LAV Filters - DirectShow Media Splitter and Decoders

[Armada]

I've been using the beta of EMET4 without problems.

edit: I checked the settings and only SEHOP is unticked for the MPC-HC process.
So what is your filter graph like? Can you turn off EMET, play a file and show me the contents of the Play > Filters menu?

[sneaker_ger]

Actually, the latest LAV from 23-09-2013 CCCP experimental (http://www.cccp-project.net/beta/do_not_touch_this/Combined-Community-Codec-Pack-BETA-2013-09-23.exe) seems to fix it. (It also adds colorimetry signaling for e.g. EVR, if anyone wants to test.) I suppose any latest git version would have done the trick? Or could this be compiling related?

[Owyn]

Just a newb question:
Should I enable LAV Splitter in MPC HC? I enabled LAV video\audio decoders for 10bit video to work, but why do I need LAV splitter there? I tried with it and without it and didn't notice anything different playing an .mkv video file + external audio .mka one.

[gilic]

Actually, the latest LAV from 23-09-2013 CCCP experimental (http://www.cccp-project.net/beta/do_not_touch_this/Combined-Community-Codec-Pack-BETA-2013-09-23.exe) seems to fix it. (It also adds colorimetry signaling for e.g. EVR, if anyone wants to test.) I suppose any latest git version would have done the trick?

Well I was already using the latest MPC-HC nightly build.
@Armada Are you still interested in my filter graph? Just now pondering how to actually turn off EMET again...

[Mangix]

I'm pretty sure that EMET is not intended for specific use-case applications like MPC-HC. General use-case things like browsers are more valuable. Although Google Chrome has problems with EMET'S EAF mitigation.

[sneaker_ger]

I don't really want to discuss what EMET is "intended for", but media players are regularly affected by security holes and are regularly used for playing non-self-created contents as well so it can make sense to enable it for them. The only reason Microsoft has not enabled this by default is for compatibility reasons. This is my last post on the usefulness of EMET for players, not least because it is off-topic in this thread.

[sneaker_ger]

Well I was already using the latest MPC-HC nightly build.

The internal version of LAV included in the MPC-HC 7851 nightly crashes for me as well. Are these based on the latest git or on the latest stable release? Does anyone have a link to a build of the latest git if they aren't based on those?

/edit:
Ok, I found roytam's builds (http://roy.orz.hm/lavf-w32-nightlies/) and they crash as well. I asked the CCCP guys what could be the cause and where we can find the source.

[nevcairiel]

Not absolutely the latest, but close (build last night for testing)
http://files.1f0.de/lavf/LAVFilters-0.58.2-73-g4eb76e7.zip

The only difference might be that CCCP builds with VS2010 and MPC-HC builds with VS2012, on the other hand, 0.58.2 release was also build with VS2010

[sneaker_ger]

That one also crashes.

[sneaker_ger]

Nothing special on their side it seems:
I think TheRyuu had EMET installed and MPC-HC started crashing for him, but that got (more or less) fixed on MPC-HC's side. The new hooking library and how it was used in MPC-HC made it herp a derp. We've not had any LAV related EMET crashes here at the CCCP so far.

The only other change is what came with the 2013-09-17 beta that makes LAV clean up after itself a bit better (tray icons etc.), otherwise the "external as internal" stuff would crash if LAV's tray icon was enabled because MPC-HC was so quick to kill/unload the library. As far as compilation goes, LAV is still built with VS2010. MPC-HC actually uses VS2012 U3 now as its build.bat is used.

The source code, as always, is available on my github repository (https://github.com/jeeb/lavfilters/releases/tag/2013-09-23-beta).

[sneaker_ger]

The CCCP 2013-09-05 BETA (which seems to use the vanilla git) does not crash as well. Currently testing on GraphStudioNext to rule out MPC-HC. GSN crashes as soon as I add LAV (does not matter which filter). Guess there got to be a difference concerning the compilation after all. :confused:

[JEEB]

The CCCP 2013-09-05 BETA (which seems to use the vanilla git) does not crash as well. Currently testing on GraphStudioNext to rule out MPC-HC. GSN crashes as soon as I add LAV (does not matter which filter). Guess there got to be a difference concerning the compilation after all. :confused:
Uhh, no. In any case we do at the very least have differing defaults and possibly some other changes depending on what gets merged upstream. So not fully vanilla (link (https://github.com/jeeb/lavfilters/commits/2013-09-05-beta)).

