Kodi Community Forum
AudioEngine branch - DO NOT REQUEST BINARY BUILDS - Printable Version

+- Kodi Community Forum (http://forum.kodi.tv)
+-- Forum: Development (/forumdisplay.php?fid=32)
+--- Forum: Development (/forumdisplay.php?fid=93)
+--- Thread: AudioEngine branch - DO NOT REQUEST BINARY BUILDS (/showthread.php?tid=78289)



RE: AudioEngine branch - DO NOT REQUEST BINARY BUILDS - grote002 - 2012-04-28 02:16

(2012-04-26 22:24)gnif Wrote:  Those that have been having issues with their AVR locking onto DTS-HD MA streams, please give the latest version a go... we believe it is fixed now.

@gnif The DTS HD MA track is working flawless now Smile Thanks so much.

Xbmc crashed a couple times already with the latest version.
Last crash shows the following line in my syslog, nothing in xbmc.log between the video starting playing and the segmentation fault.

Code:
[  302.740736] xbmc.bin[2576]: segfault at 74656e72 ip b5bdfcd6 sp 96efa2d0 error 4 in libnvidia-tls.so.290.10[b5bdf000+3000]

Do i need a newer nvidia-driver or is this an xbmc-problem?


RE: AudioEngine branch - DO NOT REQUEST BINARY BUILDS - gnif - 2012-04-28 06:08

@grote002 - please provide a debug back trace and full log.


RE: AudioEngine branch - DO NOT REQUEST BINARY BUILDS - d_hero - 2012-05-02 01:45

ok im new to compiling and this thread but have been following it it closely for ages. Im not sure if im posting in the correct place as i think this problem occured in XBMC beforehand but seeing as you are developing a new audioengine i thought it would fit in here.

So ive compiled from your latest source running under Linux (Maybe i should post in the Linux only thread?) With regards to HD audio, that seems to be working great for this files I have tested (DTS-HD MA, TrueHD)

The problem ive noticed seems to only be with Dolby Digital 2.0 sources. If I allow my receiver to decode DD2.0, during quiet scenes, where there is low background noise when someone stops talking etc, the sound cuts out completely, then comes back as soon as the level increases. Its almost like the sound is being passed through a really heavy noise gate and its cutting off all sound below a certain level. Its very noticeable especially when people are talking quietly and there is background music, when they stop talking, the sound cuts out and it ruins it. I hope that explains what i mean and you understand.

At first I thought it was my receiver, or the file i was playing, but I have now tried multiple files, weekly TV shows that I have and its the same. Its also the same when using a different receiver so I can rule that out (Although the receiver is the same brand, Onkyo, but completely different and ages)

The strange thing is that when I tell XBMC that I dont have a DD capable receiver, so that it decodes the stream internally, the resulting PCM stream plays back fine, no dropouts or anything at all, and I have tested that again with multiple files.

I dont want to start a war by having posted this in the wrong section, and its not a support request, just something that bugs me massively and I hope it can be fixed with your new audioengine.

If you need me to provide logs or anything please let me know.

I am running under Linux, and use a nVidia GT 520 for HDMI sound and video output. I have also tested this on a machine with nVidia G210 card and the same result occurs. I am using an Onkyo 809 receiver and I have tested with an Onkyo 807 receiver also.


RE: AudioEngine branch - DO NOT REQUEST BINARY BUILDS - gnif - 2012-05-07 16:18

@d_hero - DD is not AC3 which is what I think you mean here, so I will continue based on that assumption. It sounds like a hardware fault, AE does not touch the stream at all when it is in pass-through mode. Please provide a clean log via pastebin, and if possible a sample file that replicates the issue reliably. I also have an Onkyo AVR and I have not noticed this behaviour.


RE: AudioEngine branch - DO NOT REQUEST BINARY BUILDS - d_hero - 2012-05-07 19:03

ok unfortunately I dont have AE installed any longer, so the log would probably be not much use, but heres a sample of an episode of Fringe. I can hear, when he is talking, in between where the background music drops to a certain level, the sound just cuts out completely for split seconds, this happens a number of times throughout this clip. If i tell XBMC that my receiver does not support DD, then the clip plays fine with no dropouts.

Sorry about the audio sync issues, I chopped this out of a full episode and didnt spend any time on the sync


Fringe.mkv

Sorry i cant provide a log and be more help but it would be great if you could at least tell me whats going on. Ive tried this same clip on 2 machines now, different AVRs and graphics card and the same result

Thanks


RE: AudioEngine branch - DO NOT REQUEST BINARY BUILDS - D3LTA - 2012-05-08 14:08

I have te follow error in Openelec with 290 drivers...(how can i check drivers?)

