I'd be interested to hear more about the PD/Csound shootout.

Regards,
Mike

-----Original Message-----
>From: Iain Duncan 
>Sent: Mar 24, 2006 11:11 PM
>To: csound@lists.bath.ac.uk
>Subject: Re: [Csnd] ChucK?? (was Re: [Csnd] The future of computer music? (was Re: [Csnd] The future of winsound))
>
>
>> In the battle of so-called 'elegance' and sound-production, I will always
>> look for the latter. But then again that's the way I was schooled and
>> don't expect anyone to feel that way (and that's why I use csound...)
>
>And there is also efficiency. Csounds inflexibility is part and parcel
>of it's close to the machine architecture, which means that if you don't
>mind coding all your options ahead of performance ( as one would
>normally do building a synth in hardware for example ) the performance
>increase is significant. Glyn and I did a PD/csound shoot out a couple
>of years ago for a few tasks and it wasn't even close.
>
>Iain
>-- 
>Send bugs reports to this list.
>To unsubscribe, send email to csound-unsubscribe@lists.bath.ac.uk

Iain Duncan wrote:

> Crap, it was quite a while a go, so I can't remember the details. I
> think we were checking out how many oscillators we could do in real time
> under various set ups for live use.

There was a thread in the chuck list a while back. Here are that highlights:

Graham Percival posted a specefic piece of code, that would play 
smoothly with 47 sines (1 ghz G4 powerbook, 512 megs ram, OSX 10.3.9).

Ge Wang got 62 sine on an equal system, and noted something about 
setting "processor performance" to "highest".

With changed settings Graham could also do 62 sines.

I (Atte Jensen) could get 136 sines (2.4Ghz PIV laptop, 512MB ram that 
runs debian/unstable). I did a quick test in csound and could get 660 sines.

Axel Balley got 430 sines on SuperCollider (Powerbook G4 800/1GB).

Michal Sata got 100 chuck-sines (AMD AthlonXP 1.2G) and 505 sines in pd 
(same system).

-- 
peace, love & harmony
Atte

http://www.atte.dk

Istvan Varga wrote:

> How did you do this test ?

Can't find the source for that, but it was done very, very fast. I think 
  I just made an instrument playing a sine using oscili and a score that 
played that instrument. For more instances of the sine I used copy/paste 
in the score...

-- 
peace, love & harmony
Atte

http://www.atte.dk

PD vs. Csound? isn't it kinda weird since Csound is a synth and PD a 
control environnement? on what basis can they be compared? i actually 
use both, but for very different things. i mean sure, it's gonna be 
harder to get good synthesis out of PD, but it wasn't designed with that 
in mind so isn't it kinda useless to compare it to Csound?

i say use the tools that suit your work and let others do the same, 
screw the app war.


just a thought ;)

jm


Michael Gogins wrote:

>I'd be interested to hear more about the PD/Csound shootout.
>
>Regards,
>Mike
>
>-----Original Message-----
>  
>
>>From: Iain Duncan 
>>Sent: Mar 24, 2006 11:11 PM
>>To: csound@lists.bath.ac.uk
>>Subject: Re: [Csnd] ChucK?? (was Re: [Csnd] The future of computer music? (was Re: [Csnd] The future of winsound))
>>
>>
>>    
>>
>>>In the battle of so-called 'elegance' and sound-production, I will always
>>>look for the latter. But then again that's the way I was schooled and
>>>don't expect anyone to feel that way (and that's why I use csound...)
>>>      
>>>
>>And there is also efficiency. Csounds inflexibility is part and parcel
>>of it's close to the machine architecture, which means that if you don't
>>mind coding all your options ahead of performance ( as one would
>>normally do building a synth in hardware for example ) the performance
>>increase is significant. Glyn and I did a PD/csound shoot out a couple
>>of years ago for a few tasks and it wasn't even close.
>>
>>Iain
>>-- 
>>Send bugs reports to this list.
>>To unsubscribe, send email to csound-unsubscribe@lists.bath.ac.uk
>>    
>>
>
>
>
>  
>

> I'd be interested to hear more about the PD/Csound shootout.

Crap, it was quite a while a go, so I can't remember the details. I
think we were checking out how many oscillators we could do in real time
under various set ups for live use. I do vividly remember thought that
we were surprised at how much faster csound was for our test, and ( at
the time ) how much faster Istvan's 4.23 ran on linux than on windows.

It would be interesting to compare again with csound5 compiled and
optimized for this chip ( AMD64 ) but unfortunately free time is the one
thing I don't have right now and what little I have I'd like to use for
the api examples and modules.

Iain

> PD vs. Csound? isn't it kinda weird since Csound is a synth and PD a
> control environnement? on what basis can they be compared? i actually
> use both, but for very different things. i mean sure, it's gonna be
> harder to get good synthesis out of PD, but it wasn't designed with that
> in mind so isn't it kinda useless to compare it to Csound?
> 
> i say use the tools that suit your work and let others do the same,
> screw the app war.

That was the point, to find out for us which tool fit the job at hand,
the job having various features that favoured one or the other. It was
done for our own purposes, not to tell the world about csound vs PD.

Iain

Atte André Jensen  writes:

> Michal Sata got 100 chuck-sines (AMD AthlonXP 1.2G) and 505 sines in
> pd (same system).

I guess I could try csound, too, I've installed it again a few weeks
ago (after a nearly 4 year break).

Do you have your orc/sco handy for the test?

./MiS

Jean-Michel Dumas wrote:

> PD vs. Csound? isn't it kinda weird since Csound is a synth and PD a 
> control environnement? on what basis can they be compared? i actually 
> use both, but for very different things. i mean sure, it's gonna be 
> harder to get good synthesis out of PD, but it wasn't designed with 
> that in mind so isn't it kinda useless to compare it to Csound?
>
PD has all kinds of objects to make synthesizers with, plus a graphical 
interface.
Miller Puckette, the originator of pd, has an online book that 
demonstrates the use of pd in computer music generation: 
http://crca.ucsd.edu/~msp/techniques.htm
By using csoundapi~ in pd you get even more audio generation 
possibilities. I think pd _was_ designed as a synthesis language, mainly 
to be used with MIDI. The main difference is that pd is programmed in 
the gui by connecting boxes while csound is programmed in a text editor. 
But you can implement a gui in csound or run pd with no gui. On a single 
machine the gui is in competition with the sound engine for processor 
cycles, so usually something that runs with a command-line interface can 
be expected to run faster and with fewer glitches.

Martin