Other than those changes, I really don't know what could affect EMET going bonkers on vanilla compared to my compiles, I really don't do anything special other than build ffmpeg with an older GCC 4.7.3-based toolchain, and compile LAV with an up-to-date VS2010.

[sneaker_ger]

Uhh, no. In any case we do at the very least have differing defaults and possibly some other changes depending on what gets merged upstream. So not fully vanilla (link (https://github.com/jeeb/lavfilters/commits/2013-09-05-beta)).
My bad. I saw the oldest CCCP patch in the 9-23 git (https://github.com/jeeb/lavfilters/commits/2013-09-23-beta) being titled with "Sep 14, 2013" and missed some patches actually being way older than that.

[nevcairiel]

Public Service Announcement

I'll be on vacation for the next three and a half weeks, and hopefully not look here at all in that period. :p
There won't be any code changes, and no responses. But before anyone starts wondering if LAV died because of the inactivity, it didn't, and after the vacation I will also have more time again to pick up the remaining topics and push out a release again.

[ThurstonX]

Public Service Announcement

I'll be on vacation for the next three and a half weeks, and hopefully not look here at all in that period. :p
There won't be any code changes, and no responses. But before anyone starts wondering if LAV died because of the inactivity, it didn't, and after the vacation I will also have more time again to pick up the remaining topics and push out a release again.

Have a great time, nev!

Cheers

[jkauff]

Have a very restful and refreshing vacation. Stay away from Minnesota, though. :)

[JarrettH]

Where are you staying? :D

[nevcairiel]

I'll be in Australia, Melbourne and Sydney, and some places in between.

[Carpo]

pulled the latest git, and when trying to build LAVFilters I am getting this issue -

Error 23 error C1083: Cannot open include file: 'libavutil/avconfig.h': No such file or directory c:\lavfilters\ffmpeg\libavutil\common.h 39 1 Demuxers

now unless I am blind I cannot see that file there, nor is it there in MPC-HC, is this file from another place? WindowsSDK??

I believe I have everything installed that I need, as I had the same things in place before I re-installed.

[andyvt]

pulled the latest git, and when trying to build LAVFilters I am getting this issue -

Error 23 error C1083: Cannot open include file: 'libavutil/avconfig.h': No such file or directory c:\lavfilters\ffmpeg\libavutil\common.h 39 1 Demuxers

now unless I am blind I cannot see that file there, nor is it there in MPC-HC, is this file from another place? WindowsSDK??

I believe I have everything installed that I need, as I had the same things in place before I re-installed.

IIRC, these files are generated during the ffmpeg build process. Make sure you execute build.bat (and setup the GCC environment) before trying to build LAV in VS.

[Carpo]

GCC is fine as MPC-HC and laf built fine, running build.bat now and the file is there, learn something new everyday :D

[andyvt]

What is strange is that the files build okay when I built MPC-HC........:angry:

I will give gcc another go, I'm sure I set it the same as before, will triple check

Depending on how your PC is setup you may want to create a custom build.bat.

Mine has


@set PATH=%MINGW32%\bin;%MINGW64%\bin;%MSYS%\bin;%PATH%


right after


call "%VS100COMNTOOLS%vsvars32.bat"

[Carpo]

thanks, will give it a try in a bit, coffee is much needed first :D

[clsid]

Uhh, no. In any case we do at the very least have differing defaults and possibly some other changes depending on what gets merged upstream. So not fully vanilla (link (https://github.com/jeeb/lavfilters/commits/2013-09-05-beta)).

Other than those changes, I really don't know what could affect EMET going bonkers on vanilla compared to my compiles, I really don't do anything special other than build ffmpeg with an older GCC 4.7.3-based toolchain, and compile LAV with an up-to-date VS2010.I can confirm crash with official 0.58.2 build. The exploit that EMET detects is SimExecFlow.

Compiling latest git with GCC 4.7.3 gives a working build.

GraphStudioNext works fine with any build, so the problem might be related to the mhook change in MPC-HC.

[dlmh]

Is it possible to somehow change the color and/or transparency of the subtitles? I'm asking because my plasma screen is showing a bad case of image burn-in cause by subtitles. Maybe making them a little bit less opaque and white will lessen the effect.

