Re: SoundRecorder

> I downloaded the 3.9 image you pointed to, ran it on my Mac, and  
> discovered:
>
> (a)  The sound-in-event-recorder bug *does* occur.  But it's  
> slightly different:  although the speed of playback is nearly  
> doubled, as in OLPC, the *pitch* does not change.  So it is not a  
> Donald Duck sound.
>
> (b)  Also, the same bug now (i.e. in 3.9) also occurs in the  
> vanilla SoundRecorderMorph!  (In the OLPC/Mac image, the vanilla  
> SoundRecorderMorph produces good playback on the Mac -- it's only  
> the use of sound within an EventRecorderMorph that's  
> problematical.  In the 3.9/Mac image *every* SoundRecorderMorph use  
> has this problem.  But again, the overly-fast playback occurs  
> *without* a pitch rise.)
>
> So evidently 3.9 went "final" with the SoundRecorder completely  
> broken on the Mac.
>
>
> Hopefully the above observations will provide some useful clues.  
> After all, we now have examples of images in which:
>
> [a]  Plain SoundRecorder and EventRecorder-Sound both work fine  
> [Squeakland3.8]
>
> [b]  Plain SoundRecorder works fine but EventRecorder-Sound is  
> broken, donald-duck style [OLPC]
>
> [c]  Plain SoundRecorder and EventRecorder-Sound are *both*  
> broken;  playback too fast but no donald-duck-style pitch [Squeak 3.9]
>
>
> BTW:  all my tests were run with the same Mac VM, namely Squeak  
> 3.8.12beta4U.app.
>
>
> Cheers,
>
>  -- Scott
>
>
> On Nov 8, 2006, at 10:31 AM, Yoshiki Ohshima wrote:
>
>>   ... It looks like the code we use is pretty much similar in the one
>> Squeak 3.9.  Can you check if you can reproduce it in an image like:
>>
>> http://ftp.squeak.org/3.9/Squeak3.9-final-7067.zip
>>
>> -- Yoshiki
>>
>> On Oct 31, 2006 at 2:27 PM, Scott Wallace wrote:
>>
>>> The sound-recording option of the EventRecorder in OLPC1.0 has  
>>> somehow gotten broken, in that when you play back an event-
>>> recording that was recorded with sound, the sound comes out  
>>> Donald-Duck style, almost twice as fast as real time.

> Hi, John, and all,
>
> It perhaps bears mentioning that my tests were all run on an Intel  
> Mac.
>
> -- S
>
> On Nov 8, 2006, at 5:32 PM, Scott Wallace wrote:
>
>> I downloaded the 3.9 image you pointed to, ran it on my Mac, and  
>> discovered:
>>
>> (a)  The sound-in-event-recorder bug *does* occur.  But it's  
>> slightly different:  although the speed of playback is nearly  
>> doubled, as in OLPC, the *pitch* does not change.  So it is not a  
>> Donald Duck sound.

> My notes indicate that at 3.8.6b5 is where I switch to the unix/
> quartz unix sound code that uses core audio.
>
> You should try with a 3.8.6b4 or earlier VM and see if you have the  
> same problems.
>
>
> On 8-Nov-06, at 7:22 PM, Scott Wallace wrote:
>
>> Hi, John, and all,
>>
>> It perhaps bears mentioning that my tests were all run on an Intel  
>> Mac.
>>
>> -- S
>>
>> On Nov 8, 2006, at 5:32 PM, Scott Wallace wrote:
>>
>>> I downloaded the 3.9 image you pointed to, ran it on my Mac, and  
>>> discovered:
>>>
>>> (a)  The sound-in-event-recorder bug *does* occur.  But it's  
>>> slightly different:  although the speed of playback is nearly  
>>> doubled, as in OLPC, the *pitch* does not change.  So it is not a  
>>> Donald Duck sound.
> --
> ======================================================================
> =====
> John M. McIntosh <[hidden email]>
> Corporate Smalltalk Consulting Ltd.  http://
> www.smalltalkconsulting.com
> ======================================================================
> =====
>
>

