Re: [racket-users] Playing sound

['John Clements' via Racket Users]

> On Apr 6, 2016, at 4:31 PM, Jordan Johnson  wrote:
> 
> On Apr 6, 2016, at 15:18, John Clements  wrote:
>> 1) The “stop” procedure is a giant end-of-the-world hammer; I wouldn’t 
>> expect it to be useful while a program is running.
> 
> OK. It had that feel.
> 
>> 2) I hesitate to suggest this, but there is a “pstream-set-volume!” function 
>> that can silence a pstream. (that was the caveat.)
> 
> I think this points to a hackish solution that’ll work for us for now:
>   https://gist.github.com/RenaissanceBug/da7027d8c55a1819f9bae274e189ccac
> The only concern I have thought of so far is whether this’ll cause memory 
> leakage over the course of a number of switches between pstreams; I don’t 
> know if there’s anything that will keep the GC from cleaning up the pstream 
> when it becomes unreachable.

Yes, I would expect it to be a problem. I wouldn’t worry so much about memory 
leakage as I would about the sound system gradually grinding to a halt. An 
ongoing pstream will continue to generate data, and sequencer.rkt will continue 
to compute it, multiply it all by zero, and then add it to the out-going 
stream. 

> 
>> 3) You are also correct in your supposition that the interface I proposed is 
>> not implemented. It looks like it could be done quickly, meaning probably on 
>> the order of 10-20 hours of work for me.
> 
> I’m glad I asked rather than just dive in trying to figure it out in the 
> middle of a busy semester.
> 
>> The relevant code to look at here would probably be
>> https://github.com/jbclements/RSound/blob/master/rsound/sequencer.rkt
>> [...]
>> Adding the stop-playing! call would probably involve adding some king of 
>> message queue to the sequencer to receive fade-out messages, and updating 
>> the ‘entry’ structure to contain information about sounds that are currently 
>> fading in or out.
> 
> I’ll look into it at some point, time permitting. But I think the gist above 
> will obviate my need for it, at least for the time being.

Let me know how it goes.

Best,

John Clements



-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] Playing sound

[Jordan Johnson]

On Apr 6, 2016, at 15:18, John Clements  wrote:
> 1) The “stop” procedure is a giant end-of-the-world hammer; I wouldn’t expect 
> it to be useful while a program is running.

OK. It had that feel.

> 2) I hesitate to suggest this, but there is a “pstream-set-volume!” function 
> that can silence a pstream. (that was the caveat.)

I think this points to a hackish solution that’ll work for us for now:
https://gist.github.com/RenaissanceBug/da7027d8c55a1819f9bae274e189ccac 

The only concern I have thought of so far is whether this’ll cause memory 
leakage over the course of a number of switches between pstreams; I don’t know 
if there’s anything that will keep the GC from cleaning up the pstream when it 
becomes unreachable.

> 3) You are also correct in your supposition that the interface I proposed is 
> not implemented. It looks like it could be done quickly, meaning probably on 
> the order of 10-20 hours of work for me.

I’m glad I asked rather than just dive in trying to figure it out in the middle 
of a busy semester.

> The relevant code to look at here would probably be
> https://github.com/jbclements/RSound/blob/master/rsound/sequencer.rkt 
> 
> [...]
> Adding the stop-playing! call would probably involve adding some king of 
> message queue to the sequencer to receive fade-out messages, and updating the 
> ‘entry’ structure to contain information about sounds that are currently 
> fading in or out.

I’ll look into it at some point, time permitting. But I think the gist above 
will obviate my need for it, at least for the time being.

Thanks again!

Best,
jmj

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] Playing sound

['John Clements' via Racket Users]

> On Apr 6, 2016, at 11:26 AM, Jordan Johnson  wrote:
> 
> Hey, so I’m now looking at how to set up background music in rsound with the 
> ability to switch tracks. Below is the suggestion from John Clements that 
> seems right:
>>> I’m guessing you’re looking for an imperative API that looks something like 
>>> this:
>>> 
>>> (start-playing! stream sound at-time #:fade-in time-duration) => token
>>> (stop-playing! stream token at-time #:fade-out time-duration)
> 
> AFAICT this isn’t already in existence in the rsound library. My impression 
> is that we will need to build this on top of signals and networks — but I 
> want to check that, before we pour the time into doing it.
> 
> I’m thinking signals & networks because we tried to build it using pstreams, 
> and ran into a snag when trying to stop playback and switch to a new track:
> 
> ;; Given a pstream bgmusic and rsounds bgsong1, bgsong2, the following 
> Interactions
> ;; occur:
> > (pstream-play bgmusic bgsong1)
> "sound is queued"
> > (stop)
> > (pstream-play bgmusic bgsong2)
> "sound is queued";; BUT NO SOUND PLAYS.
> > 
> ;; - end interactions -
> 
> So, am I correct in thinking that pstreams aren’t sufficient for what I’m 
> attempting? Is there another way to stop playback, other than calling (stop)?

