Looking for info on VST/LADSPA/Nyquist plug-ins.

[Wes Groleau]

I'm not an audio guy normally, but I do have a particular audio task
that I would like to do more quickly than my current find, select, do,
do, do, find, select, etc. (Involves changing lengths of silent parts.)

From browsing subject lines, it seems this is not the group to ask.

What would be a good group in which to ask for pointers to a plug-in
that does this or that to audio files?

thanks

--
Wes Groleau

Ostracism: A practice of sticking your head in the sand.

[Phil W[_3_]]

Wes Groleau:

> I'm not an audio guy normally, but I do have a particular audio task
> that I would like to do more quickly than my current find, select, do,
> do, do, find, select, etc. (Involves changing lengths of silent parts.)
>
> From browsing subject lines, it seems this is not the group to ask.
>
> What would be a good group in which to ask for pointers to a plug-in
> that does this or that to audio files?

Not a group, but try:
http://www.kvraudio.com/q.php

KVR is a "database", where most plug-ins are listed.


Good luck, finding whatever you are looking for. ;-)

Phil

[Wes Groleau]

On 11-23-2013, 18:54, Phil W wrote:
> Wes Groleau:
>
>> I'm not an audio guy normally, but I do have a particular audio task
>> that I would like to do more quickly than my current find, select, do,
>> do, do, find, select, etc. (Involves changing lengths of silent parts.)
>>
>> From browsing subject lines, it seems this is not the group to ask.
>>
>> What would be a good group in which to ask for pointers to a plug-in
>> that does this or that to audio files?
>
> Not a group, but try:
> http://www.kvraudio.com/q.php
>
> KVR is a "database", where most plug-ins are listed.

:-) I was hoping no one would say that. I don't know enough to figure
out what to put in their daunting search engine.

--
Wes Groleau

“Missing a train is only painful if you run after it!”
— Nassim Nicholas Taleb

[Wes Groleau]

On 11-23-2013, 19:33, Mike Rivers wrote:
> On 11/23/2013 6:21 PM, Wes Groleau wrote:
>> I'm not an audio guy normally, but I do have a particular audio task
>> that I would like to do more quickly than my current find, select, do,
>> do, do, find, select, etc. (Involves changing lengths of silent parts.)
>
>> From browsing subject lines, it seems this is not the group to ask.
>
> Maybe yes, maybe no. What is it that you're trying to do, do, do, find,
> and select? Or more simply, what is it that you're trying to measure?
> Nyquist did a lot of stuff and his name is associated with a number of
> theories and processes.

It is also associated with a programming language for making plug-ins to
audio processing programs. So it still looks like I picked the wrong group.

But since you asked,

I understand a lot about physics of sound from my prior career in sonar,
but I know nothing at all about they way people do things in the
performance audio world.

I want to do some time editing of spoken word with minimal affect on the
quality. These will be mono tracks.

Too much shortening or lengthening can make it sound less
natural--people who talk slower lengthen vowels more than consonants. At
even higher amounts, harmonics are affect making even weirder sounds.

To minimize this, plus make the next step easier, I would like to affect
only "silences" in step one. I'll describe what I would code if I were
on my own, but I'm amenable to using something already written if it
comes close enough to the over-all goal.

1. Find all sound sections that are effectively silent, i.e., nothing
but noise and very little of that, and that are longer than some number
of milliseconds.

2. Make them completely silent with a very brief fade out/in

3. Shorten or lengthen them to make the entire selection a
user-specified duration, except that if that if this makes a particular
silence longer than a set maximum or shorter than a set minimum, use
that max/min length. This is to avoid turning two words into one or
having unnatural pauses between phrases.

The limits may prevent reaching the desired length, but by getting that
much closer, I reduce the amount of change to the non-silent parts in
the next step.

I read that there are "thousands of plugins" out there, but web searches
seem to be finding lists that are more like "dozens" and they seem to be
heavily weighted for musical effects.

I would prefer Nyquist, so that I can tweak it without having to
separately compile. String two or more in sequence is OK if one plugin
doesn't do it all.

Suggestions?

And then after the above gets as close as feasible,

Looking for a plug-in, preferably Nyquist, that will do something like
the "Elastic Audio" in Pro Tools.

