CoreELEC 19.4-Matrix Discussion

Maybe, Amlogic did change a lot in audio area in new kernel. I have no AVR so I can’t check it.
The new lines are there because of the new kernel:

Ok - I thought you had used an AVR to test the output. If I mixdown the 6 channels to 2.0 or fix the sample rate in the Kodi settings, then I get playback, or if I switch off the TV and switch it on again to renew the HDMI handshake the 2.0 playback is restored. The references to SPDIF-A and SPDIF-B are presumably for an Amlogic model with multiple S/PDIF outputs. My N2+ has header pins for the attachment of S/PDIF interface (optical or coaxial), but I don’t use it. I wonder is this something for @tobetter on the Hardkernel forum? There was a previous issue last year where the N2 was outputting 22.05kHz PCM output which almost no AVR would accept and the driver settings were changed so the PCM was in the normal 32kHz - 192kHz range. Maybe this new kernel has introduced some incompatibilities with certain AVRs?

How do you “mix-down” 6 channels to 2.0?
2.0 is not a mix-down setting, it’ simply an option to let ALL channels pass-through to your AV receiver, which then does what’s it is needed to decode all channels contained in audio stream.

I have eliminated the AVR by using a direct HDMI connection to the TV. My TV is just stereo, so the multichannel (5.1) audio sounds not quite correct but it does play. Once I stop 5.1 playback, I can’t hear the GUI clicks, nor can I playback 2.0 audio. I have to select a 2.0 file which has a sampling rate of 48kHz or higher, then when it plays and I press stop the GUI sound returns and I can play regular 44.1kHz audio also. Something wrong with my N2+ I think! It is an early model where the USB3 ports were damaged shortly after I got it. It has been working ok since but without USB3. I only use the USB2 OTG port in the front for my RF keyboard dongle. The recent builds seem to have partly resurrected functionality on the rear USB ports, though.

For clarity, this is PCM multichannel playback, not DD/DTS passthrough. When I connect a sink which can only handle two PCM channels (with Kodi set to 5.1 or 7.1(recommended setting)) then it will only play back the FR and FL channels. That’s what happens when I connect my TV directly to the N2+. If I change Kodi to 2.0 speaker configuration, then it “mixes down” all channels to 2.0 and it sounds ok on the TV.
With my AVR connected, if I have 2.0 speaker configuration selected in Kodi and playback a 5.1 PCM source, the Kodi makes a fairly good attempt at mixing down the six input channels to two output channels. The mixed sound is heard through the L and R speakers only and the Centre and Rear Surround channels are included.

Did you read any recommended sound settings in Kodi Wiki?
There is a specific line explaining on choosing “Number of channels” when not using pass-through, which is your situation when N2 is connected to your TV:
Select the number of channels supported by the audio connection”, which obliges you to set 2.0, unless your TV has (and advertises) more than 2 audio channels. If your TV has only 2 audio channels, then all is behaving as designed and expected; no fault in N2 or Kodi…

A completely different situation is when you connect your N2 to AVR, because then your receiver informs Kodi of it’s multi channel possibilities, and it surely does a much better job of remixing complex multichannel audio to front 2 speaker stereo. So why let N2/Kodi do that?

@Sholander - I think you are missing the point. I am entirely familiar with multichannel playback and how Kodi works with an AVR. The problem I am experiencing with my N2+ and 5.1 playback followed by 2.0 “non-playback” is entirely new and is either a software problem or a hardware issue specific to my N2+. My workaround to silent 2.0 playback after 5.1 playback on my AVR is either to fix the sample rate or play with the speaker settings in Kodi. I never wanted to get my AVR to mix down 6 channels to two. I was just illustrating how I got around the silent playback of 2.0 audio after playing 5.1 audio on the the latest builds of 19.4. Playback on 19.3 is perfect, so it’s related to the new kernel audio settings and possibly my somewhat damaged N2+. When I set Kodi to playback 2.0 speaker config (to suit my TV) or as an experiment on my AVR which Kodi actually mixes down correctly six channels to two, then I don’t have an issue of silent playback. It was just an illustration of a workaround.

OK, understood; sorry I’m so slow :slight_smile:
But, if your N2+ works perfectly with 19.3 than it’s surely not at fault. Hope it gets fixed in CE/Kodi.
Did you try with any CE 20 nightly builds? Or with any Android distro with Kodi.nerds.maven builds (Matrix and Nexus for Android) to track down if it’s Kodi or underlying OS that’s causing such behaviour?

No problem. I kinda knew we were talking at cross purposes. :wink:
I have tested the Nexus (CE20) nightlies and it’s the same. Only the 19.3 and previous (9.2.8) builds are working for me. It seems that it may be related to the newer kernel and recent Amlogic sound changes, but I’m getting good support from @vpeter and @Portisch. What makes me think that my hardware may be a contributory factor is that @Tim_Taylor has tested on his N2+ and he doesn’t have an issue. My N2+ is slightly crippled in that the USB3 ports are damaged, so I can’t use them reliably, although theoretically this should not affect HDMI sound playback. I can work around the problem, though, so it’s not a complete show-stopper for me.

@Joe_90 please try with this test image and report back:

1 Like

Thought I’d give you some quick feedback - I think you have fixed it, thanks :clap:
I have tested with various 5.1 and 5.0 files and GUI navigation sounds are played immediately after and I can switch between 5.1 and 2.0 audio files without any issues.
May I ask what has changed in the build? The logfile looks pretty similar, but I haven’t turned on any debugging.

