Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

kX Audio Driver (Sound Blaster Live! and Audigy series driver) OS X PPC porting project

Amethyst1 said:
Would a Thunderbolt PCIe box not be enough?
Click to expand...

Unfortunately no, because as far as i am aware pcie devices connected using thunderbolt adapters are still recognized as removable devices and so the driver must have special support for this configuration or the system wouldn't even let the driver be used.
 
Tratkazir_the_1st said:
@ITzTravelInTime

I don't remember SB cards (SBLive & Audigy series) on PCI-Ex. Your driver supports more recent devices?
Click to expand...

There are at least 3 cards that i know of that have chipsets from the audigy/live cards and are on pcie:
1) the Audigy RX which is basically an Audigy 4 non-pro with a pcie bridge chip on the same board and it's still produced and sold by creative.
2) The emu E-0404 pcie (basically a standard EMU E-0404 with a pcie bridge on the same board as far as i am aware)
3) Those Chinese sound cards using salvaged EMU 10k1 chips from sound blaster live cards again using a built-in pcie bridge.

Using a PCIe to PCI adapter to use an older pci card should also work, but it's an unreliable and junky setup to be honest.

Also i have been using the Audigy RX with success on the pcie g5 and the various intel mac pros with pcie slots, if you want something to use those cards with, i am also maintaining intel support in my code.
 
  • Like
Reactions: reukiodo
Amethyst1 said:
Does this boil down to adding the IOPCITunnelCompatible flag or is it more involved?
Click to expand...

As far as i am aware there needs to be done changes regarding how the driver talks to the card, because those sound cards just uses dma, hardware interrupts, the pci config space and old style io ports to communicate with the system, while modern pci devices usually ditches the io port in favour of memory mapped buffers. Also the kx audio driver just directly uses the io ports address space (or the equivalent memory mapped io ports on powerpc) to send the io messages to the card, which i don't know it's possible to do via thunderbolt.

Another thing to mention is that dma transfers are not done trough buffer pointers allocated directly, but you have to create various IODMACommand objects for each dma buffer.

Also changes are needed to address the extra latency and the plug and play nature of the thunderbolt connectivity.

Because of those reasons i think this would require a lot of effort just to pull off, i am not saying it isn't possible, but it would require a lot of work and time, which i am not sure i want to put for such a small use case.
 
  • Like
Reactions: reukiodo and Amethyst1
eastone said:
It works pretty nice (CT4830) I finally have one card that works fine with my triple boot G4 (Debian, MorphOS, Leopard). Thank you for your hard work @ITzTravelInTime
Click to expand...

It should also work on OS9 too! check this out: https://archive.org/details/sblive-mac-patch

Those sound blasters are very versatile, one of the reasons i like them (even if they are a real pain to make drivers for, tbh).

BTW i have a new driver version in development with several improvements, the major ones are:
- Support for multiple sound cards at the same time.
- Now using system api to comunicate with the sound card (unlike the current release wich just makes memory operations to the card, bypassing any system routines)
- New audio conversion and clipping code for the DMA buffers (this means that the driver is less cpu intensive while using the sound card for playback and the sound quality should be a bit better too with the new code).
 
  • Like
Reactions: reukiodo and eastone
Tratkazir_the_1st said:
What is it's model number?
Click to expand...

It's the sb1550 and it looks like this (see attached picture).

It's however untested by me on powerpc pc because i don't have the pcie g5, you might hear a hissing noise while connecting headphones or speakers to the card once the driver is loaded and for that i use a control app on intel macs that i haven't figure out yet how to port it to powerpc (i am not good in obj-c and it need to port it to obj-c 1).
 

Attachments

  • A19BD69A-A702-4AFB-B924-251C4FF5BD2D.png
    A19BD69A-A702-4AFB-B924-251C4FF5BD2D.png
    323.2 KB · Views: 127
  • Like
Reactions: reukiodo
ITzTravelInTime said:
It's the sb1550 and it looks like this (see attached picture).

It's however untested by me on powerpc pc because i don't have the pcie g5, you might hear a hissing noise while connecting headphones or speakers to the card once the driver is loaded and for that i use a control app on intel macs that i haven't figure out yet how to port it to powerpc (i am not good in obj-c and it need to port it to obj-c 1).
Click to expand...
You need a tester? I can go ahead and get one.
 
ifrit05 said:
You need a tester? I can go ahead and get one.
Click to expand...

honestly idk, but it's an expensive card and it's, in many ways, an upgrade over the on-board audio on the g5, so it comes to you to justify this purchase, it's also a good card for windows 10 and Linux, if you want to repurpose it, i have even a version of the kx audio driver for macOS ventura.

