[fritsch]

Again the question: Does it play the master audio, do you see this on your AVR or does it play the core only?

[bimme]

It is playing the Master Audio, Or HighResolution Audio...

[baijuxavior]

I want to see it with my own eyees first :-) Can you post that file? Or the specific config part?

Problem seems to be related to the channel count. It was 8 when I had the problem. Changing to 2 didn't work, changed to 1 and problem solved. However the bug is still there. You can reproduce it by setting hdmi as audio and passthrough device, set channel to 5.1 or 7.1. Play a movie. Then change the device to SPDIF, AAC is not converted to DD.

The audio settings that didn't work for AAC to DD:

Code:

<audiooutput>
        <ac3passthrough>true</ac3passthrough>
        <audiodevice>WASAPI:{7FCA8888-E2F1-43A5-AD8B-104369151FE5}</audiodevice>
        <channels>8</channels>
        <config>2</config>
        <dtshdpassthrough>true</dtshdpassthrough>
        <dtspassthrough>true</dtspassthrough>
        <eac3passthrough>false</eac3passthrough>
        <guisoundmode>1</guisoundmode>
        <normalizelevels>false</normalizelevels>
        <passthrough>true</passthrough>
        <passthroughdevice>WASAPI:{7FCA8888-E2F1-43A5-AD8B-104369151FE5}</passthroughdevice>
        <processquality>20</processquality>
        <samplerate>48000</samplerate>
        <stereoupmix>false</stereoupmix>
        <streamsilence>1</streamsilence>
        <truehdpassthrough>true</truehdpassthrough>
    </audiooutput>

Setting that converted AAC to DD:

Code:

<audiooutput>
        <ac3passthrough>true</ac3passthrough>
        <audiodevice>WASAPI:{7FCA8888-E2F1-43A5-AD8B-104369151FE5}</audiodevice>
        <channels>1</channels>
        <config>2</config>
        <dtshdpassthrough>false</dtshdpassthrough>
        <dtspassthrough>true</dtspassthrough>
        <eac3passthrough>false</eac3passthrough>
        <guisoundmode>1</guisoundmode>
        <normalizelevels>true</normalizelevels>
        <passthrough>true</passthrough>
        <passthroughdevice>WASAPI:{7FCA8888-E2F1-43A5-AD8B-104369151FE5}</passthroughdevice>
        <processquality>30</processquality>
        <samplerate>48000</samplerate>
        <stereoupmix>false</stereoupmix>
        <streamsilence>1</streamsilence>
        <truehdpassthrough>false</truehdpassthrough>
    </audiooutput>

[FernetMenta]

should be fixed when this goes in: https://github.com/xbmc/xbmc/pull/3681

[edrikk]

Awesome! Will give this a run tomorrow night when I get home (after nightly is built).

Great work guys!

[steph_tsf]

On your side, is there a slight chance, getting the beginning of an API for inserting external audio DSP in the audio path?
On my side, for writing the audio DSP, shall I rely on PortAudio or on some other library or framework ? What do you advise?
Basically I would write some external audio DSP taking the two audio channels of a stereo downmix, then splitting it over two times four channels, for implementing a digital crossover for stereo active 4-way speakers. The eight audio channels should be injected into HDMI using the HDMI multichannel LPCM modality. This avoids the need for interfacing fancy USB audio multichannel attachments.
When it's not a movie that's playing, there would be an image carried out on HDMI, for keeping HDMI alive.
The listening volume would be managed using the HDMI CEC.

[FernetMenta]

I have replied to your thread in the feature section. I recommend that we continue discussion there and not hijacking this testing thread.

[steph_tsf]

I have replied to your thread in the feature section. I recommend that we continue discussion there and not hijacking this testing thread.

Sorry about that. My mistake. Thanks.

[edrikk]

Awesome! Will give this a run tomorrow night when I get home (after nightly is built).

Great work guys!

sorry for the late reply, but I did want to confirm that this appears fixed.
Awesome, and thank you!

[ibazzi]

The navigation sounds is delayed about 1s on my android device(rk3188).

Sorry my cat has a headache, cause of no debug log.

sorry, log is here
http://xbmclogs.com/show.php?id=86475

Hi devs.
Is there any news about this issue?

[fritsch]

I talked with our Android dev and he told: "Second sound should be faster, when buffer is already filled a bit" :-)

[ibazzi]

I talked with our Android dev and he told: "Second sound should be faster, when buffer is already filled a bit" :-)

thanks for the reply.
So what should I do? or there will be a patch to fix this?

[lmyllari]

I'm having HDMI PCM output trouble on Haswell NUC. fritsch already checked my sample, and didn't encounter the same issues.

5.1 FLAC audio track from https://www.dropbox.com/s/1jsimqjlllkhu2...ion%29.mkv gets output as 7.1 (speaker configuration set to 7.1, output "Best Match"). Log at http://sprunge.us/HYjb

I set the speaker configuration to 5.1, rebooted and tried again. This time output is 5.1, but sound is distorted. Log at http://sprunge.us/RBDD

I set the speaker configuration back to 7.1, disabled pass-through, rebooted and tried a 5.1 truehd track. Playback was fine. I stopped, switched speaker configuration to 5.1 and tried again, getting the same distorted output as with 5.1 FLAC. Log of these both at http://sprunge.us/LUIS

I tried a 5.1 FLAC from 2l.no. Both 7.1 and 5.1 speaker configurations produced distorted output. Log: http://sprunge.us/dcJT

These are with current OpenELEC master. I haven't noticed any problems with ac3 and dts pass-through. HD pass-through fails sometimes when resolution switching is on, producing no output and leaving the system in a state where it's no longer possible to open the audio device. I will try get logs of this too, although I suspect it's a driver issue.

[FernetMenta]

The driver has an incorrect channel map. You may need to create asound.conf and correct it.

[lmyllari]

The driver has an incorrect channel map. You may need to create asound.conf and correct it.

This pull request from fritsch fixes the 5.1 output as 7.1 issue: https://github.com/xbmc/xbmc/pull/3721

Anssi asked me to test "speaker-test -c6 -Dhdmi:CARD=MID,DEV=0 -F S32_LE -p 18958 -P 3" and that produces the distorted output too. The noise is spread through multiple channels.

Trying a few different speaker-test commands puts the device in a state where no output will be generated any more, and speaker-test gives the following error:

Code:

Write error: -5,Input/output error
xrun_recovery failed: -5,Input/output error
Transfer failed: Input/output error

I hope fixing this bug will also fix the HBR audio pass-through.

These are clearly not xbmc issues.