Utility to speed up mp3 playback?

[root]

I would like to playback some spoken mp3 files to speed the rate
in order to determine the limit of intelligibility. Is there
some linux capability to achieve this end?

Thanks.

[wobbles]

Post by root
I would like to playback some spoken mp3 files to speed the rate
in order to determine the limit of intelligibility. Is there
some linux capability to achieve this end?
Thanks.

Audacity seems to provide this capability; perhaps you could experiment with
some settings.
http://manual.audacityteam.org/man/change_speed.html

[root]

Post by wobbles

Post by root
I would like to playback some spoken mp3 files to speed the rate
in order to determine the limit of intelligibility. Is there
some linux capability to achieve this end?
Thanks.

Audacity seems to provide this capability; perhaps you could experiment with
some settings.
http://manual.audacityteam.org/man/change_speed.html

Thanks for responding. I should have mentioned that mplayer allows
for speed-up but the tone is shifted up so that the voices are
as if one had been breathing helium. I would like normal voices.

I don't have audacity, but i will give it a try.

[root]

Post by root

Post by wobbles

Post by root
I would like to playback some spoken mp3 files to speed the rate
in order to determine the limit of intelligibility. Is there
some linux capability to achieve this end?
Thanks.

Audacity seems to provide this capability; perhaps you could experiment with
some settings.
http://manual.audacityteam.org/man/change_speed.html

Thanks for responding. I should have mentioned that mplayer allows
for speed-up but the tone is shifted up so that the voices are
as if one had been breathing helium. I would like normal voices.
I don't have audacity, but i will give it a try.

I give up on audacity. I am spinning my wheels trying to resolve
wxWidgets and WzPython.

[Dan C]

Post by root

Post by wobbles

I would like to playback some spoken mp3 files to speed the rate in
order to determine the limit of intelligibility. Is there some linux
capability to achieve this end?
Thanks.

Audacity seems to provide this capability; perhaps you could
experiment with some settings.
http://manual.audacityteam.org/man/change_speed.html

Thanks for responding. I should have mentioned that mplayer allows for
speed-up but the tone is shifted up so that the voices are as if one
had been breathing helium. I would like normal voices.
I don't have audacity, but i will give it a try.

I give up on audacity. I am spinning my wheels trying to resolve
wxWidgets and WzPython.

I take it you've never heard of slackbuilds.org ?

[root]

Post by Dan C
I take it you've never heard of slackbuilds.org ?

Of course I have. Audacity doesn't make because it wants 3.0
of wxWdigets (or something like that) and 64 bit 14.2 only has 2.8.
I spent some time trying to remedy that problem before giving up.

[Dan C]

Post by Dan C
I take it you've never heard of slackbuilds.org ?

Of course I have. Audacity doesn't make because it wants 3.0 of
wxWdigets (or something like that) and 64 bit 14.2 only has 2.8.
I spent some time trying to remedy that problem before giving up.

Strange. The Slackbuilds site shows that dependency, and provides v3.0.4
of wxWidgets right there. I think we have a case of operator error here.
<shrug>

[root]

Post by Dan C
Strange. The Slackbuilds site shows that dependency, and provides v3.0.4
of wxWidgets right there. I think we have a case of operator error here.

Very true. I first tried getting audacity from slonly using slpkg. That
site failed a checksum error. Then I tried to build the package from
Slackbuilds without checking the dependencies. All this was a digression
since I was only interested in seeing how an audio lecture would sound
if sped up. I just didn't want to spin my wheels screwing around with
what might have been a chain of dependencies from Slackbuilds.

Slackbuilds is my last choice since it doesn't list dependencies of
dependencies.

[Rich]

Post by root
Slackbuilds is my last choice since it doesn't list dependencies of
dependencies.

Oh but it does. But you have to go to the page for the dependency to
see the dependencies of that dependency. Which you have to do anyway
if you want to download the scripts via the web interface.