Consider this: on OS X and on PowerPC many features of the driver are broken, for instance, the recording has never been tested on powerpc due to the lack of the above mentioned setup application that i can't backport to Obj-c 1.x

BTW i have afriend with an audigy rx and a pcie g5, and i can just go to his house if am desperate to test it out.
 
So guys i have some news:

- I have implemented a proper method to have multiple sound cards working (since the old one i had initially tested was broken in macOS ventura on the intel test rig, so i have done some research and actually found the intended way of doing it).

- I have been working on input support for the e-mu professional sound cards (still looking for people to help backporting the kx audio driver's ac97 setup program to obj-c 1 and os x tiger, so we can test inputs on consumer cards too).

- Fixing bugs and driver crashes that happened to me while doing some testing.

- Implementing the fixes on the other sound card drivers i have forked and ported to powerpc.

I have also got more hardware, mainly the e-mu 1212m, an e-mu 1616m (of witch i can only use the external dock with a supported pci/pcie sound card) and a dock matching my e-mu 1010m pci i have in my g5 right now.

I have no eta on when i will release a new driver build, i'd like to make somenthing more stable and that works better than my previous builds.
 
  • Like
Reactions: reukiodo, r34per, dextructor and 4 others
First of all happy new year guys!

Second: i can say that with my current source code, the recording is working on par with intel on the PowerPC side, me and a friend just tested it using some gear and my E-MU 1010m with the 1820 dock.

Third: i am looking into putting some functionality from the config utilities (that haven't been ported to powerpc yet) into the driveritself as default audio controls.
 
  • Like
Reactions: reukiodo, dextructor, r34per and 2 others
I just discovered this driver and was thinking to try it on my Hackintosh Snow Leopard machine.
I have a Creative SB Audigy 2 ZS platiunum pro PCI, I believe a very early version.

From my investigations into the code (have built it) and IORegistryExplorer, it appears that only the
11020004 device is supported, i.e. matches what is currently in the Info.plist for the kXAudioDriver.kext.

My question is, what about the firewire port (11024001) and the game port (11027003).

Can I assume that there is no support for these devices ?

I have not yet loaded the driver, only inserted the card and explored the device tree with
IORegistryExplorer.

Thanks

-rick
 
eratta said:
My question is, what about the firewire port (11024001) and the game port (11027003).

Can I assume that there is no support for these devices ?
Click to expand...
Basically you have system drivers for the firewire part and it's picked up automatically by those and works out of the box, for the gameport part there is basically no support, and i don't want to write a driver from scratch for it, this project is mostly to get the audio part working decently.

That said, feel free to work it and pubblish all your changes, this project is under GNU GPL so it's better to keep your changes public anyway, those can also help me a lot too.

I am also working on some more audio drivers, like the CMI8738 one (i have added some basic CMI8338 support to it and i'd like to get my hands on other similar chips to get them supported too) and the Envy24HT one, sadly i don't have lots of time to do all the work i'd like to make on them and i would also love to get more cards to test and add support for.

If you have any more questions feels free to post them or to DM me.

Have a nice day.

- Pietro.

PS:
For snow leopard there should be a target for it in the kx audio driver xcode project (open the one located into the macosx folder), 64 bit targets needs kernel version 9 or 10 library dependancies or later to work also with modern macOS, so that's why it's separated from the 32bit ones (but also because the 64 bit one can build just from snow leopard or later), you might have to change your sdk, compiler and built target settings for the floating point conversion 64 bit library and the kext 64 bit target itself, if you want i can also make a build for you when i have some time.
 
Tratkazir_the_1st said:
I thought it works OOB. What is shown in System Profiler in FireWire section?
Click to expand...
I hadn't done any testing.
But you are correct.
I connected a firewire device to the firewire port, and indeed OS X attached a driver.

I was expecting a "special" driver, but I guess it depends on what is connected.

Thanks.

-rick
 
  • Like
Reactions: ITzTravelInTime
kindly help Emu1212m mac osx mojave no ADAt input only ADAT out put🙃and thanks for you👌r work
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back