[Windows] AudioEngine testers - Windows platform only

  Thread Rating:
  • 3 Votes - 3.67 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Post Reply
DDDamian Offline
Team-XBMC Developer
Posts: 3,030
Joined: Sep 2011
Reputation: 254
Location: Canada
Post: #121
@wints - thanks for joining in (there's a lot more people downloading than commenting which isn't that helpful lol). Both you and Martijn have reported a crash switching to optical/coax for the first time. It appears that when first run the audio output device string may be gibberish, and when switching to optical it tries to load a device that is just garbage data. I'll take a look at the initialization code and force it to "Default" if the device doesn't match what Windows reports is available.

It sounds like once a device is selected everything carries on okay, so in the meantime I think the workaround is to select an output device first, then the output type, but it's something I'll fix. Thanks for confirming that and let me know how it works for you now that the device is set!

@Martijn - that sucks. See my comments to @wints about the optical selection crash. I believe once you're past that it runs okay though, correct? Just the DTS files with the x264 L5 encoding were off/choppy?
find quote
DDDamian Offline
Team-XBMC Developer
Posts: 3,030
Joined: Sep 2011
Reputation: 254
Location: Canada
Post: #122
(2012-04-07 13:38)Aemstel Wrote:  
(2012-04-06 21:19)DDDamian Wrote:  @Aemstel - for sure get that skin out and re-check your logs to make sure you don't see any sign of that spam-bug (pvr.anything) as it will affect playback.

I ran your new build in portable mode so there would be no Aeon skin, and the pvr. errors were indeed gone. It did not help with the "switching-bug" though Sad It still happens with DTS HDMA audio. Again, Dolby TrueHD seems fine.

I posted the errors in the log for a dts hdma and a dolby true hd sample here: http://pastebin.com/XHBwfceG
I don't see much difference, except the discontinuity error seems larger for the dts hdma sample...?

Hope it helps.

This is curious - especially the line:

Code:
3.13:15:54 T:1540   ERROR: CAESinkWASAPI::InitializeExclusive: IsFormatSupported failed (AUDCLNT_E_UNSUPPORTED_FORMAT) - trying to find a compatible format

That means the details supplied about the audio stream don't meet the spec for DTS-MA formatting or your card/driver doesn't support DTS-MA/TrueHD. Could you let me know what card and driver version you have?

Also check that DTS-HD and Dolby TrueHD show up in your driver settings as in the first picture below. In the second picture make sure the two check-boxes are ticked to enable WASAPI exclusive which is required for these formats.

Also everyone, pls post full logs as they give me the whole picture. There's a great new addon to make this easier - see thread here.

[Image: 33zfods.jpg]

[Image: k3n788.jpg]

System: XBMC HTPC with HDMI WASAPI & AudioEngine - Denon AVR-3808CI - Denon DVD-5900 Universal Player - Denon DCM-27 CD-Changer
- Sony BDP-S580 Blu-Ray - X-Box 360 - Android tablet wireless remote - 7.1 Streem/Axiom/Velodyne Surround System
If I have been able to help feel free to add to my reputation +/- below - thanks!
find quote
DDDamian Offline
Team-XBMC Developer
Posts: 3,030
Joined: Sep 2011
Reputation: 254
Location: Canada
Post: #123
For anyone experiencing the switching bug:

I was hoping the change to the older header format would fix this - have only heard back from one person so far, so not sure yet if it helped. I am beginning to suspect that because we are using the most advanced of the WASAPI output modes (event-driven exclusive) older hardware does not have drivers which support this relatively new and fairly advanced dual-buffering system.

This was changed from the more common push-mode method a few months ago, and obviously works well for many, including myself.

So far the only concrete data I have is from ix400 that it happens on an older ATI 6310, and not sure yet if this new build fixes his yet, but I also note that he had troubles way back in Sept 2011 per this post here, which was before we switched to the event-driven mode, making me think it's his older card causing the issue. Hopefully we'll hear back from him with the new build.

If I'm going to track this down it is imperative that I get your full "clean" Debug Log, meaning open XBMC with debugging already enabled. wait 30sec for start-up tasks to finish, reproduce the issue and exit without a lot of extra activity.

I also need to know the audio driver date (go to the windows audio panel, select properties of the device, properties again then "Driver") to make sure you have the latest version.

