Re: [alsa-devel] no front speaker sound with ALC262
At Fri, 29 Feb 2008 17:24:03 +0100, Rene Dohmen wrote:
No I still don't know where the speakers are conencted. I asked the support engineers at tangent to send me a detailed blockdiagram of their sound solution and ik think i'll receive that sooner or later.
So, for now, I just tried passing the EAPD option to all NIDS in the codec#0 file.
I attached the code#0 file while working with very low volume
When I run sudo hda-verb/hda-verb /dev/snd/hwC0D0 0x03 SET_EAPD 0x02 or sudo hda-verb/hda-verb /dev/snd/hwC0D0 0x02 SET_EAPD 0x02
Doh, I actually misunderstood that your codec were ALC268 but it's ALC262. EAPD command can be issued only on widgets that support. ALC262 has no EAPD pin caps at all. so it doesn't work like that.
Instead, ALC262 (and other Realtek codecs) has a special sequence with coefficients to the vendor-defined widget. Yes, it's a black magic.
Try the following sequence:
0x20 SET_COEF_INDEX 0x07 0x20 SET_PROC_COEF 0x3060
0x3060 can be different numbers like 0x3070 or 0x3050. If you set a different number, reset the coef index again (i.e. always call as a pair).
Takashi
Hi List,
I tried it with the NID's i think are connected to headphone and speaker but so far no luck.
I'll have more time next friday but i was just wondering: Is there a way to be 100% sure that this model is a new, not supported model yet?
Can you explain what numbers i can use instead of the 0x3060?
I am trying to get the datasheets of the soundsystem used, i think some custom board is used tot wire up the soundcard tot the internal speakers.
Thanks
Grz Rene
Op Feb 29, 2008, om 5:57 PM heeft Takashi Iwai het volgende geschreven:
Instead, ALC262 (and other Realtek codecs) has a special sequence with coefficients to the vendor-defined widget. Yes, it's a black magic.
Try the following sequence:
0x20 SET_COEF_INDEX 0x07 0x20 SET_PROC_COEF 0x3060
0x3060 can be different numbers like 0x3070 or 0x3050. If you set a different number, reset the coef index again (i.e. always call as a pair).
At Tue, 4 Mar 2008 14:37:39 +0100, Rene Dohmen wrote:
Hi List,
I tried it with the NID's i think are connected to headphone and speaker but so far no luck.
The coef verbs are only for the realtek vendor-specific widget, not for other pins.
I'll have more time next friday but i was just wondering: Is there a way to be 100% sure that this model is a new, not supported model yet?
No. You'd better to ask the hardware vendor than me. If the HG version still doesn't work, it's not supported by ALSA driver yet. That's all what I can tell.
Can you explain what numbers i can use instead of the 0x3060?
Ask Realtek. It's their black magic :)
I am trying to get the datasheets of the soundsystem used, i think some custom board is used tot wire up the soundcard tot the internal speakers.
That's uncommon. It's possible that BIOS or ACPI setup influences on the behavior, but it's a rare case as well.
Takashi
Thanks
Grz Rene
Op Feb 29, 2008, om 5:57 PM heeft Takashi Iwai het volgende geschreven:
Instead, ALC262 (and other Realtek codecs) has a special sequence with coefficients to the vendor-defined widget. Yes, it's a black magic. Try the following sequence: 0x20 SET_COEF_INDEX 0x07 0x20 SET_PROC_COEF 0x3060 0x3060 can be different numbers like 0x3070 or 0x3050. If you set a different number, reset the coef index again (i.e. always call as a pair).
Rene Dohmen wrote:
Hi List,
I tried it with the NID's i think are connected to headphone and speaker but so far no luck.
I'll have more time next friday but i was just wondering: Is there a way to be 100% sure that this model is a new, not supported model yet?
Can you explain what numbers i can use instead of the 0x3060?
I am trying to get the datasheets of the soundsystem used, i think some custom board is used tot wire up the soundcard tot the internal speakers.
Thanks
Grz Rene
Op Feb 29, 2008, om 5:57 PM heeft Takashi Iwai het volgende geschreven:
Instead, ALC262 (and other Realtek codecs) has a special sequence with coefficients to the vendor-defined widget. Yes, it's a black magic.
Try the following sequence:
0x20 SET_COEF_INDEX 0x07 0x20 SET_PROC_COEF 0x3060
0x3060 can be different numbers like 0x3070 or 0x3050. If you set a different number, reset the coef index again (i.e. always call as a pair).
Rene, Just wanted to let you know that if you still haven't gotten sound working with the Tangent 8000S (we haven't), you should look into its counterpart. The Tangent 8000S is really the MSI Crystal 945 barebones unit, with parts installed and optionally Windows installed. MSI has a newer unit with identical case and external connections - the Crystal A410 . It uses an ATI chipset, but still an ALC262 codec. We got one in today and audio "just worked" for us! I suspect that the BIOS on the Crystal A410 is giving ALSA the info it needs when it auto-probes the BIOS, unlike the BIOS on the Crystal 945 (Tangent 8000S). So, you might want to check it out. Unfortunately for us, we already have several of the Crystal 945's deployed to customers, so we still need to try and resolve the audio issue with those units.
Regards,
participants (3)
-
Matthew R Hurne -
Rene Dohmen -
Takashi Iwai