM, Paul Davis wrote:
2010/12/20 Jörn Nettingsmeier
<nettings(a)folkwang-hochschule.de>de>:
there is if they are all samples of the same instrument.
that's why i said "good enough for government work" :)
ok, i took that as a general remark...
but still: chances are multiple dynamic variants of the same instrument
have the same peak value, if the maker used peak limiting to avoid s/n
issues on very dynamic instruments (which is quite likely).
and you will not be able to map multiple sounds to perceptually equal
loudness levels. say you want to create a "drum map" for your keyboard:
you will want the toms to be just as loud as snare and bass drum when
you play them with "average" velocity.
maybe the "replay gain" measurement features present in some audio
players could be used for the job (perhaps some are even scriptable)?
i'd expect them to have a very long time window though, so it's probably
best to feed them 10-30 repetitions of the same sound for the
measurement system to stabilize.
there is
a complication: loudness is no identical to maximum sample
value, but the relationship is good enough for government work, so to
speak.
no, it's not. there is absolutely no meaningful relation whatsoever
between perceived loudness and peak value. 
21 Dec
21 Dec
1:10 a.m.
New subject: peak has nothing to do with loudness (check out EBU R128)
Excerpts from Jörn Nettingsmeier's message of 2010-12-20 19:27:28 +0100:
On 12/20/2010 03:49 PM, Paul Davis wrote:
<...>
Wow, amazing that someone from ORF manage to write something useful. I
thought ORF doesn't employ anything but a board of directors and freelancers,
so whenever something goes wrong they can blame it on contractors.
I'm really very surprised that there's someone inside this beast who
actually does something.
The
documentation of the natural drum sample library is quite good. The only
thing missing is the "loudness" of each sample in order to map each sample
to a velocity level from 0-127.
What would you recommend in order to calculate the "peek" of each drum
sample automatically? Is there a library which could do this? I would also
be happy with a command line tool like this:
sndfile-info /music/misc/onewayjam_organ.ogg Signal Max : 0.805912 (-92.18 dB)
<<<<<<<<< HEADS UP!
Works for any audio fle that libsndfile can read, which is just about anything.
there is a complication: loudness is no identical to maximum sample
value, but the relationship is good enough for government work, so to
speak.
no, it's not. there is absolutely no meaningful relation whatsoever
between perceived loudness and peak value. that's why the EBU is (after
all those years) moving to k-weighted loudness as a metering standard
rather than QPPM. true, peak measurement used to be the norm, but it was
never good for anything (not even government work) except to avoid
clipping downstream.
florian camerer of orf presented a very enlightening talk at the
tonmeistertagung in leipzig this year, which i had meant to post earlier
but forgot. thanks for this reminder :)
http://tech.ebu.ch/docs/techreview/trev_2010-Q3_loudness_Camerer.pdf
it starts a bit boastful, but rightfully so, since it's in effect _the_
peace treaty to end the loudness war.
bottom line: those who have been using bob katz' k-system (for example
as implemented by fons adriaensen in jkmeter) don't have to change much...
but if you're interacting with radio stations regularly, you might want
to implement the EBU R128 standard, whose loudness measuring method is
slightly different. 
20 Dec
20 Dec
4:29 p.m.
måndag 20 december 2010 10.27.23 skrev Oliver Jaun:
Hello
I bought the natural drum samples (http://www.naturaldrum.com/). It
contains WAVs and presets for Kontakt and Halion. Now I'd like to create
some gigasampler files in order to use it with linuxsampler.
...
I could then write a C++-App using libgig.
Any ideas? Libraries? Algorithms?
This is not what you asked for but FYI, this is what I did and still works on
on NDK: -I converted almost every single sample to mono and generally choosed
the "best" 8 to 16 sample layers for every drum item by ear. I made (making)
onn gig file for every item like one for hh13 (with brushes, sticks and hand
and so on) and one rock snare and so on and then combines them (gig files) to
be Natural Drum Kits.
Jostein
27 Dec
27 Dec
9:06 a.m.
Hello
Thanks to all for the good hints. I succeeded yesterday in creating a first
"closed-hi-hat-gigasampler" file with my "application". I'm using
currently
libsndfile to calculate the "signal max" (as proposed by Paul Davis).
I will publish the source code in case that something usable will result...
Regards
Oliver
________________________________
From: Jostein Chr. Andersen <jostein(a)vait.se>
To: linux-audio-dev(a)lists.linuxaudio.org
Cc: Oliver Jaun <olijaun(a)yahoo.com>
Sent: Mon, December 20, 2010 4:20:49 PM
Subject: Re: [LAD] Identify Peek in Drum Samples
måndag 20 december 2010 10.27.23 skrev Oliver Jaun:
Hello
I bought the natural drum samples (http://www.naturaldrum.com/). It
contains WAVs and presets for Kontakt and Halion. Now I'd like to create
some gigasampler files in order to use it with linuxsampler.
...
I could then write a C++-App using libgig.
Any ideas? Libraries? Algorithms?
This is not what you asked for but FYI, this is what I did and still works on
on NDK: -I converted almost every single sample to mono and generally choosed
the "best" 8 to 16 sample layers for every drum item by ear. I made (making)
onn gig file for every item like one for hh13 (with brushes, sticks and hand
and so on) and one rock snare and so on and then combines them (gig files) to
be Natural Drum Kits.
Jostein
4879
days inactive
4886
days old
linux-audio-dev@lists.linuxaudio.org
9 comments
7 participants
participants (7)
-
Jeremy -
Jostein Chr. Andersen -
Jörn Nettingsmeier -
Oliver Jaun -
Paul Davis -
Philipp Überbacher -
Stefano D'Angelo