All of lore.kernel.org
 help / color / mirror / Atom feed
From: Carlos Rafael Giani <dv@pseudoterminal.org>
To: Joshua Kurland <joshua.kurland@adtecdigital.net>
Cc: "meta-freescale@yoctoproject.org" <meta-freescale@yoctoproject.org>
Subject: Re: Regarding GStreamer-1.x status on iMX6Q
Date: Tue, 10 Dec 2013 20:01:42 +0100	[thread overview]
Message-ID: <52A76516.2060705@pseudoterminal.org> (raw)
In-Reply-To: <CAPuotkeqzQd+VtKRTpr0TUem7=-E=rNC2ZF+bZ+W1-NWu-2TmA@mail.gmail.com>

On 2013-12-10 19:52, Joshua Kurland wrote:
> I think the problem I am having is because of the X11 requirement. 
>  Without it, running a simple pipeline using eglvivsink immediately 
> crashes.  Is it possible to run without X?  I don't currently use it 
> on my system.  I will have to try again with X running on a new build 
> later today, but until I get that working would you mind if I ask you 
> a few more questions?

Not yet. It is on my list. For the time being, you can try out 
imxipusink instead. Please note that this sink has one known problem 
(due to an IPU limitation) and does not work well together with X11 
(both render to the framebuffer). But for testing and debugging 
purposes, it should work fine.

> 1.)  The project I have been working on made use of the multi overlay 
> support from mfw_isink.  Specifically it was used for overlaying 
> static jpeg images on top of a playing video.  Will this or something 
> like it be possible using eglvivsink?

How exactly did you use this? I am not that familiar with this feature. 
mfw_isink is similar to imxipusink though, so I suppose it could work.

> 2.)  When I was using mfw_isink, I noticed that their was a five 
> second buffer time between two consecutive pipelines.  What kind of 
> delays can be expected from eglvivsink?  On my x86_64 development 
> desktop with gstreamer-1.0.8 I am able to use features in playbin to 
> seamlessly play files back to back by changing the source URI.  Would 
> this be possible using the new framework?

Five seconds? Hm, thats not what I saw. I didn't run such tests yet, but 
I suspect the buffer time should be shorter. Seamless playback should 
also work. Did this only happen with mfw_isink ? Note that playbin 
itself also has several buffers and queues, which have nothing to do 
with the sink you use.

Also, be careful about interlaced video. While the IPU can deinterlace - 
and it is strongly recommended to do so instead of deinterlacing in 
software, at least for real-time playback - playbin currently has the 
software deinterlacer hardcoded. I talked with GStreamer developers 
about this, and they are thinking about a way to plug in custom 
deinterlacers, color space converters etc. For now, I have an idea I 
didnt have the chance to test yet if you use eglvivsink: specify 
"imxipuvideotransform ! eglvivsink" . If you use the imxipusink instead, 
you don't have to do any of this, because the deinterlacing is included.

cheers


  reply	other threads:[~2013-12-10 19:00 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-11-18 15:48 Regarding GStreamer-1.x status on iMX6Q Joshua Kurland
2013-11-18 16:07 ` Otavio Salvador
2013-11-18 16:54   ` Carlos Rafael Giani
2013-11-18 18:20     ` Joshua Kurland
2013-11-20 13:24       ` Carlos Rafael Giani
2013-11-20 15:25         ` Joshua Kurland
2013-11-20 18:52           ` Carlos Rafael Giani
2013-11-21  1:59           ` Philip Craig
2013-12-02  1:36       ` Carlos Rafael Giani
2013-12-09 20:05         ` Joshua Kurland
2013-12-09 20:31           ` Carlos Rafael Giani
2013-12-09 21:14             ` Joshua Kurland
2013-12-09 21:17               ` Carlos Rafael Giani
2013-12-09 21:26                 ` Joshua Kurland
2013-12-09 21:31                   ` Carlos Rafael Giani
2013-12-10 18:52                     ` Joshua Kurland
2013-12-10 19:01                       ` Carlos Rafael Giani [this message]
2013-12-10 19:15                         ` Joshua Kurland

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=52A76516.2060705@pseudoterminal.org \
    --to=dv@pseudoterminal.org \
    --cc=joshua.kurland@adtecdigital.net \
    --cc=meta-freescale@yoctoproject.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.