[detmek]

LAV Filters does not show subtitles. It is player's feature. So, change that in your player.

[dlmh]

LAV Filters does not show subtitles. It is player's feature. So, change that in your player.

Thanks. I'm using XBMC and remembered changing the settings for subtitles didn't change anything when used with DSPlayer and LAV Splitter/video. Thought this had to do with the internal renderer being bypassed.

[Asmodian]

Maybe you have xy-VSFilter installed and is being used by XBMC to render subtitles?

[mindbomb]

heads up, the new intel atom has full bitstream hardware decoding of vp8.

[shpitz]

Probably, a sample would still help to support it eventually.

Sorry about the late response, but here's a sample that doesn't rotate:
VID_20130928_012143.mp4 (http://www39.zippyshare.com/v/28485654/file.html)

Here's the MediaInfo detail about it:
Video
ID : 1
Format : AVC
Format/Info : Advanced Video Codec
Format profile : Baseline@L4.1
Format settings, CABAC : No
Format settings, ReFrames : 1 frame
Format settings, GOP : M=1, N=29
Codec ID : avc1
Codec ID/Info : Advanced Video Coding
Duration : 16s 625ms
Source duration : 16s 640ms
Bit rate : 15.1 Mbps
Width : 1 280 pixels
Height : 720 pixels
Display aspect ratio : 16:9
Rotation : 90°
Frame rate mode : Variable
Frame rate : 30.000 fps
Minimum frame rate : 20.979 fps
Maximum frame rate : 30.040 fps
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Bits/(Pixel*Frame) : 0.547
Stream size : 30.0 MiB (97%)
Source stream size : 30.0 MiB (97%)
Title : VideoHandle
Language : English
Encoded date : UTC 2013-09-28 06:22:01
Tagged date : UTC 2013-09-28 06:22:01
mdhd_Duration : 16625

Note the entry for 'Rotation' of '90°'. A file that doesn't need rotation does not have the 'Rotation' flag at all.

I'm on latest LAV (v0.85.2) using MPC-BE with LAV splitter/audio/video.

[wanezhiling]

Confirmed.
Mpc-be and PotPlayer support embedded rotation info

[octal9]

Thanks. I'm using XBMC and remembered changing the settings for subtitles didn't change anything when used with DSPlayer and LAV Splitter/video. Thought this had to do with the internal renderer being bypassed.
try using xysubfilter with madvr renderer - can set up sub editing on the fly via xysubfilter and set up global default subs to yr. liking (or just edit sub transparency of existing subs), including transparency, font, size, color, etc...

[dandy]

hi
first of all - thank you for lav

I have an old machine with a pinacle pctvsat card
for the card I have to build the ds filter chain manually
simplefied
- source filters (card specific and transform)
- mpeg2 demultiplixer (ms)
- audio and video decoder
- renderers

