28 Nov
2003
28 Nov
'03
11:51 p.m.
Hi all!
I've been trying to stretch a soundfile. It is a piano track with mostly two
voices only. I did it with an fft plain phase vocoder. Now there are a lot of
parameters and I got some undesireable delay. In general the stretched sound
was good, but the delay is disturbing. So here's the question, how should I
set the following parameters to get the best result?
fft-length (as high as possible?), windows per second, window length,
windowtype (hamming, triangular, kasier with several coeffecients).
Any ideas or practical experience with that? - I'm NOT doing it in realtime,
so CPU-time is no problem. I've got time! :-)
Thanbks for any help or good ideas!
Kindest regards
Julien
Julien Patrick Claassen
jclaassen(a)gmx.de
julien(a)c-lab.de
http://www.geocities.com/jjs_home
SBS C-LAB
Fuerstenallee 11
33102 Paderborn
Phone: (+49) 5251 60 6060
Fax: (+49) 5251 60 6065
www.c-lab.de
29 Nov
29 Nov
6:53 a.m.
Hi, if you have short echoes that annoy you, you should decrease
the window length. For your relatively simple spectra and
relatively fast dynamics (piano attacks) I would guess that
1024 or perhaps 2048 should be good, and similar for fft length.
*Not* as high as possible!
Window type doesn't matter much - I usually use Hanning (not Hamming).
I don't know what program you are using, but I assume that "windows per
second" has to do with overlap. If you do very long stretches, you might
like to increase this value, but otherwise it shouldn't matter much.
Dr. Oyvind Hammer, senior researcher
PGP - Physics of Geological Processes
University of Oslo
I've been trying to stretch a soundfile. It is a
piano track with mostly two
voices only. I did it with an fft plain phase vocoder. Now there are a lot of
parameters and I got some undesireable delay. In general the stretched sound
was good, but the delay is disturbing. So here's the question, how should I
set the following parameters to get the best result?
fft-length (as high as possible?), windows per second, window length,
windowtype (hamming, triangular, kasier with several coeffecients).
Any ideas or practical experience with that? - I'm NOT doing it in realtime,
so CPU-time is no problem. I've got time! :-)
8:52 a.m.
el Sat, 29 Nov 2003 11:51:42 +0100 (MET)
Oyvind Hammer <oyvind.hammer(a)nhm.uio.no> escribió:
I would guess that
1024 or perhaps 2048 should be good, and similar for fft length.
*Not* as high as possible!
i second that opinion (not that it matters much, i'm no dsp expert..).
Window type doesn't matter much [...]
here i disagree. triangular would be inconvenient, as well as kaiser with a
low coefficient (approaching a rectangular window). i'd try something
between 0.35 and 0.6 to start.
- I usually use Hanning (not Hamming).
those are good bets, as well as blackmann windows, which i usually prefer
for certain applications.
lj
-
9:49 a.m.
Window type
doesn't matter much [...]
here i disagree. triangular would be inconvenient, as well as kaiser with a
low coefficient (approaching a rectangular window). i'd try something
between 0.35 and 0.6 to start. 
1 Dec
1 Dec
7:27 a.m.
Julien Claassen wrote
Hi all!
I've been trying to stretch a soundfile. It is a piano track with mostly two
voices only. I did it with an fft plain phase vocoder. Now there are a lot of
parameters and I got some undesireable delay. In general the stretched sound
was good, but the delay is disturbing. So here's the question, how should I
set the following parameters to get the best result?
fft-length (as high as possible?), windows per second, window length,
windowtype (hamming, triangular, kasier with several coeffecients).
Any ideas or practical experience with that? - I'm NOT doing it in realtime,
so CPU-time is no problem. I've got time! :-)
Audacity 1.2
--
Alexandre Prokoudine | "When you set yourself on fire and aim
ALT Linux Documentation Team | for the sky, you hope to leave behind
E-mail: avp(a)altlinux.ru | some sparks of heat and light"
JabberID: avp(a)altlinux.org | Neil Peart
2 Dec
2 Dec
7:57 a.m.
Thanks for the tip audacity 1.2. But I think it's GUI-based and so NOT very
useful for me - as I'm blind... :-)
Kindest regards
Julien
Julien Patrick Claassen
jclaassen(a)gmx.de
julien(a)c-lab.de
http://www.geocities.com/jjs_home
SBS C-LAB
Fuerstenallee 11
33102 Paderborn
Phone: (+49) 5251 60 6060
Fax: (+49) 5251 60 6065
www.c-lab.de
11:35 a.m.
