• 1
  • 2(current)
  • 3
  • 4
  • 5
  • 70
Android Passthrough Changes with v17
#16
(2016-09-14, 09:56)CiNcH Wrote: We might get Marshmallow at some point. So still no DTS. Probably in a year, if we get Nougat at all...
Shouldn't MM come vith api v23? And in first post "v23: AC3, EAC3, DTS, DTS-HD".....
So if that's the case (and mm should come soon) and we do get working ac3 and dts than that's it. We can't output anything else anyway from the TV.
#17
Ah right. Marshmallow is V23 and supports DTS and DTS-HD.
#18
Using the latest nightly all that would happen is Kodi would crash on the Shield, but using the one from the 5th (kodi-20160905-06588de-master-armeabi-v7a.apk) works great with passthrough.
Denon X6500H 7.2.4 -> LG OLED65C9P
Main:
NVIDIA Shield Pro (2019)
Other Devices: Apple TV 4K, FireStick 4K Max (2023), Homatics Box R 4K
Retired devices: Zidoo X9S, Xiaomi Mi Box, All the old RPi’s
#19
fixed now
Read/follow the forum rules.
For troubleshooting and bug reporting, read this first
Interested in seeing some YouTube videos about Kodi? Go here and subscribe
#20
Thanks for the quick turnaround @Martijn, testing 'kodi-20160918-cc9c142-master-armeabi-v7a.apk' this morning and it's working great.
Denon X6500H 7.2.4 -> LG OLED65C9P
Main:
NVIDIA Shield Pro (2019)
Other Devices: Apple TV 4K, FireStick 4K Max (2023), Homatics Box R 4K
Retired devices: Zidoo X9S, Xiaomi Mi Box, All the old RPi’s
#21
I don't know if a Kodi developer will be willing to help and test for official Kodi or a sideline Kodi project perhaps with a few test builds?
It might help with other devices that also use IEC pass-through.
I can ask users to provide logs with their testing and send feedback back to the manufacturer if there is something on the software side that need changes again. The java/jni changes for IEC 61937 was done correctly.

Some manufacturers implemented IEC Pass-through(from Android N) and backported it to Marshmallow and Lollipop.
All the audio formats like these work AC3, EAC3, DTS, DTS-HD, TrueHD with Kodi 16.1 and a few extra commits.
But on Kodi 17 betas there are currently no sound or white noise with IEC 61937 pass-through on Marshmallow and Lollipop.
I don't know if the checks for SDK>23 in Kodi 17 can be modified to be SDK >=SDK22 for IEC 61937 in a test build perhaps to get logs.

It will be a while before Android N comes to more devices, but if a test build/project can't be made for IEC 61937 lower than Android N, it is totally fine also.
#22
Already planning on doing this:

https://github.com/fritsch/xbmc/commit/b...6f0575dadd <-

Quote:But on Kodi 17 betas there are currently no sound or white noise with IEC 61937 pass-through on Marshmallow and Lollipop.

How do you use "IEC passthrough" on current beta? Cause I explicitely don't allow that anymore. If you talk of AML boxes - then they - in their great wisdom - overwrote the ENCODING_xy fields but want IEC instead, which is NOT what the Audiotrack API wants them to take. So basically they end up with getting RAW PT .... and try to interpret it as IEC.

So - if you talk of this API - then it's no backport, but a specialized "side development" misusing given API fields.
First decide what functions / features you expect from a system. Then decide for the hardware. Don't waste your money on crap.
#23
(2016-09-18, 18:03)fritsch Wrote: Already planning on doing this:

https://github.com/fritsch/xbmc/commit/b...6f0575dadd <-

Quote:But on Kodi 17 betas there are currently no sound or white noise with IEC 61937 pass-through on Marshmallow and Lollipop.