I have been manually lengthening/shortening the duration of sound
sections one at a time to make two tracks synch better with each other.
Elastic audio apparently lets you mark all the synch points, and then it
does all at once what I've been doing a few seconds at a time.


--
Wes Groleau

From the cowardice that shrinks from new truth,
From the laziness that is content with half-truths,
From the arrogance that thinks it knows all truth,
O God of Truth, deliver us.
--Leslie Dixon Weatherhead
--Rabbi Mordechai M. Kaplan
--ancient prayer
--unknown
--(no attempt at attribution)
(a thousand thanks to someone who can tell me who
really wrote it AND persuade me they're not making it up!)

[Peter Larsen[_3_]]

Hi Wes,

Wes Groleau wrote:

> On 11-23-2013, 19:33, Mike Rivers wrote:
>> On 11/23/2013 6:21 PM, Wes Groleau wrote:

>>> I'm not an audio guy normally, but I do have a particular audio task
>>> that I would like to do more quickly than my current find, select,
>>> do, do, do, find, select, etc. (Involves changing lengths of
>>> silent parts.)

How large a change of duration is needed, I think you are making something
simple complicated. Some audio editing software allows you to change the
duration of a piece without changing the pitch, that would constitute a
simple scaling. In my preconceived opion that is what is most likely to be
least audible within the range you are likely to get away with, ie. some 5
to at most 10 percent alteration of duration.

I already have Audition 3, so that would be my choice, but I reckon the
competing software will have similar options. Adobe tends to offer
demo-versions, dunno if they still do, but that is the way to find out what
is possible. It is btw. good practice never to install a demo version on a
box that you want to install a full version on later ...

Kind regards

Peter Larsen

[Mike Rivers[_2_]]

On 11/24/2013 1:30 AM, Wes Groleau wrote:

> It is also associated with a programming language for making plug-ins to
> audio processing programs.

Oh, that Nyquist. Most of us here aren't programmers. Since it's one of
the languages that people use to create synthesized sounds, you might
find some kindred spirits in the rec.music.making.synth newsgroup,
though I haven't visited there in quite some time. Another place to look
might be among the community of Audacity users. I know that there's some
integration between Audacity and Nyquist programming (there are some
ready-made Nyquist plug-ins available). Try

http://forum.audacityteam.org/viewforum.php?f=39

You might find someone working in Nyquist there who can get you started
on your quest.

> I would like to affect
> only "silences" in step one. I'll describe what I would code if I were
> on my own, but I'm amenable to using something already written if it
> comes close enough to the over-all goal.
>
> 1. Find all sound sections that are effectively silent, i.e., nothing
> but noise and very little of that, and that are longer than some number
> of milliseconds.
>
> 2. Make them completely silent with a very brief fade out/in

That much could be done with a gate, and there are plenty of gate
plug-ins available, mostly as part of a dynamics plug-in
(compressor/limiter/expander). I trust you've done some of this manually
and have a sense of how it sounds, but in my experience, dead silence
between words sounds pretty unnatural.

> 3. Shorten or lengthen them to make the entire selection a
> user-specified duration,

When I'm editing speech, I just chop out what I don't want, and if a
longer pause is called for, grab a virtual bit of tape off the virtual
cutting room floor and stick it in, then trim it to a natural length.
But I see no reason why you couldn't time-stretch a clean pause after
you've removed the stray noises that are identifiable as something other
than ambient noise.

Happy hunting.

--
For a good time, visit http://mikeriversaudio.wordpress.com

[Peter Larsen[_3_]]

Mike Rivers wrote:

> When I'm editing speech, I just chop out what I don't want, and if a
> longer pause is called for, grab a virtual bit of tape off the virtual
> cutting room floor and stick it in, then trim it to a natural length.
> But I see no reason why you couldn't time-stretch a clean pause after
> you've removed the stray noises that are identifiable as something
> other than ambient noise.

I read this question to be about bulk processing.

> Happy hunting.