You are correct. with one tiny caveat. 

1) The “stop” procedure is a giant end-of-the-world hammer; I wouldn’t expect 
it to be useful while a program is running.
2) I hesitate to suggest this, but there is a “pstream-set-volume!” function 
that can silence a pstream. (that was the caveat.)
3) You are also correct in your supposition that the interface I proposed is 
not implemented. It looks like it could be done quickly, meaning probably on 
the order of 10-20 hours of work for me. The relevant code to look at here 
would probably be

https://github.com/jbclements/RSound/blob/master/rsound/sequencer.rkt

In fact, the pstream library (in stream-play.rkt) is just a thin skin over the 
sequencer code.

Adding the stop-playing! call would probably involve adding some king of 
message queue to the sequencer to receive fade-out messages, and updating the 
‘entry’ structure to contain information about sounds that are currently fading 
in or out.

John 

> 
> Thanks,
> Jordan
> 



-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] Playing sound

[Jordan Johnson]

Hey, so I’m now looking at how to set up background music in rsound with the 
ability to switch tracks. Below is the suggestion from John Clements that seems 
right:
>> I’m guessing you’re looking for an imperative API that looks something like 
>> this:
>> 
>> (start-playing! stream sound at-time #:fade-in time-duration) => token
>> (stop-playing! stream token at-time #:fade-out time-duration)

AFAICT this isn’t already in existence in the rsound library. My impression is 
that we will need to build this on top of signals and networks — but I want to 
check that, before we pour the time into doing it.

I’m thinking signals & networks because we tried to build it using pstreams, 
and ran into a snag when trying to stop playback and switch to a new track:

;; Given a pstream bgmusic and rsounds bgsong1, bgsong2, the following 
Interactions
;; occur:
> (pstream-play bgmusic bgsong1)
"sound is queued"
> (stop)
> (pstream-play bgmusic bgsong2)
"sound is queued";; BUT NO SOUND PLAYS.
> 
;; - end interactions -

So, am I correct in thinking that pstreams aren’t sufficient for what I’m 
attempting? Is there another way to stop playback, other than calling (stop)?

Thanks,
Jordan

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] Playing sound

[Jordan Johnson]

Hi John,

Thanks for the reply.

> If you’re concerned about test failures in pkg.racket-lang.org… well, my test 
> cases try to load the shared libraries and play sound, so that’s not possible 
> in pkg.racket-lang.org’s environment.

Oh, good. That's less of a worry then.

Will the separate shared libraries likely cause any extra difficulties when he 
bundles his code up into Win/Mac executables?

> I’m guessing you’re looking for an imperative API that looks something like 
> this:
> 
> (start-playing! stream sound at-time #:fade-in time-duration) => token
> (stop-playing! stream token at-time #:fade-out time-duration)

Yep, that should do what he wants.


Best,
Jordan

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] Playing sound

['John Clements' via Racket Users]

> On Jan 27, 2016, at 6:02 PM, Lehi Toskin  wrote:
> 
> On Wednesday, January 27, 2016 at 11:41:06 AM UTC-8, Jordan Johnson wrote:
>> Hi all,
>> 
>> 
>> I'm looking at audio again because I have a student doing a game project for 
>> which he wants background music to play — and I'm feeling some trepidation 
>> based on most of the audio-tagged packages on pkgs.racket-lang.org showing 
>> some test failures or other build problems.
>> 
>> 
>> His minimal goal is to loop a sound file from the start of play until the 
>> end, and have it work on Windows and Mac. We tried play-sound from 
>> racket/gui/base, and it almost fits: we just need to be able to kill the 
>> sound at the end, and it seems that killing the sound-playing thread is not 
>> enough. Is there a way to stop the audio?
>> 
>> 
>> His ideal goal is to be able to play different background music depending on 
>> the room the player is in. AFAICT this is beyond play-sound's ability, at 
>> least if he wants to do smooth transitions. Is there another library which 
>> would facilitate cross-fading from one audio file to another?
>> 
>> 
>> More generally, what are currently our best tools for incorporating music 
>> and sound effects into Racket games?
>> 
>> Best,
>> Jordan
> 
> This may be larger than the scope of the project, but there is an audio 
> library called RSound, which uses PortAudio (via a wrapper). It's more robust 
> and I'm certain there are things you could do with it that you couldn't with 
> play-sound.

Hi there! Author of rsound here.

If you’re concerned about test failures in pkg.racket-lang.org… well, my test 
cases try to load the shared libraries and play sound, so that’s not possible 
in pkg.racket-lang.org’s environment. I suppose I should refactor those tests 
so that only the ones that can run without portaudio get run.

RSound can certainly give you the ability to start, stop, and cross-fade sounds.

I’m guessing you’re looking for an imperative API that looks something like 
this:

(start-playing! stream sound at-time #:fade-in time-duration) => token
(stop-playing! stream token at-time #:fade-out time-duration)


… right?

John



-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.