What it does not do is list the full tree of dependencies at the top
level (i.e., at the audicaty page level for example). Yeah, that would
be a useful feature.

[Doug713705]

Le 16-03-2018, Rich nous expliquait dans
alt.os.linux.slackware

Post by Rich

Post by root
Slackbuilds is my last choice since it doesn't list dependencies of
dependencies.

Oh but it does. But you have to go to the page for the dependency to
see the dependencies of that dependency. Which you have to do anyway
if you want to download the scripts via the web interface.
What it does not do is list the full tree of dependencies at the top
level (i.e., at the audicaty page level for example). Yeah, that would
be a useful feature.

This python script of mine does list the whole dependencies chain of a
package:
https://github.com/doug-letough/sbo_deps.py

As sbopkg is also written in python, it *should* be "easy" to integrate
this but I have no time for this.

[root]

Post by Doug713705
Le 16-03-2018, Rich nous expliquait dans
alt.os.linux.slackware

Post by Rich

Post by root
Slackbuilds is my last choice since it doesn't list dependencies of
dependencies.

Oh but it does. But you have to go to the page for the dependency to
see the dependencies of that dependency. Which you have to do anyway
if you want to download the scripts via the web interface.
What it does not do is list the full tree of dependencies at the top
level (i.e., at the audicaty page level for example). Yeah, that would
be a useful feature.

This python script of mine does list the whole dependencies chain of a
https://github.com/doug-letough/sbo_deps.py
As sbopkg is also written in python, it *should* be "easy" to integrate
this but I have no time for this.

I downloaded your script. The options suggest one doesn't need
any further integration since the B option will build all
the necessary stuff?

Since I have long ago waded through ffmpeg, I tried that
just to see (L) how the dependencies were resolved.

I get:
Looking for ffmpeg dependencies.
Please read the README file for the following packages:
- ffmpeg
What next ?
[I] - Install ffmpeg and all its dependencies
[B] - Build ffmpeg (no installation) and all its dependencies
[L] - List ffmpeg dependencies
[A] - Abort

The README file has nothing about ffmpeg.

As I said previously slpkg does what should be done, but there
is a problem, especially using the slonly repository. Slonly
works like Slackware current: it always has the leading edge
stuff. I don't always want to upgrade all previous packages
just to install something new.

There is another problem with dealing with pre-built packages:
the config file isn't always what you might want.

[Doug713705]

Le 16-03-2018, root nous expliquait dans
alt.os.linux.slackware

Post by root

Post by Doug713705
Le 16-03-2018, Rich nous expliquait dans
alt.os.linux.slackware

Post by Rich

Post by root
Slackbuilds is my last choice since it doesn't list dependencies of
dependencies.

Oh but it does. But you have to go to the page for the dependency to
see the dependencies of that dependency. Which you have to do anyway
if you want to download the scripts via the web interface.
What it does not do is list the full tree of dependencies at the top
level (i.e., at the audicaty page level for example). Yeah, that would
be a useful feature.

This python script of mine does list the whole dependencies chain of a
https://github.com/doug-letough/sbo_deps.py
As sbopkg is also written in python, it *should* be "easy" to integrate
this but I have no time for this.

I downloaded your script. The options suggest one doesn't need
any further integration since the B option will build all
the necessary stuff?
Since I have long ago waded through ffmpeg, I tried that
just to see (L) how the dependencies were resolved.
Looking for ffmpeg dependencies.
- ffmpeg
What next ?
[I] - Install ffmpeg and all its dependencies
[B] - Build ffmpeg (no installation) and all its dependencies
[L] - List ffmpeg dependencies
[A] - Abort
The README file has nothing about ffmpeg.
As I said previously slpkg does what should be done, but there
is a problem, especially using the slonly repository. Slonly
works like Slackware current: it always has the leading edge
stuff. I don't always want to upgrade all previous packages
just to install something new.
the config file isn't always what you might want.

ffmpeg is an exception that this will NOT handle (yes ffmpeg is a
difficult package to build if you want some specific libraries support).

