gtk-play issues

gtk-play issues

[Gary Thomas]

I'm having issues with the new media player (gtk-play and gst-play)
on my i.MX6 based systems.  In order to try and track the problems,
I'm trying to build these programs outside of Yocto (so I can hopefully
see *how* they are meant to work).  Once I can build & test on something
other than my i.MX6, I hope to be able to fix the behaviour there.

I'm trying to build on the latest Ubuntu (15.04/x86_64), but have
run into a block in ./configure:
   No package 'gstreamer-video-1.0' found.
   No package 'gstreamer-pbutils-1.0' found.

Does anyone know what package(s) I need to install to be able
to complete the build?

Thanks for any pointers

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------

Re: gtk-play issues

[Burton, Ross]

[-- Attachment #1: Type: text/plain, Size: 764 bytes --]

On 14 May 2015 at 15:14, Gary Thomas <gary@mlbassoc.com> wrote:

> I'm trying to build on the latest Ubuntu (15.04/x86_64), but have
> run into a block in ./configure:
>   No package 'gstreamer-video-1.0' found.
>   No package 'gstreamer-pbutils-1.0' found.
>
> Does anyone know what package(s) I need to install to be able
> to complete the build?
>

gst-play is a glorified wrapper around the gstreamer PlayBin, so gst-launch
will let you replicate its behaviour.

But for you Ubuntu problem, search the package database for eg
gstreamer-video-1.0.pc:

http://packages.ubuntu.com/search?searchon=contents&keywords=gstreamer-video-1.0.pc&mode=exactfilename&suite=trusty&arch=any

libgstreamer-plugins-base1.0-dev is what you're after.

Ross

[-- Attachment #2: Type: text/html, Size: 1620 bytes --]

Re: gtk-play issues

[Gary Thomas]

On 2015-05-14 08:22, Burton, Ross wrote:
>
> On 14 May 2015 at 15:14, Gary Thomas <gary@mlbassoc.com <mailto:gary@mlbassoc.com>> wrote:
>
>     I'm trying to build on the latest Ubuntu (15.04/x86_64), but have
>     run into a block in ./configure:
>        No package 'gstreamer-video-1.0' found.
>        No package 'gstreamer-pbutils-1.0' found.
>
>     Does anyone know what package(s) I need to install to be able
>     to complete the build?
>
>
> gst-play is a glorified wrapper around the gstreamer PlayBin, so gst-launch will let you replicate its behaviour.

As I figured.  Is there a way to get the pipeline out of gtk-play so I
know how it's launching the bits?

>
> But for you Ubuntu problem, search the package database for eg gstreamer-video-1.0.pc:
>
> http://packages.ubuntu.com/search?searchon=contents&keywords=gstreamer-video-1.0.pc&mode=exactfilename&suite=trusty&arch=any
>
> libgstreamer-plugins-base1.0-dev is what you're after.

That helped, thanks.

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------

Re: gtk-play issues

[Burton, Ross]

[-- Attachment #1: Type: text/plain, Size: 514 bytes --]

On 14 May 2015 at 15:29, Gary Thomas <gary@mlbassoc.com> wrote:

> As I figured.  Is there a way to get the pipeline out of gtk-play so I
> know how it's launching the bits?
>

Read the source, but it really is just the playbin:

https://github.com/sdroege/gst-player/blob/master/lib/gst/player/gstplayer.c#L2028

If you want to know what elements get dynamically created then I believe
turning on debug (see the packageconfig) will let you write out dot files
for each stage of the pipeline.

Ross

[-- Attachment #2: Type: text/html, Size: 1228 bytes --]

Re: gtk-play issues

[Gary Thomas]

On 2015-05-14 08:39, Burton, Ross wrote:
>
> On 14 May 2015 at 15:29, Gary Thomas <gary@mlbassoc.com <mailto:gary@mlbassoc.com>> wrote:
>
>     As I figured.  Is there a way to get the pipeline out of gtk-play so I
>     know how it's launching the bits?
>
>
> Read the source, but it really is just the playbin:
>
> https://github.com/sdroege/gst-player/blob/master/lib/gst/player/gstplayer.c#L2028

Interesting - the version in OE-core is 84 revisions behind the master/HEAD.
Should I bother debugging what's in OE-core or see what's up with the latest
code?

>
> If you want to know what elements get dynamically created then I believe turning on debug (see the packageconfig) will let you write out dot files for each stage of the pipeline.

I don't see any 'debug' PACKAGECONFIG settings, only GSTREAMER_1_DEBUG
variable and that seems to only affect the core.

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------

Re: gtk-play issues

[Burton, Ross]

[-- Attachment #1: Type: text/plain, Size: 675 bytes --]

On 14 May 2015 at 16:27, Gary Thomas <gary@mlbassoc.com> wrote:

> Interesting - the version in OE-core is 84 revisions behind the
> master/HEAD.
> Should I bother debugging what's in OE-core or see what's up with the
> latest
> code?
>
>
Try updating to the latest commit (and send the patch...).


> If you want to know what elements get dynamically created then I believe
>> turning on debug (see the packageconfig) will let you write out dot files
>> for each stage of the pipeline.
>>
>
> I don't see any 'debug' PACKAGECONFIG settings, only GSTREAMER_1_DEBUG
> variable and that seems to only affect the core.
>

Erm yeah, that's the one.

Ross

[-- Attachment #2: Type: text/html, Size: 1458 bytes --]

Re: gtk-play issues

[Gary Thomas]

On 2015-05-14 09:29, Burton, Ross wrote:
>
> On 14 May 2015 at 16:27, Gary Thomas <gary@mlbassoc.com <mailto:gary@mlbassoc.com>> wrote:
>
>     Interesting - the version in OE-core is 84 revisions behind the master/HEAD.
>     Should I bother debugging what's in OE-core or see what's up with the latest
>     code?
>
>
> Try updating to the latest commit (and send the patch...).