Kind regards

Peter Larsen

[Wes Groleau]

On 11-24-2013, 02:22, Peter Larsen wrote:
> Wes Groleau wrote:
>> On 11-23-2013, 19:33, Mike Rivers wrote:
>>> On 11/23/2013 6:21 PM, Wes Groleau wrote:
>>>> I'm not an audio guy normally, but I do have a particular audio task
>>>> that I would like to do more quickly than my current find, select,
>>>> do, do, do, find, select, etc. (Involves changing lengths of
>>>> silent parts.)
>
> How large a change of duration is needed, I think you are making something
> simple complicated. Some audio editing software allows you to change the
> duration of a piece without changing the pitch, that would constitute a

Yes, I know. But as I said, too much of that makes it sound weird.
That's why I said I want to do as much as I can to the pauses first--to
minimize distortion of the voice.

> simple scaling. In my preconceived opion that is what is most likely to be
> least audible within the range you are likely to get away with, ie. some 5
> to at most 10 percent alteration of duration.
>
> I already have Audition 3, so that would be my choice, but I reckon the
> competing software will have similar options. Adobe tends to offer

I am not looking for a software package. I have two, and both can
handle plug-ins.

I can learn XLISP/Nyquist and write my own plug-inif I have to, but with
allegations that "thousands" of them exist, it just made sense
to see whether I could find one.

I could tell right off that this is not the focus of this group, hence
my original question of what group. I am going to take a look in the
groups that someone else suggested.

--
Wes Groleau

Why does everyone call it a “fanny pack" ?
When was the last time you saw one on a fanny?

[Wes Groleau]

On 11-24-2013, 07:15, Mike Rivers wrote:
> On 11/24/2013 1:30 AM, Wes Groleau wrote:
>> It is also associated with a programming language for making plug-ins to
>> audio processing programs.
>
> Oh, that Nyquist. Most of us here aren't programmers. Since it's one of
> the languages that people use to create synthesized sounds, you might
> find some kindred spirits in the rec.music.making.synth newsgroup,
> though I haven't visited there in quite some time. Another place to look
> might be among the community of Audacity users. I know that there's some
> integration between Audacity and Nyquist programming (there are some
> ready-made Nyquist plug-ins available). Try
>
> http://forum.audacityteam.org/viewforum.php?f=39
>
> You might find someone working in Nyquist there who can get you started
> on your quest.

Thanks. I'll check. My favorite program (similar to Audacity) has such
a forum, but there's not a big enough user base to get much info
that way.

>> 2. Make them completely silent with a very brief fade out/in
>
> That much could be done with a gate, and there are plenty of gate
> plug-ins available, mostly as part of a dynamics plug-in
> (compressor/limiter/expander). I trust you've done some of this manually
> and have a sense of how it sounds, but in my experience, dead silence
> between words sounds pretty unnatural.

Yes, it was recorded in professional facilities, so it is already
effectively silent. The fade out/in is an attempt to eliminate any
original editing clicks (which are audible) and flattening the
wave form is intended to prevent creating more when I cut or paste.

> > 3. Shorten or lengthen them to make the entire selection a
> > user-specified duration,
>
> When I'm editing speech, I just chop out what I don't want, and if a
> longer pause is called for, grab a virtual bit of tape off the virtual
> cutting room floor and stick it in, then trim it to a natural length.

Yes, I have done that, even in ancient history with two cassette
machines. But in this case, I am trying to synchronize two recordings
of the same thing in different languages. So that it almost sounds like
simultaneous translation.

> But I see no reason why you couldn't time-stretch a clean pause after
> you've removed the stray noises that are identifiable as something other
> than ambient noise.

Right. I've done it a lot, one pause at a time. Very slow. But it's
such a simple operation, with no sound quality to worry about, that I
figure it can be better automated. By doing it to the silences, I
reduce the amount of time changes I have to make to the voice.


--
Wes Groleau

Why does everyone call it a “fanny pack" ?
When was the last time you saw one on a fanny?

[Wes Groleau]

On 11-24-2013, 07:37, Peter Larsen wrote:
> I read this question to be about bulk processing.

You read correctly. Takes me five seconds or less to do a single pause
manually, which can add up to a lot in a five minute voice sample. And
there will be hundreds of files of one to twenty minutes.


--
Wes Groleau

Why does everyone call it a “fanny pack" ?
When was the last time you saw one on a fanny?

