Hi Michael! On Fri, 24 Oct 2008 19:25:13 +0200, Michael 'Mickey' Lauer wrote: > thanks for looking into this issue. I'm glad someone actually does > benchmarks, so we can reference your work when this issue comes up > again.
I'd really like to solve not only the Debian bug, but the upstream one, too [1], that's why I spent so much time on it ;-) > If we are talking about "just" adding ogg support, The short term fix is that yes, at least to allow frameworkd in Debian main because of the Arkanoid.sid license not being DFSG-free. In the long term, however, I see the decoder matrix as suboptimal, because you need to precisely specify each codec supported, while with decodebin (or any other similar solution) the code is really general. But if I read correctly your last sentence (I left it at the bottom of this mail) we already agree on this :-) > + ogg: oggdemux ! audioconvert > (is that correct?) My fault, I forgot that ogg needs also a "real" decoder [2]. The Git patch against upstream frameworkd is attached. Once applied, the overall situation is now the following (video at [3], raw data at [4]): |--------------------+------+-------+------| | | vib | ring | lat | |--------------------+------+-------+------| | a) on FSO | | | | | 1. idle | 3.25 | 5.25 | 2.00 | | 2. active | 2.75 | 5.50 | 2.75 | |--------------------+------+-------+------| | b) on Debian | | | | | I- upstream | | | | | 3. idle | 3.50 | 6.00 | 2.50 | | 4. active | 3.00 | 6.00 | 3.00 | | II- decodebin | | | | | 5. Arkanoid.sid | 2.50 | 6.50 | 4.00 | | 6. nothing2.ogg | 2.00 | 7.25 | 5.25 | | III- oggdemux | | | | | 7. nothing2.ogg | 2.75 | 7.25 | 4.50 | |--------------------+------+-------+------| As you can see, it seems there's an improvement WRT to decodebin, but if someone would like to confirmed my analyses, please grab all the video files [2] and then post the results. As soon as my patch (or something similar but with the same effect) is applied upstream, I'll backport it to the Debian package. > Yes, that's just a quick bandaid which doesn't solve it in general, but > I think it's the way to go until someone profiled gstreamer's decodebin > and checked where it burns all these cycles. Please tell me if there's something I can do here, like reporting the bug to the GStreamer people or more tests. The best option would be to have a stand-alone test case, but I'm not sure this will be representative since it's completely detached from the frameworkd infrastructure (and thus the "payload" can be different). Thx, bye, Gismo / Luca Footnotes: [1] http://trac.freesmartphone.org/ticket/123 [2] http://trac.freesmartphone.org/ticket/123#comment:2 [3] http://pkg-fso.alioth.debian.org/ringtone-decodebin/ [4] first round: vib 3.0s, ring 7.0s, lat 4.0s second round: vib 2.5s, ring 7.5s, lat 5.0s
From 20a5c72009431fdc245eb940ea88e364cf9614a1 Mon Sep 17 00:00:00 2001 From: Luca Capello <[EMAIL PROTECTED]> Date: Sun, 26 Oct 2008 11:29:24 +0100 Subject: [PATCH] framework/subsystems/odeviced/audio.py: add Ogg decoder This in part solves ticket #123 [1] (which is vital for Debian [2]) or, better, in Michael's words: "that's just a quick bandaid which doesn't solve it in general, but I think it's the way to go until someone profiled gstreamer's decodebin and checked where it burns all these cycles" [3]. Footnotes: [1] http://trac.freesmartphone.org/ticket/123 [2] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=495668 [3] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=495668#63 --- framework/subsystems/odeviced/audio.py | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/framework/subsystems/odeviced/audio.py b/framework/subsystems/odeviced/audio.py index 3e03931..321e6e2 100644 --- a/framework/subsystems/odeviced/audio.py +++ b/framework/subsystems/odeviced/audio.py @@ -103,6 +103,7 @@ class GStreamerPlayer( Player ): "mod": "modplug", "mp3": "mad", "wav": "wavparse", + "ogg": "oggdemux ! vorbisdec ! audioconvert" } def __init__( self, *args, **kwargs ): -- 1.5.6.5
pgppTR0PpgvL1.pgp
Description: PGP signature
_______________________________________________ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