How do you use "IEC passthrough" on current beta? Cause I explicitely don't allow that anymore. If you talk of AML boxes - then they - in their great wisdom - overwrote the ENCODING_xy fields but want IEC instead, which is NOT what the Audiotrack API wants them to take. So basically they end up with getting RAW PT .... and try to interpret it as IEC.

So - if you talk of this API - then it's no backport, but a specialized "side development" misusing given API fields.

Hi, it's not on Amlogic.
If there is anything that needs to change or improve, I'll be happy to pass it on.
If a test build can be made that would be great.
I send you a PM.
#24
I ignore pms. Just post a Debug Log of your box with the pr I opened. If you cannot build, I will do a testbuild tomorrow.
First decide what functions / features you expect from a system. Then decide for the hardware. Don't waste your money on crap.
#25
(2016-09-18, 22:58)fritsch Wrote: I ignore pms. Just post a Debug Log of your box with the pr I opened. If you cannot build, I will do a testbuild tomorrow.

OK.
I will wait for your test build, that will be quicker.
I will then get some logs of the test build(Kodi 17 beta) and logs from another Kodi 16.1 build where everything worked to compare if there are still problems and post them to help.
#26
That will be yours: http://jenkins.kodi.tv/job/Android-ARM/9232/ once it is on the mirrors.

Only DTS / AC3 / EAC3 is enabled for now as those were the only formats I could test on nexus player with android N. So your Debug Log will help us to implement it properly.

Thanks in advance.
First decide what functions / features you expect from a system. Then decide for the hardware. Don't waste your money on crap.
#27
@ fritsch: Check this out: http://pastebin.com/qeckw16M (line 670). I know only 2 channels are initialized.

Does this mean that Nexus Player can support Dolby TrueHD passthrough on Android N with IEC passthrough?
#28
We already tried the Nexus and it couldn't
Read/follow the forum rules.
For troubleshooting and bug reporting, read this first
Interested in seeing some YouTube videos about Kodi? Go here and subscribe
#29
@wesk05:
Quote:ENCODING_IEC61937 (Added in API level 24)
int ENCODING_IEC61937
Audio data format: compressed audio wrapped in PCM for HDMI or S/PDIF passthrough. IEC61937 uses a stereo stream of 16-bit samples as the wrapper. So the channel mask for the track must be CHANNEL_OUT_STEREO. Data should be written to the stream in a short[] array. If the data is written in a byte[] array then there may be endian problems on some platforms when converting to short internally.

Constant Value: 13 (0x0000000d)

When I have the output from the "real backports box" guy - I will enable TrueHD / DTS-HD again, then we can see. Btw. this is not an SPMC testing thread.
First decide what functions / features you expect from a system. Then decide for the hardware. Don't waste your money on crap.
#30
(2016-09-19, 07:27)fritsch Wrote: @wesk05:
Quote:ENCODING_IEC61937 (Added in API level 24)
int ENCODING_IEC61937
Audio data format: compressed audio wrapped in PCM for HDMI or S/PDIF passthrough. IEC61937 uses a stereo stream of 16-bit samples as the wrapper. So the channel mask for the track must be CHANNEL_OUT_STEREO. Data should be written to the stream in a short[] array. If the data is written in a byte[] array then there may be endian problems on some platforms when converting to short internally.

Constant Value: 13 (0x0000000d)

When I have the output from the "real backports box" guy - I will enable TrueHD / DTS-HD again, then we can see. Btw. this is not an SPMC testing thread.

Hi

The test build crashes with ac3/dts videos.
The same problem was in SPMC.
I think Koying fixed it with this commit.
It had to do with SDK lower than 23 not having the correct write permissions.
https://github.com/koying/libdroidjni/co...0706e67930
Can you please have a look and make a new test build if needed?
When I'm back home I will try to get the logs with the new test build and also for the crashing.
  • 1
  • 2(current)
  • 3
  • 4
  • 5
  • 70

Logout Mark Read Team Forum Stats Members Help
Passthrough Changes with v178