I have a Asus en210 silent Nvidia card in a HPDC 7700

"Error - No devices Found" in System - Audio Settings


Code:
root ~ # aplay -l
**** List of PLAYBACK Hardware Devices ****
card 0: NVidia [HDA NVidia], device 3: HDMI 0 [HDMI 0]
  Subdevices: 1/1
  Subdevice #0: subdevice #0
card 0: NVidia [HDA NVidia], device 7: HDMI 0 [HDMI 0]
  Subdevices: 1/1
  Subdevice #0: subdevice #0
card 0: NVidia [HDA NVidia], device 8: HDMI 0 [HDMI 0]
  Subdevices: 1/1
  Subdevice #0: subdevice #0
card 0: NVidia [HDA NVidia], device 9: HDMI 0 [HDMI 0]
  Subdevices: 1/1
  Subdevice #0: subdevice #0

Photo of error in XBMC Openelec System-settings-audio

do i need to edit settings in asound.conf or in sound.confConfused

Got it working in eden with old audio Engine like this

asound.conf

pcm.!default {
type hw
card 0
device 7
}
ctl.!default {
type hw
card 0
device 7
}
pcm.!hdmi {
type hw
card 0
device 7
}
ctl.!hdmi {
type hw
card 0
device 7
}


RE: AudioEngine branch - DO NOT REQUEST BINARY BUILDS - gnif - 2012-05-08 14:17

@D3LTA - provide your xbmc log, cant do anything without it


RE: AudioEngine branch - DO NOT REQUEST BINARY BUILDS - D3LTA - 2012-05-08 18:58

(2012-05-08 14:17)gnif Wrote:  @D3LTA - provide your xbmc log, cant do anything without it

XBMC.LOG

Complete log zip


RE: AudioEngine branch - DO NOT REQUEST BINARY BUILDS - Drae - 2012-05-08 19:03

I originally posted this in the other AE Linux topic ... probably better here (and I can't delete the other post sigh).

I'm having similar issues since the enumeration capabilities were added as D3LTA is experiencing. I'm also using the same type of card, an Nvidia G210. I've found that initially it can work fine - the ELD data is probed and it correctly identifies the correct device and up pops my Denon receiver in the GUI - everything works fine. If I suspend and resume, again, everything continues to work fine. The moment I stop xbmc or reboot - when it returns no ELD data is found and hence no sound.

The only reliable way I've found of regaining sound via the Nvidia card is through a reinstall of the OS. Simply reinstalling xbmc has never worked in my testing. Removing/reinstalling the nvidia driver does not reliably fix things - messing with ALSA (reinstalling, upgrading, etc.) likewise does not reliably fix anything. Reinstalling the graphics driver likewise does not reliably fix anything. Rebooting - numerous times, doesn't fix it. Leaving it running - nothing, switching it off and returning later, nothing. By default I used stored EDID information when X starts, i.e. I have a line in my xorg.conf of the form:

Code:
Option          "CustomEDID" "DFP-1:/etc/X11/edid.bin"

However it makes no difference if that line is present or not or whether the amp is switched on and tuned to the media player before it boots or not.

I'm fairly certain this is related directly to the G210. Firstly we're not seeing lots of "my sound isn't working!" responses from people which would suggest a wider problem. Secondly when I switch to using the onboard graphics (provided by the Sandybridge Pentium G620) ELD data and sound work fine - from suspend/resume, from reboots, from restarting (all be it just noise for DTS-HD/TrueHD but that's a seperate issue).

So some background and logs. This is an Intel DH67CF motherboard with a Pentium G620 and this addin Nvidia G210 card. Using Ubuntu 12.04 and the 290.10 driver (but I've tried the 285 through to the latest 302 betas with the same issues). I've tried disabling the onboard graphics/hdmi/sound - same result. I've tried using a probe line for the nvidia card, same result. It's something about this card - I realise that it doesn't technically support HD sound formats but it's been working fine since Gnif's early AE builds "long ago". And besides it's the ELD data that is the issue, or apparent lack of ELD data Big Grin

I use the latest git (ae_rebase) from Gnif's repo

Output of aplay -l:

Code:
**** List of PLAYBACK Hardware Devices ****
card 0: PCH [HDA Intel PCH], device 0: ALC892 Analog [ALC892 Analog]
  Subdevices: 0/1
  Subdevice #0: subdevice #0
card 0: PCH [HDA Intel PCH], device 1: ALC892 Digital [ALC892 Digital]
  Subdevices: 1/1
  Subdevice #0: subdevice #0
card 0: PCH [HDA Intel PCH], device 3: HDMI 0 [HDMI 0]
  Subdevices: 1/1
  Subdevice #0: subdevice #0
card 0: PCH [HDA Intel PCH], device 7: HDMI 1 [HDMI 1]
  Subdevices: 1/1
  Subdevice #0: subdevice #0
card 1: NVidia [HDA NVidia], device 3: HDMI 0 [HDMI 0]
  Subdevices: 1/1
  Subdevice #0: subdevice #0
card 1: NVidia [HDA NVidia], device 7: HDMI 0 [HDMI 0]
  Subdevices: 1/1
  Subdevice #0: subdevice #0
card 1: NVidia [HDA NVidia], device 8: HDMI 0 [HDMI 0]
  Subdevices: 1/1
  Subdevice #0: subdevice #0
card 1: NVidia [HDA NVidia], device 9: HDMI 0 [HDMI 0]
  Subdevices: 1/1
  Subdevice #0: subdevice #0

