Four or so comments in-line.

Art Hunkins

----- Original Message -----
From: "Istvan Varga" 
To: 
Sent: Friday, March 11, 2005 7:03 AM
Subject: Re: [Cs-dev] New file release for Csound 5 and CsoundVST on Windows


> Art Hunkins wrote:
>
> > in CsoundAV (in contrast to all others) I *never* have to fool
>  > with buffer settings.
>
> CsoundAV does not, and will probably never run on anything other
> than Windows, making it much easier to find default buffer settings.

IMHO we *very much need workable default settings*. AV has them; CS5 must
achieve similar results (I know it will be more difficult as it's a
cross-platform setting).

> > In CS5 Windows, the buffer issue is quite unclear and changing
>
> More details on this ?

OK - CS5 docs currently state (needs revision) that both -b and -B are
required and must be set to ksmps. I *think* now (for Windows - not
apparently Linux?) -b is not required, but -B must still be set to ksmps.
I've bugged Michael endlessly to get rid of -B - since it must be ksmps
anyway. (Apparently that will happen at some time in the unspecified
future.) What's recommended on Windows as to the relationship between -b
and -B is a complete mystery to me. You talk a lot in Linux about powers of
two and the smaller the better latency. I've no idea whether that applies to
Windows or not. How's *that* for confusion? (And I just want good latency
and no zipper noise - plus defaults that work in 90% of real-time cases for
Windows.)

>  > plus seems to be different on different platforms.
>
> Well, but Csound5 at least actually runs on different platforms.
> And things like buffer settings are highly machine dependent.
>
> > Also, since only ASIO is implemented, regular soundcard users have to
>  > install and configure ASIO4ALL.
>
> PortAudio is not limited to ASIO, it also supports DirectSound and MME
> (as well as WDM but that is experimental). You may want to ask Michael
> Gogins to include a PortAudio module that is compiled to use DirectSound
> in the next Windows file release.

I've also bugged Michael endlessly for this. He already spent a great amount
of time and energy doing this; unfortunately his results gave quite poor MME
and DirectSound results, and he reverted to ASIO only. He's already promised
another try at it. Victor has also indicated he'll try at some point if no
one does it first. This is a *huge* issue for me; without its resolution,
CS5 will always be less than robust for my purposes.

> > I'll also add that Gab Maldonado has been active in maintaining and
> > debugging CsoundAV, and personally responsive to my particular needs
(i.e.,
> > realtime performance).
>
> Now if only all the efforts on maintaining the various platform specific
> Csound variants derived from old versions of Csound4 would be spent on
> Csound5, we would eventually get one good version instead of several
> broken ones, all with their own set of bugs that have been fixed in
others.
> And the fact is that in recent years CsoundAV was only actively developed
> in the areas of real time audio, OpenGL, and FLTK, while core language
> support is still only 4.19.

FWIW - Gab has just introduced 0.0432 which updates the roster of canonical
opcodes and moves to the latest Portaudio. As CS5 comes out of beta, he
plans to update further.
>
>
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Art Hunkins wrote:

> OK - CS5 docs currently state (needs revision) that both -b and -B are
> required and must be set to ksmps. I *think* now (for Windows - not
> apparently Linux?) -b is not required, but -B must still be set to ksmps.
> I've bugged Michael endlessly to get rid of -B - since it must be ksmps
> anyway. (Apparently that will happen at some time in the unspecified
> future.) What's recommended on Windows as to the relationship between -b
> and -B is a complete mystery to me. You talk a lot in Linux about powers of
> two and the smaller the better latency. I've no idea whether that applies to
> Windows or not. How's *that* for confusion? (And I just want good latency
> and no zipper noise - plus defaults that work in 90% of real-time cases for
> Windows.)

You'd better ask Michael Gogins about buffer sizes on Windows as I do
not use Csound5 (or any other version) on that platform. However, the
fact that buffer sizes may need to be set differently on various platforms
is not a good argument for saying that CsoundAV is "better": after all,
it is not likely to ever support more than one operating system, and you
only need to care about settings on your platform and not others that
you do not use.

> I've also bugged Michael endlessly for this. He already spent a great amount
> of time and energy doing this; unfortunately his results gave quite poor MME
> and DirectSound results, and he reverted to ASIO only.

Csound5 implements real time audio using DLLs that are loaded at run-time.
There is no technical reason why multiple DLL files for different interfaces
(ASIO, DirectSound, etc.) could not be included with the package.
This also includes the possibility of having plugins for both PortAudio
v18 and v19.

> FWIW - Gab has just introduced 0.0432 which updates the roster of canonical
> opcodes

I was not talking about opcodes, but rather than that, important but
often overlooked parts of Csound such as the API, parser, event processing
loop, and the like. This of course implies that opcodes that depend
on features of newer engines are not present even in 0.0432.

> and moves to the latest Portaudio.

Which, I assume, will have many of the same problems as in Csound5...


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net