I tried the time stretching effect in Audacity 1.2, and even for small
amounts of stretching ( less than 1.5 ) I found that it gives very poor
results, with lots of artifacts. Hardly usable, in my opinion.
Julien, what software are you using? Do you use csound? Some time ago I
wrote a simple instrument for time stretching and/or pitch transposing,
using the newer streaming phase vocoding opcodes. I got much better results
than with Audacity, anyway.
Best,
lj
--
3 Dec
3 Dec
3:14 a.m.
luis jure wrote:
I tried the time stretching effect in Audacity 1.2,
and even for small
amounts of stretching ( less than 1.5 ) I found that it gives very poor
results, with lots of artifacts. Hardly usable, in my opinion.
It appears the latest cvs version of ardour has better support. I just
stretched Kermit and it sounds real nice (tm)
Julien, what software are you using? Do you use
csound? Some time ago I
wrote a simple instrument for time stretching and/or pitch transposing,
using the newer streaming phase vocoding opcodes. I got much better results
than with Audacity, anyway.
Best,
lj
--
Patrick Shirkey - Boost Hardware Ltd.
Http://www.boosthardware.com
Http://www.djcj.org - The Linux Audio Users guide
========================================
Apparently upon the beginning of the barrage, the donkey broke
discipline and panicked, toppling the cart. At that point, the rockets
disconnected from the timer, leaving them strewn around the street.
Tethered to the now toppled cart, the donkey was unable to escape before
the arrival of U.S. troops.
United Press International
Rockets on donkeys hit major Baghdad sites
By P. MITCHELL PROTHERO
Published 11/21/2003 11:13 AM
10:02 a.m.
luis jure wrote:
el Wed, 03 Dec 2003 16:13:04 +0900
Patrick Shirkey <pshirkey(a)boosthardware.com> escribió:
No I mean ardour ;)
--
Patrick Shirkey - Boost Hardware Ltd.
Http://www.boosthardware.com
Http://www.djcj.org - The Linux Audio Users guide
========================================
Apparently upon the beginning of the barrage, the donkey broke
discipline and panicked, toppling the cart. At that point, the rockets
disconnected from the timer, leaving them strewn around the street.
Tethered to the now toppled cart, the donkey was unable to escape before
the arrival of U.S. troops.
United Press International
Rockets on donkeys hit major Baghdad sites
By P. MITCHELL PROTHERO
Published 11/21/2003 11:13 AM
It appears the latest cvs version of ardour has
better support.
i guess you mean audacity (not ardour)? i'll try the cvs later, thank you
for the info.
9:39 a.m.
Julien Claassen wrote
Thanks for the tip audacity 1.2. But I think it's
GUI-based and so NOT very
useful for me - as I'm blind... :-)
A-ha, I didn't know that. Well, there is a stretch effect in sox
--
Alexandre Prokoudine | "When you set yourself on fire and aim
ALT Linux Documentation Team | for the sky, you hope to leave behind
E-mail: avp(a)altlinux.ru | some sparks of heat and light"
JabberID: avp(a)altlinux.org | Neil Peart
1:01 p.m.
ecasound's pitch shifter might be useful. From man ecasound:
-ei:pitch-shift-%
Pitch shifter. Modifies audio pitch by altering its length.
I haven't tried it, yet. I seem to recall something about it not working
in realtime, but I could be making that up ... looking at your original
post I see that isn't a requirement anyway.
-Eric Rz.
On Wed, Dec 03, 2003 at 04:37:37PM +0300, Alexandre Prokoudine wrote:
Julien Claassen wrote
Thanks for the tip audacity 1.2. But I think
it's GUI-based and so NOT very
useful for me - as I'm blind... :-)
A-ha, I didn't know that. Well, there is a stretch effect in sox
--
Alexandre Prokoudine | "When you set yourself on fire and aim
ALT Linux Documentation Team | for the sky, you hope to leave behind
E-mail: avp(a)altlinux.ru | some sparks of heat and light"
JabberID: avp(a)altlinux.org | Neil Peart
9 Dec
9 Dec
8:49 a.m.
Hi!
Ecasound's pitch-shift can work in realtime. Only you'd need a file as
input, not a device. You could even do some modulation on shift-% via the
-k-something options.
thanks for the idea anyway! If you'd like to try feel free to, it's working
fine!