for the mpeg2 demultiplexer I create output pins for sd and hd:
H264VideoFormat : array[0..87] of Byte=(
$00, $00, $00, $00, // .hdr.rcSource.left = 0x00000000
$00, $00, $00, $00, // .hdr.rcSource.top = 0x00000000
$00, $00, $00, $00, // .hdr.rcSource.right = 0x00000000
$00, $00, $00, $00, // .hdr.rcSource.bottom = 0x00000000
$00, $00, $00, $00, // .hdr.rcTarget.left = 0x00000000
$00, $00, $00, $00, // .hdr.rcTarget.top = 0x00000000
$00, $00, $00, $00, // .hdr.rcTarget.right = 0x00000000
$00, $00, $00, $00, // .hdr.rcTarget.bottom = 0x00000000
$00, $00, $00, $00, // .hdr.dwBitRate = 0x00000000
$00, $00, $00, $00, // .hdr.dwBitErrorRate = 0x00000000
$00, $00, $00, $00, $00, $00, $00, $00, // .hdr.AvgTimePerFrame = 0x0000000000000000
$1C, $00, $00, $00, // .hdr.bmiHeader.biSize = 0x00000028
$80, $07, $00, $00, // .hdr.bmiHeader.biWidth = 1920
$38, $04, $00, $00, // .hdr.bmiHeader.biHeight = 1080
$00, $00, // .hdr.bmiHeader.biPlanes = 0x0000
$18, $00, // .hdr.bmiHeader.biBitCount = 0x0000
$41, $56, $43, $31, // .hdr.bmiHeader.biCompression = "AVC1"
$00, $00, $00, $00, // .hdr.bmiHeader.biSizeImage = 0x00000000
$00, $00, $00, $00, // .hdr.bmiHeader.biXPelsPerMeter = 0x00000000
$00, $00, $00, $00, // .hdr.bmiHeader.biYPelsPerMeter = 0x00000000
$00, $00, $00, $00, // .hdr.bmiHeader.biClrUsed = 0x00000000
$00, $00, $00, $00 // .hdr.bmiHeader.biClrImportant = 0x00000000
);
Mpeg2ProgramVideo : array[0..217] of Byte=(
$00, $00, $00, $00, //00 .hdr.rcSource.left = $00000000
$00, $00, $00, $00, //04 .hdr.rcSource.top = $00000000
$D0, $02, $00, $00, //08 .hdr.rcSource.right = $000002d0 //720
$40, $02, $00, $00, //0c .hdr.rcSource.bottom = $00000240 //576
$00, $00, $00, $00, //10 .hdr.rcTarget.left = $00000000
$00, $00, $00, $00, //14 .hdr.rcTarget.top = $00000000
$D0, $02, $00, $00, //18 .hdr.rcTarget.right = $000002d0 //720
$40, $02, $00, $00, //1c .hdr.rcTarget.bottom = $00000240// 576
$00, $09, $3D, $00, //20 .hdr.dwBitRate = $003d0900
$00, $00, $00, $00, //24 .hdr.dwBitErrorRate = $00000000
$80, $1A, $06, $00, $00, $00, $00, $00, //28 .hdr.AvgTimePerFrame = $0000000000051763 ->1000000/ 40000 = 25fps
$00, $00, $00, $00, //2c .hdr.dwInterlaceFlags = $00000000
$00, $00, $00, $00, //30 .hdr.dwCopyProtectFlags = $00000000
$04, $00, $00, $00, //34 .hdr.dwPictAspectRatioX = $00000004
$03, $00, $00, $00, //38 .hdr.dwPictAspectRatioY = $00000003
$00, $00, $00, $00, //3c .hdr.dwReserved1 = $00000000
$00, $00, $00, $00, //40 .hdr.dwReserved2 = $00000000
$28, $00, $00, $00, //44 .hdr.bmiHeader.biSize = $00000028
$D0, $02, $00, $00, //48 .hdr.bmiHeader.biWidth = $000002d0 //720
$40, $02, $00, $00, //4c .hdr.bmiHeader.biHeight = $00000240 //576
$00, $00, //50 .hdr.bmiHeader.biPlanes = $0000
$00, $00, //54 .hdr.bmiHeader.biBitCount = $0000
$00, $00, $00, $00, //58 .hdr.bmiHeader.biCompression = $00000000
$00, $00, $00, $00, //5c .hdr.bmiHeader.biSizeImage = $00000000
$D0, $07, $00, $00, //60 .hdr.bmiHeader.biXPelsPerMeter = $000007d0
$27, $CF, $00, $00, //64 .hdr.bmiHeader.biYPelsPerMeter = $0000cf27
$00, $00, $00, $00, //68 .hdr.bmiHeader.biClrUsed = $00000000
$00, $00, $00, $00, //6c .hdr.bmiHeader.biClrImportant = $00000000
$98, $F4, $06, $00, //70 .dwStartTimeCode = $0006f498
$00, $00, $00, $00, //74 .cbSequenceHeader = $00000000
$02, $00, $00, $00, //78 .dwProfile = $00000002
$02, $00, $00, $00, //7c .dwLevel = $00000002
$00, $00, $00, $00, //80 .Flags = $00000000
$00, $00, $01, $B3, $2D, $01, $E0, $24,
$09, $C4, $23, $81, $10, $11, $11, $12,
$12, $12, $13, $13, $13, $13, $14, $14,
$14, $14, $14, $15, $15, $15, $15, $15,
$15, $16, $16, $16, $16, $16, $16, $16,
$17, $17, $17, $17, $17, $17, $17, $17,
$18, $18, $18, $19, $18, $18, $18, $19,
$1A, $1A, $1A, $1A, $19, $1B, $1B, $1B,
$1B, $1B, $1C, $1C, $1C, $1C, $1E, $1E,
$1E, $1F, $1F, $21, $00, $00, $01, $B5,
$14, $82, $00, $01, $00, $00 );

with ffdshow (old version - rev2547_20081228 / newer don't work) for SD and with CoreAVC for HD all is working well

now: lav is connecting to the demultiplexer (sd or hd output pin) allright - but there is no picture