There is too many possibilities depending on many variables you may
or not enable.

The README file to read is the ffmpeg SBo README, not the script one !

[Doug713705]

Le 16-03-2018, root nous expliquait dans
alt.os.linux.slackware

Post by root

Post by Doug713705
Le 16-03-2018, Rich nous expliquait dans
alt.os.linux.slackware

Post by Rich

Post by root
Slackbuilds is my last choice since it doesn't list dependencies of
dependencies.

Oh but it does. But you have to go to the page for the dependency to
see the dependencies of that dependency. Which you have to do anyway
if you want to download the scripts via the web interface.
What it does not do is list the full tree of dependencies at the top
level (i.e., at the audicaty page level for example). Yeah, that would
be a useful feature.

This python script of mine does list the whole dependencies chain of a
https://github.com/doug-letough/sbo_deps.py
As sbopkg is also written in python, it *should* be "easy" to integrate
this but I have no time for this.

I downloaded your script. The options suggest one doesn't need
any further integration since the B option will build all
the necessary stuff?
Since I have long ago waded through ffmpeg, I tried that
just to see (L) how the dependencies were resolved.
Looking for ffmpeg dependencies.
- ffmpeg
What next ?
[I] - Install ffmpeg and all its dependencies
[B] - Build ffmpeg (no installation) and all its dependencies
[L] - List ffmpeg dependencies
[A] - Abort
The README file has nothing about ffmpeg.
As I said previously slpkg does what should be done, but there
is a problem, especially using the slonly repository. Slonly
works like Slackware current: it always has the leading edge
stuff. I don't always want to upgrade all previous packages
just to install something new.
the config file isn't always what you might want.

ffmpeg is an exception that this script will NOT handle (yes ffmpeg is a
difficult package to build if you want some specific libraries support).

There is too many possibilities depending on many variables you may
or not enable.

The README file to read is the ffmpeg SBo README, not the script one !

[Peter Chant]

Post by Doug713705
This python script of mine does list the whole dependencies chain of a
https://github.com/doug-letough/sbo_deps.py

Should have read futher down the thread!

Post by Doug713705
As sbopkg is also written in python, it *should* be "easy" to integrate
this but I have no time for this.

[root]

Post by Rich

Post by root
Slackbuilds is my last choice since it doesn't list dependencies of
dependencies.

Oh but it does. But you have to go to the page for the dependency to
see the dependencies of that dependency. Which you have to do anyway
if you want to download the scripts via the web interface.

Tracing out the dependencies is a terrible nuisance. Once you have
done that you have to build them from the bottom up: first you
have to build all the elements which have no further dependencies.
Then you work up the stack to, finally, build the item you started
out to get.

Post by Rich
What it does not do is list the full tree of dependencies at the top
level (i.e., at the audicaty page level for example). Yeah, that would
be a useful feature.

It is a real pain to trace the dependencies for something like ffmpeg.
In most cases slpkg does the job for me. However, in this case
I was put astraddle (I think that is a word) of two version of
qt stuff. Before trying to fetch audacity, I tried an update of
calibre. This involved updating some of the qt stuff. That
broke the requirements for Slackbuilds audacity. Trying to
fix that broke calibre.

Today I ripped out anything to do with audacity, and deleted
the calibre from slonly. I fetched a working version of calibre
from Alien. Slonly offers Calibre 3 while Alien offers
Calibre 2. I am guessing that has to do with different
versions of Python. My slackware still runs Python 2.

[Rich]

Post by root

Post by Rich

Post by root
Slackbuilds is my last choice since it doesn't list dependencies of
dependencies.

Oh but it does. But you have to go to the page for the dependency
to see the dependencies of that dependency. Which you have to do
anyway if you want to download the scripts via the web interface.

