Re: [linux-audio-dev] PTAF link and comments

New Message Reply About this list Date view Thread view Subject view Author view Other groups

Subject: Re: [linux-audio-dev] PTAF link and comments
From: Steve Harris (S.W.Harris@ecs.soton.ac.uk)
Date: Mon Feb 03 2003 - 14:32:46 EET


Commenting on your comment, I'm about to read the doc.

On Mon, Feb 03, 2003 at 12:18:10 +0100, David Olofson wrote:
> * Assuming that audio and sequencer stuff is in different
> threads, and even have plugins deal with the sync sounds
> like a *very* bad idea to me.

Yes, I very much doubt this would work.

> * GUI code in the same binaries is not even possible on
> some platforms. (At least not with standard toolkits.)

Agreed, this is not ideal, ohm force are a host based processing shop, so
haven't thought about these issues. The TDM people will obviously have
different ideas.

> * I think the VST style dispatcher idea is silly. A table
> of function pointers, with a bunch of reserved NULLs
> would be simpler, faster and just as extensible for all
> practical matters.

Ideally this would be abstracted in the API, ELF sytems wouldn't need this
anyway, its just because COFF is broken.

> * Is using exceptions internally in plugins safe and
> portable?

Exceptions are a C++ feature.

> * Hosts assume all plugins to be in-place broken. Why?

It doesn't really matter what the default is as long as you can override
it. That way is probably safer.

> * No mix output mode; only replace. More overhead...

But how many hosts use mix? As a plugin author, having to implement both
is a pain. I use code generation to make both functions.

> * Buffers 16 byte aligned. Sufficient?

This should just be a host guideline shouldn't it.

> * Events are used for making connections. (But there's
> also a function call for it...)

If the events are a signalling and the funciton calls make the actual
connections then this is convienient.

> * Note IDs are just "random" values chosen by the sender,
> which means synths must hash and/or search...

Yup, this is bad. VVIDs are not an obvious solution though.

> * Hz is not a good unit for pitch...
> * Why both pitch and transpose?

Presumably because they dont understand how to represent pitch.
Base frequnecy + octave cleans this up.

> * Why [0, 2] ranges for Velocity and Pressure?

1.0 for default/middle kinda makes sense, but I agree its wrong.

> * The "save state chunk" call seems cool, but what's
> the point, really?

This could be useful for state storing in hosts, but how many would use
it? I'd certainly be reluctant to implemnt it.

- Steve


New Message Reply About this list Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b28 : Mon Feb 03 2003 - 14:39:17 EET