I don't know if this is of any interest for other people, but perhaps it's easy to fix
and I do want to use lav for viewing tv!

thank you

[clsid]

So rev2583 (http://sourceforge.net/projects/ffdshow-tryout/files/SVN%20builds%20by%20clsid/very%20old%20builds/ffdshow_rev2583_20090105_clsid.exe/download) is broken? Have you tried both libavcodec and libmpeg2?

[dandy]

@clsid
thank you for your reply
I've tested some of your old builds and can now say (live mpeg2 ts from BBC One)
- rev3572_20100913 works ok with both libavcodec and libmpeg2 (libmpeg2 seems smoother)
- rev3611_20101006 doesn't work (both libavcodec and libmpeg2 / ffdshow doesn't connect to the Microsoft mpeg2-demultiplexer)
nothing else is changed during the tests
latest build - same as 3611
I don't remember why I kept rev2547

with lav the graph builds ok but no picture

[NikosD]

Using latest official LAV 0.58.2 and an old laptop with Intel iGPU 945GM (mobile 945 express chipset), I found out that LAV doesn't support it in DXVA mode.

MPEG-2 is the only HW accelerated codec exposed by that chipset (MPEG2_VLD) and it's working fine with PotPlayer in DXVA mode.

LAV says it's available in DXVA mode but never goes "in use".
LAV decodes MPEG-2 in software (fall back) if you select DXVA.

I tried all of my MPEG-2 clips (progressive and interlaced)

[wanezhiling]

MPEG-2 is the only HW accelerated codec exposed by that chipset (MPEG2_VLD)

http://download.intel.com/products/graphics/intel_graphics_guide.pdf
http://i.imgur.com/LPdr0no.png
GMA 950 only supports MPEG2_MC...


Could you post a screenshot of dxva checker?

[NikosD]

13752

Maybe your post is more accurate than mine.

I can't blame DXVA Checker, probably an Intel's driver error.
I'm using driver version 8.15.10.1930 (latest).

Probably the driver tricks LAV Video decoder too, because LAV says DXVA available in properties.

I have to say though, that during playback PotPlayer info says DXVA_VLD not some other partial acceleration.

CPU utilization is more than usual for VLD decoding, but because it's a rather slow platform, I can't say for sure if it's a partial or full acceleration during PotPlayer's playback.

For now, I'm staying with the impression of tricky drivers, exposing something that the chipset can't support.

[wanezhiling]

Driver error i think. should only support MC not VLD.

[oddball]

I cannot play this clip.

http://pdl.warnerbros.com/wbmovies/gravity/trailer2/GRAVITY_TRAILER_5-2k.mov

It comes out red and black in MPC-HC using LAV.

EDIT: Nevermind it's MadVR that's causing it.

[sneaker_ger]

JFYI that clip is ProRes (10 Bit, 4:2:2) and I have no problem with LAV + madVR playing it with P210 output.

/edit:
http://forum.doom9.org/showthread.php?p=1646239#post1646239 (v210 broken, not sure if LAV or madVR fault)

[Aleksoid1978]

Driver error i think. should only support MC not VLD.

Intel 950 support MPEG2_VLD DXVA2 mode.
I work with it's chipset some times ago, and MPC-BE must support this.

[sdancer75]

Hi,

I added a new simple Interface (IFileOffset.h) like below.

static const IID IID_IFileOffset = { 0xF9A5586A, 0xEFC5, 0x4b46, { 0x99, 0x18, 0x03, 0xA7, 0x8B, 0x71, 0x5B, 0x81 } };


[uuid("F9A5586A-EFC5-4b46-9918-03A78B715B81")]
interface IFileOffset : public IUnknown {
virtual STDMETHODIMP SetFileOffset(ULONGLONG offset)= 0;
virtual STDMETHODIMP GetFileOffset(ULONGLONG *offset)= 0;
};

I changed the CLAVSplitterSource, declaring and implementing the above two virtual methods.


