No HDMI sound from iGPU

[deeveedee]

@LockDown please take this case.  I need to sign off and I feel you and agastrato are very close to a working solution.

[agstrato]

1 minute ago, deeveedee said:

 

If it's any consolation, I'm currently patching the frame buffer for an HP Elitebook 850 G7 laptop.  There are numerous EFIs shared for this laptop.  I haven't found one of them without mistakes and not a single one with the correct, fully patched frame buffer.  This may be a shock, but not everything you find on the internet is true.

Nothing shocks me anymore, i had a lot of EFI tests as well but this was the only one that actually gave me hardware acceleration so after 2 days of battling with it was like yeah ok this might be good. Tbh i don't really need this hdmi sound to work right now because i am not moving this PC to the living room yet and it will take some time for me to build a new one but i just wanted to have it done. I am gonna test more tomorrow. Too bad there is no tool to pretty much automate this procedure of checking buffers. Even if i would have to reboot again and again, as long as the tool goes through all of them until the good one is found would definitely be amazing.

[Oyecomova]

10 hours ago, agstrato said:

Hola a todos, parece que ahora puedo publicar en cualquier sitio, así que aquí va mi hilo. Tengo una placa base Asus Maximus XI Extreme Z390 con un procesador 9900KS y, tras parchear los framebuffers, todo funciona bien. Uso la iGPU conectada a otro monitor para mi hackintosh, y el problema es que el control deslizante de volumen del monitor está desactivado. macOS detecta dos monitores, lo cual es incorrecto, y no, no los estoy monitorizando ni nada parecido. No sé por qué los detecta. Adjunto mi archivo EFI y una captura de pantalla; tened en cuenta que he ido añadiendo y modificando cosas, solo quiero solucionar el problema del sonido HDMI. Creo que el problema reside en la detección incorrecta de los monitores. ¿Alguna sugerencia?

Z390-Asus-M11Extreme.zip 4,68 MB · 1 descarga

Your config.plist file has many validation errors.
How does it work like that?

[deeveedee]

5 minutes ago, Oyecomova said:

Your config.plist file has many validation errors.
How does it work like that?

 

Maybe he's not running Open Core 1.0.5?  That is a good catch.  I completely forgot to check the config.plist with ocvalidate - rookie mistake   -> MY rookie mistake.

Edited November 8, 2025 by deeveedee

[deeveedee]

43 minutes ago, agstrato said:

Well damn, what do you suggest, first i try the framebuffer patch for each connector and then try all alcids or ?

Sometimes a web search works...

https://github.com/acidanthera/AppleALC/wiki/Supported-codecs

[agstrato]

7 hours ago, deeveedee said:

Sometimes a web search works...

https://github.com/acidanthera/AppleALC/wiki/Supported-codecs

Oh believe me i found this long time ago when i did my first hack, i just happened to try out layout 28 after 1,3,5,17 and it sounded pretty good to me. The whatevergreen article also been found first day, i also seen the emptyness when you click spoiler for my particular processor and honestly the information there is very confusing for ex: "The bytes in Properties must be put in reversed order. For example: framebuffer 0x0166000B would be put in as 0B006601, DevID 0x1E3A would be put in as 3A1E0000". Don't get me wrong with this i'm just saying, i know how to google very well it's why i found those framebuffers because if you look around any opencore efi or config listed anywhere for my mobo is very old.

8 hours ago, Oyecomova said:

Your config.plist file has many validation errors.
How does it work like that?

This is happening because i opened and edited my config with a bunch of software and also texteditors but it is made from scratch using OpenCore 1.0.5. Once i get this hdmi sound working i`ll build a fresh one with 1.0.6 using only ProperTree. The config does work though it's made using the dortania guide the only section that changed a lot was the deviceproperties one.

[agstrato]

Ok so it's definitely con1, i just tested the con0 and con2, the monitor goes black right after opencore loads. This happened to me a lot before finding the framebuffers.

[deeveedee]

@agstrato You're off to a good start.  Looking forward to seeing how conclude this hacking experience.  Your next steps are

• Fix the open core config.plist errors to eliminate them as possible problems.  They're not affecting your frame buffer patching, but it's best to minimize the possible problems before you proceed.
• Be appreciative and say thank you to the people helping you every once in a while

Good luck.

[agstrato]

38 minutes ago, deeveedee said:

@agstrato You're off to a good start.  Looking forward to seeing how conclude this hacking experience.  Your next steps are

• Fix the open core config.plist errors to eliminate them as possible problems.  They're not affecting your frame buffer patching, but it's best to minimize the possible problems before you proceed.
• Be appreciative and say thank you to the people helping you every once in a while

Good luck.

Ok then I'll see if I can make a config from scratch today. The plan was to fix this first and then make a clean one including a fresh install but I guess I can do it first no problem. I am appreciative I am sorry if anybody thinks I'm not, probably language barrier issue although my English is not bad at all or so I'm told. Anyway I've been through all alcids and it seems 25 is a good layout because it auto detects when I insert headphones and quality is good too so for the moment I'm gonna stick with it. HDMI sound still not working so probably I'm gonna make time to build the new OpenCore 1.0.6 and test. Besides the "issues" from my config.plist I'm pretty much out of ideas.

[agstrato]

This is a clean OpenCore 1.0.6 EFI. I have simplified the config.plist a lot.

EFI-1.0.6.zip

[feartech]

1 hour ago, agstrato said:

This is a clean OpenCore 1.0.6 EFI. I have simplified the config.plist a lot.