Output from:

cat /proc/asound/NVidia/eld#1.0

Code:
monitor_present         1
eld_valid               1
monitor_name            DENON-AVAMP

connection_type         HDMI
eld_version             [0x2] CEA-861D or below
edid_version            [0x3] CEA-861-B, C or D
manufacture_id          0xee11
product_id              0x15
port_id                 0x20000
support_hdcp            0
support_ai              0
audio_sync_delay        0
speakers                [0x5f] FL/FR LFE FC RL/RR RC RLC/RRC
sad_count               7
sad0_coding_type        [0x1] LPCM
sad0_channels           8
sad0_rates              [0x1ee0] 32000 44100 48000 88200 96000 176400 192000
sad0_bits               [0xe0000] 16 20 24
sad1_coding_type        [0x7] DTS
sad1_channels           6
sad1_rates              [0x6c0] 44100 48000 88200 96000
sad1_max_bitrate        1536000
sad2_coding_type        [0x2] AC-3
sad2_channels           6
sad2_rates              [0xe0] 32000 44100 48000
sad2_max_bitrate        640000
sad3_coding_type        [0xb] DTS-HD
sad3_channels           8
sad3_rates              [0x1ec0] 44100 48000 88200 96000 176400 192000
sad4_coding_type        [0xa] E-AC-3/DD+ (Dolby Digital Plus)
sad4_channels           8
sad4_rates              [0xc0] 44100 48000
sad5_coding_type        [0xc] MLP (Dolby TrueHD)
sad5_channels           6
sad5_rates              [0x1ec0] 44100 48000 88200 96000 176400 192000
sad6_coding_type        [0xc] MLP (Dolby TrueHD)
sad6_channels           8
sad6_rates              [0x6c0] 44100 48000 88200 96000

Log of successful eld data probe: http://pastebin.com/XVeiwcSZ

Log of unsuccessful eld data probe: http://pastebin.com/aCtb8Grg

Copy of guisettings.xml from working AE install: http://pastebin.com/0KGKd0qD

Copy of guisettings.xml from "broken" AE install: http://pastebin.com/hx9K4ZMJ

Obviously getting further data is a little ... complicated ... since the only way I can reliably obtain it is via a reinstall. However I'm more than happy to do that as many times as necessary if it can help resolve this issue.


RE: AudioEngine branch - DO NOT REQUEST BINARY BUILDS - gnif - 2012-05-08 19:08

@D3LTA - The version of AE that openelec has built seems to be quite old (25th April), many many changes have gone in since then, including fixed device enumeration (your problem).

@Drae - your issue is unrelated to D3LTA's, sounds like an ALSA or kernel driver issue, any chance of getting SSH access to your machine for some debugging?


RE: AudioEngine branch - DO NOT REQUEST BINARY BUILDS - D3LTA - 2012-05-08 19:17

@GNIF

Okay are there newer builds as you know off?? would love to stay with openelec...

Can't compile them myself Sad

Or is there something else i can do ??


RE: AudioEngine branch - DO NOT REQUEST BINARY BUILDS - Drae - 2012-05-08 19:25

@gnif - most certainly, anything to help (or just fix my issue Big Grin). I can pm you access details in a bit, currently reinstalling a minimal ubuntu 12.04. I'll do a basic install as I usually do.


RE: AudioEngine branch - DO NOT REQUEST BINARY BUILDS - DDDamian - 2012-05-08 19:31

D3LTA - there's a thread for Linux AE - maybe try there. Otherwise (as the thread topic says), no binaries here. If another user cares to build one for you ofc they're free to do so.


RE: AudioEngine branch - DO NOT REQUEST BINARY BUILDS - gnif - 2012-05-08 19:32

@D3LTA - Read the title of this thread more closely, the big capital letters to be specific. Other then that, wait till its officially supported.

@Drae - Great, I have to sleep some now (3:30am here), but if you can get a complete dmesg dump before XBMC, after XBMC and after reboot it would be great.


RE: AudioEngine branch - DO NOT REQUEST BINARY BUILDS - D3LTA - 2012-05-08 21:28

@GNIF i am truly sorry, wasn't a real binair request.. just as a shared knowledge or something...

This is where i can find newest code? https://github.com/xbmc/xbmc/pull/939