Re: Sound support (2), using hardware directly

[Julius Schwartzenberg <julius@zgod.cjb.net>]

Stas Sergeev wrote:
> Softsynth involves timidity++,
> which requires ~P4 by default,
> but can be tuned even for 386
> I think.

I have ran it succesfully on my AMD Athlon @ 1GHz, and I know there are 
softsynths that need much less, although not for Linux. Hmm, this gives 
me an idea..., running a softsynth inside Dosemu... :) This probably 
won't work for games though.

>> Yes I noticed your pc-speaker driver in ALSA :)
> 
> Yes, but the kernel interface
> for it is still not in the kernel,
> thats the problem.
> 
>> Is the goal really to get full sound and music from Dosemu through a 
>> plain pc-speaker?
> 
> The goal is to make it compatible
> with the modern sound systems like
> alsa, and with the cheap soundcards
> without the HW mixing and midi. The
> pc-speaker, being the one of such a
> cards, will work just in case, but
> for me it is probably more important
> than any other cards out there:)

Hehe, it would be really interesting to see how well this will work. I 
haven't tried your driver yet (but I definetaly will), but I remember 
that the pc-speaker driver in BeOS was pretty good. Also sometimes the 
pc-speaker is somehow wired to the soundcard too, in those cases the 
quality isn't even too bad.

>> I understand the current OPL3 modules aren't suitable to be used in 
>> this case?
> 
> Of course.
> 
>> I personally indeed do not really care about needing root access, so a 
>> working solution using $_ports would work for me. Would you recommend 
>> sending an e-mail to the ALSA mailing list about this?
> 
> About that the $_ports doesn't work
> very well for the native OPL3? I
> don't think they'll help here.

Well, the problem seemed to be at the time that the (OPL3) driver from 
ALSA was conflicting with the OPL3 being accessed directly. I've now 
verified that this problem occurs on another system with different 
hardware too. It probably isn't a Dosemu problem either, so this points 
to the ALSA driver. I don't think it can be called a bug in the ALSA 
driver, but maybe it would be possible to support cases such as these in 
the driver somehow?

Thanks,
Julius