[jason]

On Sat, 23 Nov 2013 18:21:33 -0500 "Wes Groleau"
> wrote in article <l6rd9v$ubg$1@dont-
email.me>
>
> I'm not an audio guy normally, but I do have a particular audio task
> that I would like to do more quickly than my current find, select, do,
> do, do, find, select, etc. (Involves changing lengths of silent parts.)
>
> From browsing subject lines, it seems this is not the group to ask.
>
> What would be a good group in which to ask for pointers to a plug-in
> that does this or that to audio files?
>
> thanks

If you are looking to delete silent portions of the audio, Audition can
do it.

[Wes Groleau]

On 11-24-2013, 13:46, Jason wrote:
> On Sat, 23 Nov 2013 18:21:33 -0500 "Wes Groleau"
> If you are looking to delete silent portions of the audio, Audition can
> do it.

I am looking to _modify_ silent portions in the manner I described,
which included doing it with a plug-in for the software I have instead
of spending a bushel of bucks on another system.

--
Wes Groleau

Why does everyone call it a “fanny pack" ?
When was the last time you saw one on a fanny?

[Scott Dorsey]

In article >, Wes Groleau > wrote:
>On 11-24-2013, 02:22, Peter Larsen wrote:
>> Wes Groleau wrote:
>>> On 11-23-2013, 19:33, Mike Rivers wrote:
>>>> On 11/23/2013 6:21 PM, Wes Groleau wrote:
>>>>> I'm not an audio guy normally, but I do have a particular audio task
>>>>> that I would like to do more quickly than my current find, select,
>>>>> do, do, do, find, select, etc. (Involves changing lengths of
>>>>> silent parts.)
>>
>> How large a change of duration is needed, I think you are making something
>> simple complicated. Some audio editing software allows you to change the
>> duration of a piece without changing the pitch, that would constitute a
>
>Yes, I know. But as I said, too much of that makes it sound weird.
>That's why I said I want to do as much as I can to the pauses first--to
>minimize distortion of the voice.

How much timesquish are you trying to do? Reducing pauses can work but
it can also make for a very unnatural effect if you do too much of it.

There are commercial hardware devices intended for the broadcast industry
for timesquishing. I don't know of a plugin, but it may exist out there
with the radio applications.
--scott


--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Peter Larsen[_3_]]

Wes Groleau wrote:

> On 11-24-2013, 13:46, Jason wrote:
>> On Sat, 23 Nov 2013 18:21:33 -0500 "Wes Groleau"
>> If you are looking to delete silent portions of the audio, Audition
>> can do it.

> I am looking to _modify_ silent portions in the manner I described,

You thinking is broken, you assume that the short breaks between the words
are not speech. They are.

Scaling as I suggested with complete method can probably give you 5 to 10
percent shortening at the cost of unnatural sound but with pitch maintained.

Kind regards

Peter Larsen

[jason]

On Mon, 25 Nov 2013 03:36:20 +0100 "Peter Larsen" >
wrote in article >
>
> Wes Groleau wrote:
>
> > On 11-24-2013, 13:46, Jason wrote:
> >> On Sat, 23 Nov 2013 18:21:33 -0500 "Wes Groleau"
> >> If you are looking to delete silent portions of the audio, Audition
> >> can do it.
>
> > I am looking to _modify_ silent portions in the manner I described,
>
> You thinking is broken, you assume that the short breaks between the words
> are not speech. They are.
>
> Scaling as I suggested with complete method can probably give you 5 to 10
> percent shortening at the cost of unnatural sound but with pitch maintained.
>
> Kind regards
>
> Peter Larsen

Gaps in speech are not simple. If I listen to someone speaking
a language I do not understand, I hear NO gaps between words. With
English, I do. This is apparently common. So the gaps may not be
gaps at all but the result of ear/brain processing instead.

[Wes Groleau]

On 11-24-2013, 16:00, Scott Dorsey wrote:
> How much timesquish are you trying to do? Reducing pauses can work but
> it can also make for a very unnatural effect if you do too much of it.

That's why I said I would have limits, not to be shortened or lengthened
outside of a certain range. Avoiding unnatural effect
is part of the goal.

--
Wes Groleau

Is it an on-line compliment to call someone a Net Wit ?

[Wes Groleau]

On 11-24-2013, 21:36, Peter Larsen wrote:
> Wes Groleau wrote:
>> On 11-24-2013, 13:46, Jason wrote:
>>> On Sat, 23 Nov 2013 18:21:33 -0500 "Wes Groleau"
>>> If you are looking to delete silent portions of the audio, Audition
>>> can do it.
>
>> I am looking to _modify_ silent portions in the manner I described,
>
> You thinking is broken, you assume that the short breaks between the words
> are not speech. They are.