Tracing out the dependencies is a terrible nuisance. Once you have
done that you have to build them from the bottom up: first you have
to build all the elements which have no further dependencies. Then
you work up the stack to, finally, build the item you started out to
get.

Oh, I don't disagree. It is a royal pain. I've done it myself many
times.

It 'is' something that is almost trivial for a computer to do.

But none of us has come forward with the code to instruct the computer
to do it, so it is still an undone feature.

Post by root

Post by Rich
What it does not do is list the full tree of dependencies at the top
level (i.e., at the audicaty page level for example). Yeah, that
would be a useful feature.

It is a real pain to trace the dependencies for something like
ffmpeg. In most cases slpkg does the job for me. However, in this
case I was put astraddle (I think that is a word) of two version of
qt stuff. Before trying to fetch audacity, I tried an update of
calibre. This involved updating some of the qt stuff. That broke
the requirements for Slackbuilds audacity. Trying to fix that broke
calibre.

Well, short of the software having enough AI to realize that package X
needs qt-vX and package Y needs qt-vY and to install qt-vX and qt-vY in
separate locations, there's not much that could be done here (besides
the software aborting with a message that X needs Y but you have Z
which is used by Q and so upgrading Z to Y will break Q).

Post by root
Today I ripped out anything to do with audacity, and deleted the
calibre from slonly. I fetched a working version of calibre from
Alien. Slonly offers Calibre 3 while Alien offers Calibre 2. I am
guessing that has to do with different versions of Python. My
slackware still runs Python 2.

Yeah, that's Slack's standard python. Python v3 is available in
SlackBuilds (and will install alongside v2), but it is not already on
the install media from the start.

[Doug713705]

Le 16-03-2018, Rich nous expliquait dans
alt.os.linux.slackware

Post by Rich

Post by root
Today I ripped out anything to do with audacity, and deleted the
calibre from slonly. I fetched a working version of calibre from
Alien. Slonly offers Calibre 3 while Alien offers Calibre 2. I am
guessing that has to do with different versions of Python. My
slackware still runs Python 2.

Yeah, that's Slack's standard python. Python v3 is available in
SlackBuilds (and will install alongside v2), but it is not already on
the install media from the start.

Note that you can install python3 alongside Python2 without any
nuisance.

Apps using python2 will still use it while Python3 apps will use
Python3.

If you need to install a specific module from py2, you can use pipi
(available via SBo):
# pip install <module name>

For a python3 module, use pip3 (provided by python3, IIRC)
# pip3 install <module name>

[root]

Post by Doug713705
Note that you can install python3 alongside Python2 without any
nuisance.
Apps using python2 will still use it while Python3 apps will use
Python3.

That's good news because I have been afraid of P3 since I have
a suite of R,Numpy,and such programs.

Thanks.

[Doug713705]

Le 16-03-2018, Rich nous expliquait dans
alt.os.linux.slackware

Post by Rich

Post by root
Today I ripped out anything to do with audacity, and deleted the
calibre from slonly. I fetched a working version of calibre from
Alien. Slonly offers Calibre 3 while Alien offers Calibre 2. I am
guessing that has to do with different versions of Python. My
slackware still runs Python 2.

Yeah, that's Slack's standard python. Python v3 is available in
SlackBuilds (and will install alongside v2), but it is not already on
the install media from the start.

Note that you can install python3 alongside Python2 without any
nuisance.

Apps using python2 will still use it while Python3 apps will use
Python3.

If you need to install a specific module from py2, you can use pip
(available via SBo):
# pip install <module name>

For a python3 module, use pip3 (provided by python3, IIRC)
# pip3 install <module name>

[Peter Chant]

Post by Rich
What it does not do is list the full tree of dependencies at the top
level (i.e., at the audicaty page level for example). Yeah, that would
be a useful feature.

I can't remember where it came from, but:
sbodeps -i <package name>

Builds the dependencies and then the package. Very nice.

Pete

[Dan C]

Post by root

Post by Dan C
Strange. The Slackbuilds site shows that dependency, and provides
v3.0.4 of wxWidgets right there. I think we have a case of operator
error here.