Without those details I'm really just guessing and that defeats the purpose of this thread, not to mention doesn't lead to a fix!
find quote
ix400 Offline
Senior Member
Posts: 201
Joined: Aug 2010
Reputation: 0
Location: Germany
Post: #124
DDDamian, I will test the new build on monday (I'm on a short vacation right now). I will post a clean debug log then.

You mentioned my September post. I would like to add that at that time I used a different system: Sandy Bridge i7, a Z68 board from Asrock and a Powercolor 6970.

Btw, does DanielA's patch use a different WASAPI output mode?

Have a good weekend,

Chris

unRAID Server, Pioneer VSX-LX52, Panasonic 65VT30, Marusys VU+ Duo, Zotac AD10
(This post was last modified: 2012-04-07 17:52 by ix400.)
find quote
Aemstel Offline
Member
Posts: 82
Joined: Feb 2011
Reputation: 0
Post: #125
These are my audio properties:

[Image: nfh3ir.jpg]

I also have those 2 boxes ticked in the advanced tab.
I have a GeForce GTX 560 Ti graphics card with HDMI output. Tried driver version 290.53 and updated to latest 296.10 (this comes with HD audio driver 1.3.12.0). Same results.

Also, DanielaE's patched version runs all HD audio formats that I've tried just fine.

Here's a full log when running a DTS MA with the switch bug: http://pastebin.com/J3Mp2Jbd
(This post was last modified: 2012-04-07 18:37 by Aemstel.)
find quote
wints Offline
Senior Member
Posts: 201
Joined: Apr 2011
Reputation: 0
Post: #126
(2012-04-07 17:06)DDDamian Wrote:  Also everyone, pls post full logs as they give me the whole picture. There's a great new addon to make this easier - see thread here.

DDDamian, ive done a debug log and used the addon you suggested dont know if you want it or not but this is the link i got emailed http://xbmclogs.com/show.php?id=1270 hope i've done it right lol

7.1 Cinema Room HTPC with XBMC Gotham~Epson TW3200~Homemade 7ft pj screen~AVR~Yamaha RX-A2020~Speakers~Monitor Audio RX6 Fronts & RX Centre~RXFX Surrounds~180's Rears~BK XXLS400 Sub~8033C AntiMode~HarmonyOne


find quote
DDDamian Offline
Team-XBMC Developer
Posts: 3,030
Joined: Sep 2011
Reputation: 254
Location: Canada
Post: #127
Hey - awesome all of you - that's what i need! Give me some time to review.

@ix400 - great - I thought your system was different then - have a great holiday and look forward to your results Smile

The main output difference between AE and Daniela's patch for DTS-MA is that AE uses the more advanced event-driven WASAPI mode, so it's sounding more and more like we need to accomodate both this mode and the older push method. That is something I can add back in with a bit of effort. Not easy, but the indications so far are that not all GPU's / sound-cards / drivers are supporting this.

If a review of all your info supports this conclusion I'll add that pull-method in at the next build.
(This post was last modified: 2012-04-07 18:56 by DDDamian.)
find quote
ix400 Offline
Senior Member
Posts: 201
Joined: Aug 2010
Reputation: 0
Location: Germany
Post: #128
I tried a lot to find a perfect XBMC system... Haven't found it yet. My Pioneer receiver seems to be a real show stopper. I also tested the AMD MacMini for a short time: Same problem with AE builds running on Win 7.

Chris

unRAID Server, Pioneer VSX-LX52, Panasonic 65VT30, Marusys VU+ Duo, Zotac AD10
find quote
DDDamian Offline
Team-XBMC Developer
Posts: 3,030
Joined: Sep 2011
Reputation: 254
Location: Canada
Post: #129
@ix400 - Chris, beyond what I posted above, not sure if it's the event-driven mode, your GPU or your receiver which is the issue. Looks like you've had a couple of GPU's so i think we can rule that out. I can't test or rule out your receiver, but it seems like the issue was the same with both event-driven and non-event driven modes, on two different GPU's.

