Mixer support on the iBook2

Mixer support on the iBook2

[Ethan Blanton]

[-- Attachment #1: Type: text/plain, Size: 763 bytes --]

Hi all,
I have gotten the sound on my iBook2 working (via the files on
http:/www.hadess.net/files/ppc/ibook2/), and I have some questions...

I know there has been some discussion on this list that the OSS mixer
semantics are insufficient for the PPC sound architecture in modern
Macs.  I don't know if this is related or not...

In any mixer program I can find, the volume steps on the iBook2 are
simply too large.  The sound card goes from muted to too loud to use
through earphones in one step of the slider.  Is this a limitation of
the hardware, the driver architecture, or what?  Or is it too early to
be asking?  ;-)
Ethan

-- 
If I've told you once, I've told you once
And once is all that you needed.
		-- The Refreshments, "Carefree"

[-- Attachment #2: Type: application/pgp-signature, Size: 0 bytes --]

Re: Mixer support on the iBook2

[Bastien Nocera]

Ethan Blanton wrote:
> Hi all,
> I have gotten the sound on my iBook2 working (via the files on
> http:/www.hadess.net/files/ppc/ibook2/), and I have some questions...

Please use the dmasound.tar.gz , it's a better driver (anton2 on
#mklinux fixed that 8 days ago), and I'm supposed to clean it up for
integration in BenH's tree (if he and Iain think that it's clean
enough). Or maybe Olaf will just be sick of waiting and do it himself ;)

Given that my iBook is going to the repair shop (the cable to the screen
got hurt and my display goes purple), one might want to take this task.

the dmasound.tar.gz contains a replacement for drivers/sound/dmasound
you still need the i2c-keywest driver from pfh's driver, but not the
dmasound_core hacks anymore.
All the i2c calls should be #ifdef'ed for machines without i2c buses to
still work. And a nice configuration option is needed as well.

> I know there has been some discussion on this list that the OSS mixer
> semantics are insufficient for the PPC sound architecture in modern
> Macs.  I don't know if this is related or not...
>
> In any mixer program I can find, the volume steps on the iBook2 are
> simply too large.  The sound card goes from muted to too loud to use
> through earphones in one step of the slider.  Is this a limitation of
> the hardware, the driver architecture, or what?  Or is it too early to
> be asking?  ;-)

The driver...

Cheers


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Mixer support on the iBook2

[Ethan Blanton]

[-- Attachment #1: Type: text/plain, Size: 1351 bytes --]

Bastien Nocera spake unto us the following wisdom:
> Ethan Blanton wrote:
> > I have gotten the sound on my iBook2 working (via the files on
> > http:/www.hadess.net/files/ppc/ibook2/), and I have some questions...
> 
> Please use the dmasound.tar.gz , it's a better driver (anton2 on 
> #mklinux fixed that 8 days ago), and I'm supposed to clean it up for 
> integration in BenH's tree (if he and Iain think that it's clean 
> enough). Or maybe Olaf will just be sick of waiting and do it himself ;)

That's what I'm using.  It works beautifully (thanks!), but the volume
control is too coarse-grained to be useful for much...

> > In any mixer program I can find, the volume steps on the iBook2 are
> > simply too large.  The sound card goes from muted to too loud to use
> > through earphones in one step of the slider.  Is this a limitation of
> > the hardware, the driver architecture, or what?  Or is it too early to
> > be asking?  ;-)
> 
> The driver...

To clear up some confusion I have here ...  Are you saying it's a
limitation of the driver (in which case I'm more than happy to wait
for it to be fixed as I know *nothing* about DSP or sound hardware) or
that I need to snag the driver?
Ethan

-- 
If I've told you once, I've told you once
And once is all that you needed.
		-- The Refreshments, "Carefree"

[-- Attachment #2: Type: application/pgp-signature, Size: 0 bytes --]

Re: Mixer support on the iBook2

[Olaf Hering]

On Tue, Aug 21, Ethan Blanton wrote:

> Bastien Nocera spake unto us the following wisdom:
> > Ethan Blanton wrote:
> > > I have gotten the sound on my iBook2 working (via the files on
> > > http:/www.hadess.net/files/ppc/ibook2/), and I have some questions...
> >
> > Please use the dmasound.tar.gz , it's a better driver (anton2 on
> > #mklinux fixed that 8 days ago), and I'm supposed to clean it up for
> > integration in BenH's tree (if he and Iain think that it's clean
> > enough). Or maybe Olaf will just be sick of waiting and do it himself ;)
>
> That's what I'm using.  It works beautifully (thanks!), but the volume
> control is too coarse-grained to be useful for much...

I have placed the ibook2 patches here:
http://www.penguinppc.org/~olaf/ibook2/patches/

70__00_linux-2.4.8--benh.diff.gz
70__2.4.8-ibook2_dmasound.diff.gz
70__2.4.8-ibook2_zzz_dmasound-requestmodule.diff.gz

Everything is against 2.4.8, maybe it fits on the 2.4.9 tree.


What I have found so far is that you have sound working after modprobe
dmasound_pmac. when you launch kmix the volume slider is at 100%. but as
soon as you move the volume control a bit the sound gets really loud. So
I guess the driver reports the wrong volume level to the userland.



Gruss Olaf

--
 $ man clone

BUGS
       Main feature not yet implemented...

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Mixer support on the iBook2

[Iain Sandoe]

> What I have found so far is that you have sound working after modprobe
> dmasound_pmac. when you launch kmix the volume slider is at 100%. but as
> soon as you move the volume control a bit the sound gets really loud. So
> I guess the driver reports the wrong volume level to the userland.

Missed most of this thread ... I've been overseas on business...

----

kmix (in common  with all the mixers I've tried so far) does not obey the
OSS rules - which say that the mixer *should not* rely on the return value
from the driver.

a lot of chips need the driver to modify the User's parameter before
supplying it to the chip... so you might, for example, lose precision when
it is stored in the chip.

For AWACS/Screamer I have fixed this problem in the driver by maintaining a
cached version of the value the User has set and just echoing this back.

This is, perhaps, not correct behaviour for the driver really, one day I
should ping all the mixer maintainers that it doesn't work for and ask them
to fix it there.   However, it's not an extravagant or particularly bloated
work-around for the short-term... it's also justifiable for situations where
you have write-only regs, or need to restore from sleep.

---

Looking at the TAS3001C data sheet reminds me that the value range is
greater than 0..100 so there must be some scaling needed.

Maybe that is related to the problem?

OSS does *not* (AFAIR) specify whether the steps are supposed to be linear
or logarithmic - just that they represent 0 .. 100%.  From a perceptual POV
logarithmic makes most sense - but you might want to expand the 0..100 into
the full range available from TAS3001C.

It is TAS3001C we're considering here (yes?) ... same comment apply also to
iBook1 with DACA but in a less severe way because the DACA range is
smaller...

----

Has anyone made the keywest-i2c driver interrupt-driven yet?
(I assume this is what's being used to control the mixer).

----

There are quite a lot of things to sort out to integrate the new chips into
dmasound_pmac tidily... I didn't do any re-coding yet - still working out
the layout changes - and I've also got to do the work-to-eat-stuff...  ;-)

ciao,
Iain.

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/