Very true. I first tried getting audacity from slonly using slpkg. That
site failed a checksum error. Then I tried to build the package from
Slackbuilds without checking the dependencies. All this was a digression
since I was only interested in seeing how an audio lecture would sound
if sped up. I just didn't want to spin my wheels screwing around with
what might have been a chain of dependencies from Slackbuilds.

In other words, you're lazy and impatient.

Post by root
Slackbuilds is my last choice since it doesn't list dependencies of
dependencies.

As others have already told you, it does, indirectly. Very simple to do,
really.

You don't seem like a true Slacker to me.

[Peter "Shaggy" Haywood]

Groovy hepcat root was jivin' in alt.os.linux.slackware on Sun, 11 Mar
2018 12:22 pm. It's a cool scene! Dig it.

Post by root
Thanks for responding. I should have mentioned that mplayer allows
for speed-up but the tone is shifted up so that the voices are
as if one had been breathing helium. I would like normal voices.

Try it with -af scaletempo=scale=2 (to play back at twice the speed,
for example - change the 2 to whatever speed factor you want).

[root]

Post by Peter "Shaggy" Haywood
Groovy hepcat root was jivin' in alt.os.linux.slackware on Sun, 11 Mar
2018 12:22 pm. It's a cool scene! Dig it.

Post by root
Thanks for responding. I should have mentioned that mplayer allows
for speed-up but the tone is shifted up so that the voices are
as if one had been breathing helium. I would like normal voices.

Try it with -af scaletempo=scale=2 (to play back at twice the speed,
for example - change the 2 to whatever speed factor you want).

I couldn't get your suggestion to work. I tried this from man mplayer:
-speed <0.01-100>
Slow down or speed up playback by the factor given as parameter.
Not guaranteed to work correctly with -oac copy. Add -af
scaletempo to get past the 4x limit on playback.

so I had -af scaletempo=2 which would not start. -af scaletempo 2 had no
effect.

There is an additional mention of scaletempo in man mplayer, but its
usage wasnt't immediately clear.

Although mplayer is my default video player, it is not important
to my playback of audio lectures. The sox suggestion is perfect
for my application.

BTW, I find playback of spoken English perfectly comprehensible
at tempo 1.3

[Eli the Bearded]

Post by root
I would like to playback some spoken mp3 files to speed the rate
in order to determine the limit of intelligibility. Is there
some linux capability to achieve this end?

The mplayer tool can speed up and slow down playback of both audio
and video files. In the default keybindings [ and ] are the slow-down
and speed-up controls, changing the speed by ten percent each time.
So ] = 110% original, ]] = 121% original, ]]] 133%, ]]]] 146%, etc.

Relevant input.conf entries:

[ speed_mult 0.9091 # scale playback speed
] speed_mult 1.1
{ speed_mult 0.5
} speed_mult 2.0
BS speed_set 1.0 # reset speed to normal


Elijah
------
speeds up a lot of youtube how-to videos (downloaded with youtube-dl)

[root]

Post by Eli the Bearded

Post by root
I would like to playback some spoken mp3 files to speed the rate
in order to determine the limit of intelligibility. Is there
some linux capability to achieve this end?

The mplayer tool can speed up and slow down playback of both audio
and video files. In the default keybindings [ and ] are the slow-down
and speed-up controls, changing the speed by ten percent each time.
So ] = 110% original, ]] = 121% original, ]]] 133%, ]]]] 146%, etc.
[ speed_mult 0.9091 # scale playback speed
] speed_mult 1.1
{ speed_mult 0.5
} speed_mult 2.0
BS speed_set 1.0 # reset speed to normal
Elijah
------
speeds up a lot of youtube how-to videos (downloaded with youtube-dl)

The speaker's voice is barely intelligible at 1.46 speed up, but I think
it is because the tonal range changes along with the speed.