If you're still in the search for a perfect card, I can endorse the ATI 6450 as perfect for A/V sync on 24p content and for flawless support for HD audio. That and they're dirt cheap right now Smile

Also, DanielaE's patch is built on the existing XBMC audio system, which uses the pull-mode of Wasapi, and AE uses the event-driven mode. The early versions of AE used pull-mode, but that was changed on Jan. 26th of this year to event-driven. If a someone who has this switching issue volunteers to test with the old pull-method I'd be happy to do a custom build - PM me. If that confirms it then I'll make it user-selectable, if it still happens I can rule that out.
(This post was last modified: 2012-04-08 03:00 by DDDamian.)
find quote
Mav78g Offline
Junior Member
Posts: 1
Joined: Apr 2012
Reputation: 0
Post: #130
hi,
My first post so be gentle.... Blush
i tried the latest build and the previous build and it plays most type of HD and multichannel music i throw at it but having an issue with DTS stream 24bit music,. It appears to play the track but i get no sound and the track time counter appears to be accelorated i.e. it counts faster than normal/ These music types play fine in the latest HD branch.build.

Mediainfo reports the .dts file as

Container and general information
Wave 44.6 MiB, 4mm 25s
1 audio stream: DTS
First Audio Stream
1 411.2 Kbps, 44.1Khz, 24 bits, 6 channels, DTS (LE/14)

my rig is intel based with core 2 duo and Nvidia Geforce GT 520 HDMI to onkyo amp

Thanks
find quote
DDDamian Offline
Team-XBMC Developer
Posts: 3,030
Joined: Sep 2011
Reputation: 254
Location: Canada
Post: #131
@Mav78g - I'm pretty newb-friendly lol - welcome to the forums. It looks like you're playing a DTS-wav (DTS-CD) file, and we just haven't added support for that format yet - one of the next things on the list to do.

Could you please let me know what other formats you've tested and the results on those? Thanks.
find quote
abudabi Offline
Senior Member
Posts: 109
Joined: Aug 2011
Reputation: 1
Post: #132
Thanks for all the windows work DDDamian. Might persuade me to ditch lilnux for a bit Smile Do you know if the windows/linux builds are in sync?
find quote
DDDamian Offline
Team-XBMC Developer
Posts: 3,030
Joined: Sep 2011
Reputation: 254
Location: Canada
Post: #133
Hi abudabi - the Linux code is a bit behind right now as gnif (the main dev of AE) has been away for a bit. If you like you can try to compile his code from https://github.com/gnif/xbmc and post your comments in the main AE thread linked in the first post of this one.

With gnif being away and the main thread being quite large already I started this thread for testing of my contributions which are genrally Windows related and so progress could (hopefully) continue unabated Smile

My branch will compile okay in Linux, but it would have only a few advantages over gnif's right now.
find quote
abudabi Offline
Senior Member
Posts: 109
Joined: Aug 2011
Reputation: 1
Post: #134
Cool... will attempt it. I'm not a linux guru by any stretch of the imagination so have to figure out how to have 2 versions of xbmc on my machine 1st (and oh yeah.. compile my 1st program Smile ).

Keep up the good work!
find quote
liquidskin76 Offline
Skilled Skinner
Posts: 1,358
Joined: Aug 2009
Reputation: 17
Location: Arkham, Gotham City
Post: #135
Hey DDDamian,

Just got round to quickly testing your latest build. Big Grin

Strange one... some TrueHD and DTS-HD movies are spot on for AV sync (when running 'match refresh rate' at 24p). However some others are about 100ms out. Not sure why some are fine and others are out? Any ideas? All are blu-ray backups. Sync is good under normal playback circumstances on other players. Haven't tested sync yet with 'match refresh rate' off yet on your ae build.

Also, seeking is getting good! However there's still a short period where there's no audio after seek. Happens with buffer set to 50ms or the 120ms default. Only happens on hd audio bitstream by the way.

Do you need debug logs for any of these?

Otherwise all is looking great. Using your ae as my default install now!

Cheers Wink

HTPC: Streacom F1C | Asus H87I-Plus ITX | Intel Core i5-4670T | Kingston HyperX SSD | Kingston DDR3 RAM
Home AV: Sony KDL-40HX723 | Sky+HD | Sony BDP-S480 BD Player | Sony STR-DH820 AV Receiver | Sony 5.1 Speaker System | Sony PS3 Slim | Logitech Harmony Touch
find quote
Post Reply