> Hi, John,
>
> Okay -- as you suspected, I can confirm that the bug does *not*  
> occur when I switch to a sufficiently old VM.
>
> Running an old 3.8.1Beta2 VM, I tested both the SoundRecorder and  
> sound use in the EventRecorder, in both Squeak3.9 and OLPC1.0, and  
> never encountered any problems.
>
> So...?
>
> Cheers,
>
>   -- Scott
>
>

> Well it means I get to read some C code and wander thru apple  
> developer pages, and press buttons in XCode for the weekend.
>
> On 8-Nov-06, at 10:49 PM, Scott Wallace wrote:
>
>> Hi, John,
>>
>> Okay -- as you suspected, I can confirm that the bug does *not*  
>> occur when I switch to a sufficiently old VM.
>>
>> Running an old 3.8.1Beta2 VM, I tested both the SoundRecorder and  
>> sound use in the EventRecorder, in both Squeak3.9 and OLPC1.0, and  
>> never encountered any problems.
>>
>> So...?
>>
>> Cheers,
>>
>>   -- Scott
>>
>>
>
> --
> ======================================================================
> =====
> John M. McIntosh <[hidden email]>
> Corporate Smalltalk Consulting Ltd.  http://
> www.smalltalkconsulting.com
> ======================================================================
> =====
>
>

> Hi, John, and all,
>
> It perhaps bears mentioning that my tests were all run on an Intel  
> Mac.
>
> -- S
>
> On Nov 8, 2006, at 5:32 PM, Scott Wallace wrote:
>
>> I downloaded the 3.9 image you pointed to, ran it on my Mac, and  
>> discovered:
>>
>> (a)  The sound-in-event-recorder bug *does* occur.  But it's  
>> slightly different:  although the speed of playback is nearly  
>> doubled, as in OLPC, the *pitch* does not change.  So it is not a  
>> Donald Duck sound.

> I think I have found the culprit.
>
> If you listen carefully to the recorded sound, you will notice that  
> frames are dropped at regular intervals. Since this pattern is  
> quite obvious and the VM doesn't seem to screw the sound data, it  
> should be something happening at the image level.
>
> Looking for a clue in a 3.9 alpha version, I found a change made by  
> "dgd" on April 4, 2006:
>
> "workaround for OSS emulation on top
> on ALSA (on Linux environments)"
> (Delay forMilliseconds: 20) wait.
>
> If the last line is commented out, the sound recorder just works as  
> expected. The same code is present in the OLPC and Squeakland  
> images and both work pretty well without this delay.
>
> CoreAudio records the audio input in real time but SoundRecorder  
> picks the data once in a while. The end result is that the sound  
> data is stored in a different time frame that the sampling rate and  
> the sound plays faster.
>
> Now I am not sure what can be done to have a cross-platform  
> solution, but as far I have seen the problem is not on the Mac VM  
> side.

> I think I have found the culprit.
>
> If you listen carefully to the recorded sound, you will notice that  
> frames are dropped at regular intervals. Since this pattern is  
> quite obvious and the VM doesn't seem to screw the sound data, it  
> should be something happening at the image level.
>
> Looking for a clue in a 3.9 alpha version, I found a change made by  
> "dgd" on April 4, 2006:
>
> "workaround for OSS emulation on top
> on ALSA (on Linux environments)"
> (Delay forMilliseconds: 20) wait.
>
> If the last line is commented out, the sound recorder just works as  
> expected. The same code is present in the OLPC and Squeakland  
> images and both work pretty well without this delay.
>
> CoreAudio records the audio input in real time but SoundRecorder  
> picks the data once in a while. The end result is that the sound  
> data is stored in a different time frame that the sampling rate and  
> the sound plays faster.
>
> Now I am not sure what can be done to have a cross-platform  
> solution, but as far I have seen the problem is not on the Mac VM  
> side.
>