I was moved to look into this because I have a recording of someone
reading one of those disclaimer messages but at a very much increased
talking speed. The tonal range of the voice is normal, but there is
little or no dead space between words and the words seem sped up
as well. With concentration I can keep up with the speed, but I
don't know if I could do it for any length of time.

[Rich]

Post by root
The speaker's voice is barely intelligible at 1.46 speed up, but I
think it is because the tonal range changes along with the speed.

Speeding up playback, without also changing the tone to get 'donald
duck' voices, is a much more complicated problem.

Post by root
I was moved to look into this because I have a recording of someone
reading one of those disclaimer messages but at a very much increased
talking speed. The tonal range of the voice is normal, but there is
little or no dead space between words and the words seem sped up as
well. With concentration I can keep up with the speed, but I don't
know if I could do it for any length of time.

Those disclaimers are often created by a select group of folks who
actually talk that fast rather than by speeding up a recording of a
normal rate speaker. Which is likely why the tonal range is normal.
It is not a sped up recording but rather a 'very fast talker'.

You might look into sox and it's "speed" option. The sox man page
indicates it changes both pitch and tempo.

Sox also has a "tempo" filter listed in the manpage that claims to
change the playback speed but not pitch of an audio file.

[root]

Post by Rich

Post by root
The speaker's voice is barely intelligible at 1.46 speed up, but I
think it is because the tonal range changes along with the speed.

Speeding up playback, without also changing the tone to get 'donald
duck' voices, is a much more complicated problem.

Post by root
I was moved to look into this because I have a recording of someone
reading one of those disclaimer messages but at a very much increased
talking speed. The tonal range of the voice is normal, but there is
little or no dead space between words and the words seem sped up as
well. With concentration I can keep up with the speed, but I don't
know if I could do it for any length of time.

Those disclaimers are often created by a select group of folks who
actually talk that fast rather than by speeding up a recording of a
normal rate speaker. Which is likely why the tonal range is normal.
It is not a sped up recording but rather a 'very fast talker'.
You might look into sox and it's "speed" option. The sox man page
indicates it changes both pitch and tempo.
Sox also has a "tempo" filter listed in the manpage that claims to
change the playback speed but not pitch of an audio file.

Perfect: sox with tempo setting does exactly what I want.

My version of sox cannot generate mp3 so I first convert
to .wav:

sox infile.mp3 test.wav tempo 1.6 for example.
Tone of voice is natural, speed is increased. Sibilants are enhanced
somewhat.

My first impression is that we are used to having time to reflect upon
what has just been said. A speed increase of 1.6 does not give me
time to either reflect or absorb. It is as if I get mentally
exhausted very quickly.

A tempo increase of 1.3 is very much easier to comprehend. I am
going to experiment with this for a while.

Thank you very much for this solution Rich.

[Rich]

Post by root

Post by Rich

Post by root
I was moved to look into this because I have a recording of someone
reading one of those disclaimer messages but at a very much
increased talking speed. The tonal range of the voice is normal,
but there is little or no dead space between words and the words
seem sped up as well. With concentration I can keep up with the
speed, but I don't know if I could do it for any length of time.

Those disclaimers are often created by a select group of folks who
actually talk that fast rather than by speeding up a recording of a
normal rate speaker. Which is likely why the tonal range is normal.
It is not a sped up recording but rather a 'very fast talker'.
You might look into sox and it's "speed" option. The sox man page
indicates it changes both pitch and tempo.
Sox also has a "tempo" filter listed in the manpage that claims to
change the playback speed but not pitch of an audio file.

Perfect: sox with tempo setting does exactly what I want.
My version of sox cannot generate mp3 so I first convert

If you wish to have a mp3 as the final version, then lame
(http://lame.sourceforge.net/) can convert the wav back into an mp3 for
you.