Many thanks again for your time and efforts (and also to @vpeter).

1 Like

Hmm, an hour later and my N2+ seems to have stopped responding to my remote. I think it’s related to the continual kernel messages trying to start the USB3 hub. Here’s a snip of the tail end of dmesg.

[ 3109.071981@3]- usb 2-1: New USB device found, idVendor=05e3, idProduct=0620
[ 3109.071984@3]- usb 2-1: New USB device strings: Mfr=1, Product=2, SerialNumber=0
[ 3109.071985@3]- usb 2-1: Product: USB3.1 Hub
[ 3109.071987@3]- usb 2-1: Manufacturer: GenesysLogic
[ 3109.073845@3]- hub 2-1:1.0: USB hub found
[ 3109.074125@3]- hub 2-1:1.0: 4 ports detected
[ 3109.425791@3]- usb 2-1: reset SuperSpeed USB device number 31 using xhci-hcd
[ 3110.481765@3]- usb 2-1: reset SuperSpeed USB device number 31 using xhci-hcd
[ 3111.313698@3]- usb 2-1: Device not responding to setup address.
[ 3111.521136@3]- usb 2-1: Device not responding to setup address.
[ 3111.729118@3]- usb 2-1: device not accepting address 31, error -71
[ 3111.909194@3]- usb 2-1: reset SuperSpeed USB device number 31 using xhci-hcd
[ 3112.045087@3]- hub 2-1:1.0: hub_ext_port_status failed (err = -71)
[ 3112.225764@3]- usb 2-1: reset SuperSpeed USB device number 31 using xhci-hcd
[ 3113.409767@3]- usb 2-1: reset SuperSpeed USB device number 31 using xhci-hcd
[ 3114.449764@3]- usb 2-1: reset SuperSpeed USB device number 31 using xhci-hcd
[ 3115.649763@3]- usb 2-1: reset SuperSpeed USB device number 31 using xhci-hcd
[ 3115.788669@3]- hub 2-1:1.0: hub_ext_port_status failed (err = -71)
[ 3115.969762@3]- usb 2-1: reset SuperSpeed USB device number 31 using xhci-hcd
[ 3115.990267@3]- usb 2-1: device firmware changed
[ 3116.016623@3]- hub 2-1:1.0: set hub depth failed
[ 3116.124622@3]- usb 2-1: USB disconnect, device number 31
[ 3116.304714@3]- usb 2-1: new SuperSpeed USB device number 32 using xhci-hcd
[ 3120.641445@3]- usb 1-1: USB disconnect, device number 50
[ 3126.683448@3]s xhci-hcd xhci-hcd.0.auto: xHCI host not responding to stop endpoint command.
[ 3126.683452@3]s xhci-hcd xhci-hcd.0.auto: Assuming host is dying, halting host.
[ 3126.707666@3]s xhci-hcd xhci-hcd.0.auto: HC died; cleaning up
[ 3126.708403@3]- usb 2-1: unable to get BOS descriptor or descriptor too short
[ 3126.708455@3]- usb usb1-port1: couldn’t allocate usb_device
[ 3126.708465@3]- usb 1-2: USB disconnect, device number 3

Yes, it is.

I managed to borrow a UART adapter and got it connected, I am seeing the messages from the built-in bootloader but once it boots there is nothing, with a working system I get nothing until CE is up and I get a shell.

Is there some configuration/boot param that is needed ?

Thanks to @vpeter I have resolved my USB3 errors flooding dmesg by disabling it with a udev rule. The USB3 ports are faulty since I purchased the N2+ but I don’t use them in any case. It seems recent kernels are trying unsuccessfully to continually activate the USB3, but this workaround stops it.

Mod edit: This is the udev rule used to remove USB 3.1 hub on N2+ (file /storage/.config/udev.rules.d/01-disable-genesys-usb31-hub.rules):

ACTION=="add", ATTR{idVendor}=="05e3", ATTR{idProduct}=="0620", RUN="/bin/sh -c 'echo 1 >/sys/\$devpath/remove'"

Hallo,
gibt es schon Neuigkeiten zu Uart Bridge?
Grüße

Hello,
is there any news about Uart Bridge?
Regards

I haven’t got around to it yet.

Danke und Frohe Ostern :grinning:

1 Like

Hi All,

I have kernel oops with my S905X device (Tronsmart Vega S9x) with all ng builds (coreelec 9.2.8 is rock stable).
I can reproduce it with 100% success rate:

  • Open an extension what requires widewine drm
  • Download chromeos recovery with inputstreamhelper.
  • When download is around 50% device hangs and reboots after some seconds

Ramoops Panic Logger successfully captured the kernel oops:
ramoops logs: http://ix.io/3Wha
kodi logs pasted from gui: http://ix.io/3Wh9

Saw similar posts where they mentioned the Android 6 bootloader as the root cause.
I also have android 6 bootloader and no official upgrade. I tried to install generic images (aidan’s rom v9) with bootcard maker, but I got an android robot with red X only.

What is your opinion, is it relally related to the uboot version? If yes, are there any solution to upgrade it without usb burn cable?

Thanks,
madar

Yes, Android 6 is might the root cause. Try again with clean install and new boot media.
No, USB burn cable is not needed: coreelec:aml_burncard [Core Embedded Linux Entertainment Center for Amlogic Devices]

Already tried the bootcard maker, but I have only the original stock android 6 image. Custom roms got the Android robot + red x after booting to recovery.