Will do.

>
>         If you want to know what elements get dynamically created then I believe turning on debug (see the packageconfig) will let you write out dot files for each stage of the
>         pipeline.
>
>
>     I don't see any 'debug' PACKAGECONFIG settings, only GSTREAMER_1_DEBUG
>     variable and that seems to only affect the core.
>
>
> Erm yeah, that's the one.

OK, thanks.

BTW, neither gtk-play nor gst-play seem to work correctly with
an out-of-the-box qemu that I just built (from master at the
time of fido release - rev 515a8b790e08a48aaf64a6b7237e2b39390e35f6)

I'm not much of an x86 user - how would I build (which $MACHINE, etc)
to run on my generic [Dell] laptop so I can test it there?

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------

Re: gtk-play issues

[Burton, Ross]

[-- Attachment #1: Type: text/plain, Size: 478 bytes --]

On 14 May 2015 at 17:29, Gary Thomas <gary@mlbassoc.com> wrote:

> BTW, neither gtk-play nor gst-play seem to work correctly with
> an out-of-the-box qemu that I just built (from master at the
> time of fido release - rev 515a8b790e08a48aaf64a6b7237e2b39390e35f6)
>
> I'm not much of an x86 user - how would I build (which $MACHINE, etc)
> to run on my generic [Dell] laptop so I can test it there?
>

For a real x86 machine, intel-corei7-64 from meta-intel.

Ross

[-- Attachment #2: Type: text/html, Size: 892 bytes --]

Re: gtk-play issues

[Gary Thomas]

On 2015-05-14 10:31, Burton, Ross wrote:
>
> On 14 May 2015 at 17:29, Gary Thomas <gary@mlbassoc.com <mailto:gary@mlbassoc.com>> wrote:
>
>     BTW, neither gtk-play nor gst-play seem to work correctly with
>     an out-of-the-box qemu that I just built (from master at the
>     time of fido release - rev 515a8b790e08a48aaf64a6b7237e2b39390e35f6)
>
>     I'm not much of an x86 user - how would I build (which $MACHINE, etc)
>     to run on my generic [Dell] laptop so I can test it there?
>
>
> For a real x86 machine, intel-corei7-64 from meta-intel.
>

Interesting - gst-play/gtk-play works correctly on this target.
Now I have to figure out why it doesn't work on the other systems
I tested.

Should I submit bugs against the QEMU targets as they get it
even more wrong than the i.MX6 targets??

Query: my laptop [test device] has two ALSA sound devices,
HDMI and the internal PCH.  It has picked the HDMI as default
and I get no sound when no HDMI monitor attached.  How do I
select the other sound device (from sato desktop would be
best, but I'd settle for any answer)

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------

Re: gtk-play issues

[Burton, Ross]

[-- Attachment #1: Type: text/plain, Size: 1368 bytes --]

On 14 May 2015 at 20:40, Gary Thomas <gary@mlbassoc.com> wrote:

> Interesting - gst-play/gtk-play works correctly on this target.
> Now I have to figure out why it doesn't work on the other systems
> I tested.
>
> Should I submit bugs against the QEMU targets as they get it
> even more wrong than the i.MX6 targets??
>
> Query: my laptop [test device] has two ALSA sound devices,
> HDMI and the internal PCH.  It has picked the HDMI as default
> and I get no sound when no HDMI monitor attached.  How do I
> select the other sound device (from sato desktop would be
> best, but I'd settle for any answer)
>

Ah, the old audio playback problem.  There's a bug for that.

Sato doesn't have a GUI for this right now but the good news is that this
is something you can probably fiddle with alsamixer for now, and in master
we'll be integrating PulseAudio properly so ideally it re-routes the audio
when HDMI is hotplugged.   Obviously at this point you're expecting this
series of events to actually work:
1) GPU detects HDMI hotplug
2) GPU probes HDMI, detects audio channel
3) ALSA gets notified of the new audio channel
4) PulseAudio gets notified of the new channel and can reroute the channels

I think the key word here is YMMV, and a lot of it is up to your GPU driver
as that's where the HDMI interaction obviously happens.

Ross

[-- Attachment #2: Type: text/html, Size: 2059 bytes --]

Re: gtk-play issues

[Nikolay Dimitrov]

Hi Gary,

On 05/14/2015 10:40 PM, Gary Thomas wrote:
> On 2015-05-14 10:31, Burton, Ross wrote:
>>
>> On 14 May 2015 at 17:29, Gary Thomas <gary@mlbassoc.com
>> <mailto:gary@mlbassoc.com>> wrote:
>>
>>     BTW, neither gtk-play nor gst-play seem to work correctly with
>>     an out-of-the-box qemu that I just built (from master at the
>>     time of fido release - rev 515a8b790e08a48aaf64a6b7237e2b39390e35f6)
>>
>>     I'm not much of an x86 user - how would I build (which $MACHINE, etc)
>>     to run on my generic [Dell] laptop so I can test it there?
>>
>>
>> For a real x86 machine, intel-corei7-64 from meta-intel.
>>
>
> Interesting - gst-play/gtk-play works correctly on this target.
> Now I have to figure out why it doesn't work on the other systems
> I tested.
>
> Should I submit bugs against the QEMU targets as they get it
> even more wrong than the i.MX6 targets??
>
> Query: my laptop [test device] has two ALSA sound devices,
> HDMI and the internal PCH.  It has picked the HDMI as default
> and I get no sound when no HDMI monitor attached.  How do I
> select the other sound device (from sato desktop would be
> best, but I'd settle for any answer)

How about this:

gst-launch playbin ... ! alsasink device="hw:1,0"

Regards,
Nikolay