Kindest regards
Julien
Julien Patrick Claassen
jclaassen(a)gmx.de
julien(a)c-lab.de
http://www.geocities.com/jjs_home
SBS C-LAB
Fuerstenallee 11
33102 Paderborn
Phone: (+49) 5251 60 6060
Fax: (+49) 5251 60 6065
www.c-lab.de
9:53 a.m.
Julien,
Julien Claassen wrote:
Ecasound's pitch-shift can work in realtime.
Only you'd need a file as
input, not a device.
Have you found a way to use ecasound's pitch-shift to stretch a sound at
all? If so, could you post the command line you are using please?
'man ecasound' and 'ecasound --help' didn't mention anything about
time-stretch, only pitch-shift.
Be curious to know.
Best,
D.
--
derek holzer ::: http://www.umatic.nl
---Oblique Strategy # 30:
"Change specifics to ambiguities"
1:23 p.m.
I think the pitch shift works by either compressing or stretching the
time....
-ei:pitch-shift-%
Pitch shifter. Modifies audio pitch by altering its length.
which i think means that -ei:50, for instance, will make the file play
twice as long while sounding at half pitch and -ei:200 will make
the file play half as long while sounding at twice the pitch.
I did some simple tests on this at home the other night. A pitch shift
of less than 100% is equivalent to time stretching while a pitch shift
of greater than 100% is the same as time compressing.
-Eric Rz.
On Tue, Dec 09, 2003 at 03:04:34PM +0100, derek holzer wrote:
Julien,
Julien Claassen wrote:
Ecasound's pitch-shift can work in realtime.
Only you'd need a file as
input, not a device.
Have you found a way to use ecasound's pitch-shift to stretch a sound at
all? If so, could you post the command line you are using please?
'man ecasound' and 'ecasound --help' didn't mention anything about
time-stretch, only pitch-shift.
Be curious to know.
Best,
D.
--
derek holzer ::: http://www.umatic.nl
---Oblique Strategy # 30:
"Change specifics to ambiguities"
10 Dec
10 Dec
2:46 p.m.
Eric,
Eric Dantan Rzewnicki wrote:
I think the pitch shift works by either compressing or
stretching the
time....
This is, unfortunately, contrary to what I was hoping.
What I hoped was that ecasound could time-stretch or -compress
independent of pitch-shifting. At the very least, I thought there might
be a way to simultaneously slow or speed playback while pitch-shifting
to compensate. But this is a kind of windowed operation that ecasound
probably doesn't have.
Better for me to investigate CSound, I guess. Or Pure Data, but it
doens't make for a very good batch-processor.
The time-stretch and -compress options in Ardour seem to be pretty good
[although they also tend to crash Ardour if you aren't careful], but
Ardour is an even worse batch-processor than PD!!!
Best,
D.
--
derek holzer ::: http://www.umatic.nl
---Oblique Strategy # 167:
"Use cliches"
5:19 p.m.
On Wed, Dec 10, 2003 at 07:57:33 +0100, derek holzer wrote:
The time-stretch and -compress options in Ardour seem
to be pretty good
[although they also tend to crash Ardour if you aren't careful], but
Ardour is an even worse batch-processor than PD!!!
Ardour uses soundtouch http://sky.prohosting.com/oparviai/soundtouch/
which comes with batch programs IIRC.
- Steve
10:28 p.m.
Steve, thanks for pointing to this. I have a friend at work who is
learning to play bass. I showed him the soundstretch utility from that
project as a way to slow down songs he's trying to learn to play along
with without changing the pitch. He seemed pretty excited about it. :)
We tried it out and found that it works well for time
stretching/compressing with no pitch change. Some of the other options
didn't work, but it's just an example utility.
-Eric Rz.
Steve Harris wrote:
On Wed, Dec 10, 2003 at 07:57:33 +0100, derek holzer wrote:
The time-stretch and -compress options in Ardour
seem to be pretty good
[although they also tend to crash Ardour if you aren't careful], but
Ardour is an even worse batch-processor than PD!!!
Ardour uses soundtouch http://sky.prohosting.com/oparviai/soundtouch/
which comes with batch programs IIRC.
- Steve
7443
days inactive
7455
days old
linux-audio-user@lists.linuxaudio.org
17 comments
9 participants
participants (9)
-
Alexandre Prokoudine -
derek holzer -
Eric Dantan Rzewnicki -
Eric Dantan Rzewnicki -
Julien Claassen -
luis jure -
Oyvind Hammer -
Patrick Shirkey -
Steve Harris