EFI-1.0.6.zip 4.21 MB · 1 download

you could try adding:

 

<key>hda-gfx</key>
<string>onboard-1</string>

in the PciRoot(0x0)/Pci(0x2,0x0) section

 

 

 

[agstrato]

2 hours ago, feartech said:

you could try adding:

 

<key>hda-gfx</key>
<string>onboard-1</string>

in the PciRoot(0x0)/Pci(0x2,0x0) section

 

 

 

Thank you for the suggestion, adding right now.

[feartech]

9 hours ago, agstrato said:

Thank you for the suggestion, adding right now.

also this may also help:

https://github.com/corpnewt/SSDTTime

in creating your own HPET.aml and config.plist Find and Replace entries it creates

[agstrato]

On 11/9/2025 at 11:33 AM, feartech said:

also this may also help:

https://github.com/corpnewt/SSDTTime

in creating your own HPET.aml and config.plist Find and Replace entries it creates

Thanks for this! What SSDTs do you suggest for my build (besides the ones already suggested by dortania guide) ? Asus Z390 + 9900KS ? I gave up on trying to make the hdmi sound work it's pretty clear it won't and that's that. Now all i want is to polish my EFI to the max to leave it as good as it can be.

[feartech]

3 hours ago, agstrato said:

Thanks for this! What SSDTs do you suggest for my build (besides the ones already suggested by dortania guide) ? Asus Z390 + 9900KS ? I gave up on trying to make the hdmi sound work it's pretty clear it won't and that's that. Now all i want is to polish my EFI to the max to leave it as good as it can be.

SSDT's are used to fix things / make things work

[LockDown]

5 hours ago, agstrato said:

I gave up on trying to make the hdmi sound work it's pretty clear it won't and that's that.

can i have a look of your config & ioreg. You might be needing a framebuffer-flag

[theroadw]

Missing HDMI Audio could be due to a regression in AppleALC code that unfortunately affects some PCI ID's.
This patch in AppleALC is the issue if that is the case.
 

Edited November 15, 2025 by theroadw
corrected info

[agstrato]

11 hours ago, theroadw said:

Missing HDMI Audio could be due to a regression in AppleALC code that unfortunately affects some PCI ID's.
Try this patch to rule it out
 

Hi there, where exactly do i need to add this? Is it under Kernel > Patches ? And if so what is the procedure for it? Just to confirm the patch is from https://raw.githubusercontent.com/acidanthera/AppleALC/refs/heads/master/Resources/Controllers.plist ?

I'm a little confused on where and how i should add this...

[theroadw]

That is the regression in AppleALC that affects HDMI audio in some 400 Series PCH's.

To counter it, try injecting in your device properties:

	<key>PciRoot(0x0)/Pci(0x1f,0x3)</key>
	<dict>
		<key>device-id</key>
		<data>yAIAAA==</data>
	</dict>

(confirm your PciRoot if different)
you need to use AppleALC and probably will want to also use a layout id, etc...
Haven't tested this with Sequoia (probably doesn't work) but it should with Sonoma and earlier

Edited November 15, 2025 by theroadw
added info

[agstrato]

9 hours ago, theroadw said:

That is the regression in AppleALC that affects HDMI audio in some 400 Series PCH's.

To counter it, try injecting in your device properties:

	<key>PciRoot(0x0)/Pci(0x1f,0x3)</key>
	<dict>
		<key>device-id</key>
		<data>yAIAAA==</data>
	</dict>

(confirm your PciRoot if different)
you need to use AppleALC and probably will want to also use a layout id, etc...
Haven't tested this with Sequoia (probably doesn't work) but it should with Sonoma and earlier

Ok i think i understand now, i need to spoof my device because it is unsupported. It seems my unsupported device-id is "8086:a348". How would this translate to a value for my config.plist? I also found the FakePCID kexts documentation but it's 8 years old and i guess that isn't available anymore right? 

[agstrato]

On 11/14/2025 at 1:40 AM, LockDown said:

can i have a look of your config & ioreg. You might be needing a framebuffer-flag

I am extremely sorry for the late answer, I just lost myself in the responses from everybody and I haven't seen your request. Here is the IOReg and my latest config.plist

Alex’s iMac.zip config.plist

EDIT: I don't know if you need this but unsupported device-id is "8086:a348"

Edited November 15, 2025 by agstrato

[theroadw]

4 hours ago, agstrato said:

Ok i think i understand now, i need to spoof my device because it is unsupported. It seems my unsupported device-id is "8086:a348". How would this translate to a value for my config.plist? I also found the FakePCID kexts documentation but it's 8 years old and i guess that isn't available anymore right? 

Not unsupported, the issue is that there are 2 devices that have mutually exclusive patches in AppleALC, and it is possible that the one that is in the current AppleALC build may not be your audio device, so you just fake it so the patch applies to your device as well.
just make your config look like this to test

[agstrato]

1 hour ago, theroadw said:

Not unsupported, the issue is that there are 2 devices that have mutually exclusive patches in AppleALC, and it is possible that the one that is in the current AppleALC build may not be your audio device, so you just fake it so the patch applies to your device as well.
just make your config look like this to test

Ok, adding that to my config.plist makes the monitor volume bar disappear.

[theroadw]

5 hours ago, agstrato said:

Ok, adding that to my config.plist makes the monitor volume bar disappear.

That's good information, that means the issue is not an obscure PCI device patch.
So you have volume control on the monitor, are you sure hdmi audio works in windows/linux? and not in macos? a volume bar most likely means it is working? something else may be at play

 

[agstrato]

57 minutes ago, theroadw said:

That's good information, that means the issue is not an obscure PCI device patch.
So you have volume control on the monitor, are you sure hdmi audio works in windows/linux? and not in macos? a volume bar most likely means it is working? something else may be at play

 

Tomorrow i will check for this on Windows, it didn't even cross my mind. I will also try to get another cable.