I didn't use drm-tip, I just picked the single fix patch from GitHub and dropped it and the HDMI (alsa) reorder patch in packages/linux/patches and rebuilt the image; along with a general bump to RPi ffmpeg 6.1.1 sources which is in my Amlogic dev branch.
Using actual drm-tip probably isn't too hard, but you'll probably need to pull sources from GitHub or GitLab using a githash instead of using a release number. See how the Amlogic sources are handled for Linux; it'll be similar.
These test files are courtesy of HomerJau (Assuming it is the same user on QuadrophonicQuad):
https://drive.google.com/file/d/1dWB93r…iew?usp=sharing
Yea. That’s me.
Opening all the channels in that screen grab does not mean all the output channels are sending data. For the Android Kodi quad fix all the only 4 channels have data so the output is pure Quad, even though all the output channels are active (all but ‘quad’ are silent). Anyway the true test is to listen to each speaker not look at the AVR graphic.
I didn't use drm-tip, I just picked the single fix patch from GitHub and dropped it and the HDMI (alsa) reorder patch in packages/linux/patches and rebuilt the image; along with a general bump to RPi ffmpeg 6.1.1 sources which is in my Amlogic dev branch.
Using actual drm-tip probably isn't too hard, but you'll probably need to pull sources from GitHub or GitLab using a githash instead of using a release number. See how the Amlogic sources are handled for Linux; it'll be similar.
I can confirm the HD Audio/LSPCON fix is working as part of Kernel 6.6.31. The change seems slightly more involved than the single line patch file, but works as expected.
kernel/git/stable/linux.git - Linux kernel stable tree
| drm/i915/audio: Fix audio time stamp programming for DP | Chaitanya Kumar Borah |
I can confirm the HD Audio/LSPCON fix is working as part of Kernel 6.6.31. The change seems slightly more involved than the single line patch file, but works as expected.
https://git.kernel.org/pub/scm/linux/…/log/?h=v6.6.31
drm/i915/audio: Fix audio time stamp programming for DP Chaitanya Kumar Borah
Nightlies just went to 6.6.31 for RPi; hopefully the generic kernel will be updated soon too.
If you could confirm any behaviour differences for refresh rate and PCM channel support when using the 8K input on your Denon that would be really useful.
Nightlies just went to 6.6.31 for RPi; hopefully the generic kernel will be updated soon too.
If you could confirm any behaviour differences for refresh rate and PCM channel support when using the 8K input on your Denon that would be really useful.
I tried this. The Denon input options are also set to "8k Enhanced". If I set to just "Enhanced", then the behaviour is the same as my CBL/SAT (non-8k) input.
Firstly, I have an HDMI switch before my Denon, and using the 8k input in 8k Enhanced mode, I could not get LibreElec to show any output. My Sky Q works fine.
Connected directly from LibreElec to the 8k input.
Following refresh rates worked at 4k: 60/59.94/50/30/29.97/25/24.
Following refresh rates did not work at 4k: 23.98. This means 4k/24p switching does not work (I get no output).
I confirmed the channel mappings for 3.0/4.0 etc all worked as expected in all scenarios.
HTH.
Display MoreI tried this. The Denon input options are also set to "8k Enhanced". If I set to just "Enhanced", then the behaviour is the same as my CBL/SAT (non-8k) input.
Firstly, I have an HDMI switch before my Denon, and using the 8k input in 8k Enhanced mode, I could not get LibreElec to show any output. My Sky Q works fine.
Connected directly from LibreElec to the 8k input.
Following refresh rates worked at 4k: 60/59.94/50/30/29.97/25/24.
Following refresh rates did not work at 4k: 23.98. This means 4k/24p switching does not work (I get no output).
I confirmed the channel mappings for 3.0/4.0 etc all worked as expected in all scenarios.
HTH.
Thank you for trying that. I don't have the same option for my Arcam. Interestingly the Arcam/Harman units use the same brand of chipset (Panasonic) as your Denon/Marantz unit and have the same 2in/2out hdmi processor but the switching chips are the HDMI 2.0 model on your 2700h where as the HDMI 2.1 Arcams use the same HDMI 2.1 switching solution as the Denon X800h units.
I will get other two main brands of LSPCon to test with later today. Hopefully patterns/consistencies will start to emerge if more people contribute to this thread.
Well some interesting results with PS176 and mcdp2900 based LSPCons.
The PS176 product has no HDR support otherwise behaves impeccably. The mcdp2900 supports HDR and all refresh rates but is the only product after the audio patch which fails to playback bitstream audio and suffers from the same limited multichannel PCM support as the VMM7100. fezster Was the first converter you tried Megachips based?
I'm not sure - could not really find much info on it: https://www.amazon.co.uk/gp/product/B07…e?ie=UTF8&psc=1
Krobar Lets drag that LSPCon conversation here. I will dust of my i3-8100 and connect it again to my home cinema setup. One last time just for some tests. I use a N100 ans so far i am very happy (one problem, but i found a workaround).
Krobar Lets drag that LSPCon conversation here. I will dust of my i3-8100 and connect it again to my home cinema setup. One last time just for some tests. I use a N100 ans so far i am very happy (one problem, but i found a workaround).