[uuid("B98D13E7-55DB-4385-A33D-09FD1BA26338")]
class CLAVSplitterSource : public CLAVSplitter , public IFileOffset
{
public:
// construct only via class factory
CLAVSplitterSource(LPUNKNOWN pUnk, HRESULT* phr);
virtual ~CLAVSplitterSource();


// IUnknown
DECLARE_IUNKNOWN;
STDMETHODIMP NonDelegatingQueryInterface(REFIID riid, void** ppv);


//IFileOffset
STDMETHODIMP SetFileOffset(ULONGLONG offset);
STDMETHODIMP GetFileOffset(ULONGLONG *offset);



};




Testing the filter with the code below, it fails quering IID_IFileOffset with E_NOINTERFACE error.

hr = CoCreateInstance(CLSID_LAVSplitterSource, NULL, CLSCTX_INPROC_SERVER, IID_IBaseFilter, (void**) &ppF);
if (SUCCEEDED(hr))
{
hr = pGraph->AddFilter(ppF, L"LAV Splitter Source");

hr = ppF->QueryInterface(IID_IFileOffset,(void **)&pFileOffset);
if (FAILED(hr))
{
printf("ERROR - Could not obtain the File Offset interface.");
ppF->Release(); // Clean up after ourselves
ppF = NULL;
return 0;

}
}

What seems to be the problem here ?

[sebas_led]

You have to implement:
[code]
CLAVSplitterSource::QueryInterface(... clsid , void **pObject) {
if (clsid == IID_IFileOffset) {
... //
}
}
[code]

[sdancer75]

You have to implement:
[code]
CLAVSplitterSource::QueryInterface(... clsid , void **pObject) {
if (clsid == IID_IFileOffset) {
... //
}
}
[code]

Yes you have right... I forgot it :-)

[molitar]

I don't know why there is not a good boost on LAV Audio. I have to use both LAV and ffdshow as I can increase low audio with ffdshow by setting it to all unsupported and than use volume boost. It's ridiculous I have to use both but when all the different encoding being used by the different fansub groups it's the only way to have decent audio. LAV Audio needs a feature to boost ALL audio! And too boost it alot I have to run usually at 9db and if you know anything about db that is 100% gain for ever 3db so I am boosting 300%. If I do not do this than all other audio is too loud for my normal pc usage like a simple beep becomes a blare! So effectively increasing ffdshow audio boost let's me keep my pc audio at a comfortable level for everything else.

So basically the only way I have found for a good solution so far is CCCP with ffdshow.

[e-t172]

I don't know why there is not a good boost on LAV Audio. I have to use both LAV and ffdshow as I can increase low audio with ffdshow by setting it to all unsupported and than use volume boost. It's ridiculous I have to use both but when all the different encoding being used by the different fansub groups it's the only way to have decent audio. LAV Audio needs a feature to boost ALL audio!

Having a feature to boost audio (with clipping if need be) would indeed be useful as a stopgap solution for people with low-gain audio systems.

And too boost it alot I have to run usually at 9db and if you know anything about db that is 100% gain for ever 3db so I am boosting 300%.

Not quite. When talking about voltage and digital samples 6dB = x2 (100%). 9dB is x2.8 gain. You arrive at the correct number because you made a mistake in the calculation (if 3dB were 100% gain, then 9dB would be 400%, not 300%, because dB is a logarithmic scale).

If I do not do this than all other audio is too loud for my normal pc usage like a simple beep becomes a blare! So effectively increasing ffdshow audio boost let's me keep my pc audio at a comfortable level for everything else.

That's normal and expected. Most videos have very high dynamic range (average level is typically -20 dBFS) and that's A Very Good Thing©®™ when it comes to audio quality. To fix your issues you should reduce the volume of other applications, not boost and clip your video player's output, as it might significantly affect audio quality.

[Thunderbolt8]

whats the difference of using LAV stuff as internal filters of mpc-hc or using it as external filter? how can I be sure when I want to use the external filters that the internal LAV filter settings wont override or interfere?

[Soukyuu]

One difference I know of is that the internal filters can only be used by the player they're built into. So if you only plan to use LAV for playback in that player, you don't need to install the external one.

[Thunderbolt8]

ok but how do I know if inbuilt LAV is active in mpc-hc or one of their own internal codecs? what are the rules of priority here?