Post by root
sox infile.mp3 test.wav tempo 1.6 for example.
Tone of voice is natural, speed is increased. Sibilants are enhanced
somewhat.
My first impression is that we are used to having time to reflect
upon what has just been said. A speed increase of 1.6 does not give
me time to either reflect or absorb. It is as if I get mentally
exhausted very quickly.

It may also be that you are not accustomed to listening at such a rate.
Given time listening to the 'sped up' version, you may find you can
keep up.

Post by root
A tempo increase of 1.3 is very much easier to comprehend. I am
going to experiment with this for a while.
Thank you very much for this solution Rich.

You are welcome. Honestly I did not know sox had that filter until
after I searched through its manpage. I looked there because "sox does
a lot of things" and thought if there was something somewhere, it might
be there.

[Robert Riches]

Post by Rich

Post by root
The speaker's voice is barely intelligible at 1.46 speed up, but I
think it is because the tonal range changes along with the speed.

Speeding up playback, without also changing the tone to get 'donald
duck' voices, is a much more complicated problem.

Post by root
I was moved to look into this because I have a recording of someone
reading one of those disclaimer messages but at a very much increased
talking speed. The tonal range of the voice is normal, but there is
little or no dead space between words and the words seem sped up as
well. With concentration I can keep up with the speed, but I don't
know if I could do it for any length of time.

Those disclaimers are often created by a select group of folks who
actually talk that fast rather than by speeding up a recording of a
normal rate speaker. Which is likely why the tonal range is normal.
It is not a sped up recording but rather a 'very fast talker'.
You might look into sox and it's "speed" option. The sox man page
indicates it changes both pitch and tempo.
Sox also has a "tempo" filter listed in the manpage that claims to
change the playback speed but not pitch of an audio file.

Rich, I'll have to give you the benefit of the doubt that many of
the disclaimers are recorded by voice talent who can actually
talk that fast. However, there do exist methods to speed up or
slow down speech while retaining the original pitch. For one
example, Tascam makes hand-held audio recording devices that do
that. The same thing can undoubtedly be done in software on
general purpose computing devices.

Thanks,

[Rich]

Post by Robert Riches

Post by Rich

Post by root
The speaker's voice is barely intelligible at 1.46 speed up, but I
think it is because the tonal range changes along with the speed.

Speeding up playback, without also changing the tone to get 'donald
duck' voices, is a much more complicated problem.

Post by root
I was moved to look into this because I have a recording of someone
reading one of those disclaimer messages but at a very much increased
talking speed. The tonal range of the voice is normal, but there is
little or no dead space between words and the words seem sped up as
well. With concentration I can keep up with the speed, but I don't
know if I could do it for any length of time.

Those disclaimers are often created by a select group of folks who
actually talk that fast rather than by speeding up a recording of a
normal rate speaker. Which is likely why the tonal range is normal.
It is not a sped up recording but rather a 'very fast talker'.
You might look into sox and it's "speed" option. The sox man page
indicates it changes both pitch and tempo.
Sox also has a "tempo" filter listed in the manpage that claims to
change the playback speed but not pitch of an audio file.

Rich, I'll have to give you the benefit of the doubt that many of
the disclaimers are recorded by voice talent who can actually
talk that fast. However, there do exist methods to speed up or
slow down speech while retaining the original pitch. For one
example, Tascam makes hand-held audio recording devices that do
that. The same thing can undoubtedly be done in software on
general purpose computing devices.

In another branch of this thread I found and referred the OP to the
'tempo' filter in sox, which does do a speedup without changing pitch
(presumably as best it can do).

So, yes, there are ways to do it without having to hire one of the
"fast talkers".

[Rich]

Post by root
I would like to playback some spoken mp3 files to speed the rate
in order to determine the limit of intelligibility. Is there
some linux capability to achieve this end?

mplayer has a speedup/slowdown option, and it will play plain mp3's,
and it's speedup/slowdown option works for mp3's (just tested). You do
get 'donald duck' voice however.

You get 10% steps with it's fine step, and 2x/4x with its large step.