As I stated, I would not touch gaps that are sorter than some threshold,
and I would not allow gaps to be shortened below some threshold.

> Scaling as I suggested with complete method can probably give you 5 to 10
> percent shortening at the cost of unnatural sound but with pitch maintained.

I have done both manually often. As I stated, I am trying to avoid
"unnatural sound." By doing more of the compression/stretching in the
silences, I can reduce the amount done to the voice, thereby reducing
unnatural sound.

I'm sure you guys know a heck of a lot more than I do about this sort of
thing. But I DO know what I've seen and heard and done. I am just
looking for more automation of a tedious process that I have already
verified to work.

--
Wes Groleau

If you put garbage in a computer nothing comes out but garbage.
But this garbage, having passed through a very expensive machine,
is somehow ennobled and none dare criticize it.

[Scott Dorsey]

In article >, Wes Groleau > wrote:
>On 11-24-2013, 16:00, Scott Dorsey wrote:
>> How much timesquish are you trying to do? Reducing pauses can work but
>> it can also make for a very unnatural effect if you do too much of it.
>
>That's why I said I would have limits, not to be shortened or lengthened
>outside of a certain range. Avoiding unnatural effect
>is part of the goal.

How much timesquish are you trying to do?
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Peter Larsen[_3_]]

Scott Dorsey wrote:

> In article >, Wes Groleau
> > wrote:

>> On 11-24-2013, 16:00, Scott Dorsey wrote:
>>> How much timesquish are you trying to do? Reducing pauses can work
>>> but it can also make for a very unnatural effect if you do too much
>>> of it.

>> That's why I said I would have limits, not to be shortened or
>> lengthened outside of a certain range. Avoiding unnatural effect
>> is part of the goal.

> How much timesquish are you trying to do?

And: to solve what problem?

> --scott

Kind regards

Peter Larsen

[S. King]

On Mon, 25 Nov 2013 00:05:41 -0500, Wes Groleau wrote:

> On 11-24-2013, 21:36, Peter Larsen wrote:
>> Wes Groleau wrote:
>>> On 11-24-2013, 13:46, Jason wrote:
>>>> On Sat, 23 Nov 2013 18:21:33 -0500 "Wes Groleau"
>>>> If you are looking to delete silent portions of the audio, Audition
>>>> can do it.
>>
>>> I am looking to _modify_ silent portions in the manner I described,
>>
>> You thinking is broken, you assume that the short breaks between the
>> words are not speech. They are.
>
> As I stated, I would not touch gaps that are sorter than some threshold,
> and I would not allow gaps to be shortened below some threshold.
>
>> Scaling as I suggested with complete method can probably give you 5 to
>> 10 percent shortening at the cost of unnatural sound but with pitch
>> maintained.
>
> I have done both manually often. As I stated, I am trying to avoid
> "unnatural sound." By doing more of the compression/stretching in the
> silences, I can reduce the amount done to the voice, thereby reducing
> unnatural sound.
>
> I'm sure you guys know a heck of a lot more than I do about this sort of
> thing. But I DO know what I've seen and heard and done. I am just
> looking for more automation of a tedious process that I have already
> verified to work.

I know that you are looking for a VST plug-in; however, I just did a Google
search for Cool Edit 2000, the predecessor program to Adobe Audition,
which is now available free as a download from many web sites, and it does
have the "Delete Silence" function. I did not check to see if the
adjustable parameters available in Audition 3 are available in Cool Edit
2000, but I suspect that they are. The adjustable functions in A3 include
definition of silence in terms of level and duration and minimum length of
remaining silence after automated cutting operations. In 'general
preferences' one can set the fade-out/fade-in time for cuts. That seems
to cover what you want to do. Probably for free. Although not VST/
Nyquist. For what it is worth I tried the function in A3 and it works a
treat.

Steve King

[Wes Groleau]

On 11-25-2013, 10:25, Peter Larsen wrote:
> Scott Dorsey wrote:
>> How much timesquish are you trying to do?
>
> And: to solve what problem?