These are my results for multiple LSPCons:
PS175 - Internal:
DP branch device present: yes
Type: HDMI
ID: 175IB0
HW: 1.2
SW: 8.64
Max TMDS clock: 600000 kHz
Min TMDS clock: 25000 kHz
Max bpc: 12
Windows HDR: Yes
Windows 4K Refresh Rates: All with HDR; <=50HZ without HDR broken
Linux HDR: Yes
Linux BT2020: Yes
Linux HBR Audio Bitstream: Yes
Linux Multichannel PCM Support: Excellent
Linux 4K Refresh Rates: All with HDR; <=50HZ without HDR broken
PS175 (Older Firmware) - Internal:
DP branch device present: yes
Type: HDMI
ID: 175IB0
HW: 1.2
SW: 8.44
Max TMDS clock: 600000 kHz
Min TMDS clock: 25000 kHz
Max bpc: 12
Windows HDR: No (Does work on Windows 10 1809 or earlier)
Windows 4K Refresh Rates: All
Linux HDR: No
Linux BT2020: No
Linux HBR Audio Bitstream: Yes
Linux Multichannel PCM Support: Excellent
Linux 4K Refresh Rates: All
Mcdp2900:
DP branch device present: yes
Type: HDMI
ID: MC2907
HW: 1.4
SW: 11.116
Max TMDS clock: 600000 kHz
Min TMDS clock: 25000 kHz
Max bpc: 16
Windows HDR: Yes
Windows 4K Refresh Rates: ALL
Linux HDR: Yes
Linux BT2020: No
Linux HBR Audio Bitstream: No
Linux Multichannel PCM Support: Ok (Only support 5.1 & 7.1 in some setups)
Linux 4K Refresh Rates: ALL
PS176:
DP branch device present: yes
Type: HDMI
ID: 176GB0
HW: 1.0
SW: 7.85
Max TMDS clock: 600000 kHz
Min TMDS clock: 25000 kHz
Max bpc: 12
Windows HDR: NA
Windows 4K Refresh Rates: ALL
Linux HDR: NA
Linux BT2020: NA
Linux HBR Audio Bitstream: Yes
Linux Multichannel PCM Support: Excellent
Linux 4K Refresh Rates: ALL
VMM7100:
DP branch device present: yes
Type: HDMI
ID: SYNAq
HW: 1.0
SW: 7.2
Max TMDS clock: 600000 kHz
Min TMDS clock: 25000 kHz
Max bpc: 12
Windows HDR: Yes
Windows 4K Refresh Rates: ALL
Linux HDR: Yes
Linux BT2020: No
Linux HBR Audio Bitstream: Yes
Linux Multichannel PCM Support: Ok (Only support 5.1 & 7.1 in some setups)
Linux 4K Refresh Rates: ALL
PS186:
DP branch device present: yes
Type: HDMI
ID: 186GA1
HW: 0.1
SW: 2.52
Max TMDS clock: 600000 kHz
Min TMDS clock: 25000 kHz
Max bpc: 12
Windows HDR: Yes
Windows 4K Refresh Rates: ALL
Linux HDR: Yes
Linux BT2020: No
Linux HBR Audio Bitstream: Yes
Linux Multichannel PCM Support: Yes
Linux 4K Refresh Rates: ALL
Notes: Weird colour space issues n Linux. LSPCon crashes on change of audio format so not really useable
Linux HBR Audio required the recent Linux patch in all cases.
Wow that is very diligent.
I fear i can not add anything to that.
It was cheaper for me (if time is money) to use my Vero 4k+ in the last years for contents that demanded HBR and/or HDR.
Now i use my N100.
But as i said i will setup my i3-8100 one more time and do some test, but for sure not as many as you did.
Wow that is very diligent.
I fear i can not add anything to that.
It was cheaper for me (if time is money) to use my Vero 4k+ in the last years for contents that demanded HBR and/or HDR.
Now i use my N100.
But as i said i will setup my i3-8100 one more time and do some test, but for sure not as many as you did.
In retrospect it wasn't a great use of my time; I've mainly proved what was already suspected that LSPcons are near universally bugged and the bugs vary quite significantly and are a lot worse for Linux than Windows. As general advice I will say the newer VMM7100 based HDMI 2.1 converters actually seem to be the best choice right now although will be interested in any results you have (As you have different converters/LSPCons).
The display refresh restrictions of the PS175 are at least easy to workaround with no quality loss with careful whitelisting (I discovered I have no 4K/50/25 non-hdr content) so at least I have a good working solution and it is nice to be off of Windows. The last remaining for me is the lack of bit perfect multichannel PCM with recent builds but I need to try some other builds to narrow down the cause (I'm not sure if it is actually an LSPCon issue, might be Kodi or something else).
Here is a LE12 with kernel 6.9.5 to test https://sky42.libreelec.tv/release/12.0/G…020/kernel-6.9/. No encryption or lvm yet.
This image contains the patch from drm-tip (which was no issue to backport onto the LE kernel sources) and the channel allocation reorder patch that was suggested with RPi recently: https://chewitt.libreelec.tv/testing/LibreE…_64-12.80.0.tar
Let me know if the issue is resolved and/or whether the 4.0 > 5.1 workaround is required?
Could I be cheeky and ask for a build of 12.0 stable with that patch applied? I suspect this patch breaks bitperfect multichannel output in my setup (Bitperfect stereo / DTS CD is OK), this potential build would really help me confirm that and that it is not something else in the nightlies since.