On 11-24-2013, 01:30, Wes Groleau wrote:
> I have been manually lengthening/shortening the duration of sound
> sections one at a time to make two tracks synch better with each
> other.

On 11-24-2013, 12:20, Wes Groleau wrote:
> .... But in this case, I am trying to synchronize two recordings
> of the same thing in different languages. So that it almost sounds
> like simultaneous translation.

So, of course, the lengths vary. The last one I worked on was the worst
ever: 03:22.451 in Chinese and 01:42.292 in Italian.

--
Wes Groleau

Expert, n.:
Someone who comes from out of town and shows slides.

[Wes Groleau]

On 11-25-2013, 11:15, S. King wrote:
> I know that you are looking for a VST plug-in; however, I just did a Google

Actually, I would prefer Nyquist to VST, because it's easier for me to
edit if I need to.

BUT ...

> search for Cool Edit 2000, the predecessor program to Adobe Audition,
> which is now available free as a download from many web sites, and it does
> have the "Delete Silence" function. I did not check to see if the
> adjustable parameters available in Audition 3 are available in Cool Edit
> 2000, but I suspect that they are. The adjustable functions in A3 include
> definition of silence in terms of level and duration and minimum length of
> remaining silence after automated cutting operations. In 'general
> preferences' one can set the fade-out/fade-in time for cuts. That seems
> to cover what you want to do. Probably for free. Although not VST/
> Nyquist. For what it is worth I tried the function in A3 and it works a
> treat.

.... this certainly sounds worth checking out!! Thank you!

--
Wes Groleau

Expert, n.:
Someone who comes from out of town and shows slides.

[James Perrett[_4_]]

On Sun, 24 Nov 2013 20:00:07 -0000, Wes Groleau
> wrote:

> On 11-24-2013, 13:46, Jason wrote:
>> On Sat, 23 Nov 2013 18:21:33 -0500 "Wes Groleau"
>> If you are looking to delete silent portions of the audio, Audition can
>> do it.
>
> I am looking to _modify_ silent portions in the manner I described,
> which included doing it with a plug-in for the software I have instead
> of spending a bushel of bucks on another system.
>

The main problem with plug-ins is that they're not often used where you
want the length of the audio to change. Most plug-in systems assume that
the length of the audio will stay the same. That's why operations like
this have to be done in the host software.

James.
--
JRP Music - Audio restoration and mastering - http://www.jrpmusic.net

[Les Cargill[_4_]]

James Perrett wrote:
> On Sun, 24 Nov 2013 20:00:07 -0000, Wes Groleau
> > wrote:
>
>> On 11-24-2013, 13:46, Jason wrote:
>>> On Sat, 23 Nov 2013 18:21:33 -0500 "Wes Groleau"
>>> If you are looking to delete silent portions of the audio, Audition can
>>> do it.
>>
>> I am looking to _modify_ silent portions in the manner I described,
>> which included doing it with a plug-in for the software I have instead
>> of spending a bushel of bucks on another system.
>>
>
> The main problem with plug-ins is that they're not often used where you
> want the length of the audio to change.

Very much so.

> Most plug-in systems assume that
> the length of the audio will stay the same. That's why operations like
> this have to be done in the host software.
>
> James.

Wait. I'm confused.

So there are two kinds of DAW software. One consumes other audio files
and transmogrifies them into something else into another audio
file. These have variably been called "sequencers" ( they started out
when com-pu-ters were not powerful enough to do much damage ). Protools,
Cakewalk-and-descendents, Reaper & even lowly n-Track are like this.

The second class are "wave editors" like the CoolEdits and such.
These, you were presumed to want to do permanent damage to the original
sound vectors (although you could always save to a new file name ).

So for plugins ( as in VST plugins), those are deployed within
the first group of DAW. They can't really alter the number of samples
represented by the source material, at least not in any way I am aware
of. They're all about "one sample in , one sample out." They might add
something like a reverb tail to the end but can't *inherently* do any
shortening.

"The host software" is the thing that you add plugins to. So it's
just as constrained by number-of-samples as the plugins.

Perhaps the second, "destructive" class of software can have plugins;
I'm just ignorant of that possibility.

--
Les Cargill

[Luxey]

Have no nerves to read all the posts, just to say:

Each and every DAW I ever worked on had "remove silence" function.