mrw wrote:
> I should remark that I have the same outcome with R1.3 of jive_alsa. I
> haven't tried with R1.4.
I've now made some tests with R1.4, seeking to replicate your own tests
(#1 and #3) with the stock buffer size and period count. I used a 48kHz
16bit Raw PCM stream provided by the LMS Signal Generator plug-in,
rather than your flac file. Again, wirelessly.
Sadly, I experienced bass drop-out with R1.4, just as with the stock
jive_alsa. I discerned no obvious difference in the pattern of
drop-outs, but I wasn't timing it.
A rudimentary analysis of the logs:
Code:
--------------------
Test Test duration # pcm # pcm xrun # kernel
(minutes) underrun!!!
ralphy-stock-20000-2 41.2 49 73 26
per minute 1.19 1.77 0.63
ralphy-ri14-20000-2 24.25 31 38 7
per minute 1.28 1.57 0.29
mrw-stock-20000-2 38.66 79 89 34
per minute 2.04 2.30 0.88
mrw-ri14-20000-2 30.5 59 95 28
per minute 1.93 3.11 0.92
--------------------
The only patterns that I can see are:
My rev 0005 radio consistently experienced more errors that your UE rev
0007 radio. That may just reflect the difference in stream format. But
there may be something else.
My rev 0005 radio experienced pretty much the same error rates with both
versions of jive_alsa, although there was a noticeable up-tick in the
'xrun' errors with R1.4.
Your UE rev 0007 radio experienced pretty much the same 'ALSA-pcm' error
rates with both versions of jive_alsa, but showed a significant drop off
in the kernel error messages with R1.4. As well as not 'dropping out'.
So, what might one conclude ? Here's my stab at it:
R1.4 does not, in general, resolve the problem. But it may be helping
the UE rev 0007 radio for undetermined reasons. The only reason that I
can think of might be a subtle change in timing of the audio play back
loop between 'stock' and R1.4, possibly in the 1's/10's of
micro-seconds, that is (for unknown reasons) impacting play back.
I think this has been a worthwhile exercise, even though (in my view)
not conclusive as regards the UE rev 0007 radio.
I'll carry on with a longer term test of the increased buffer size,
which I hope (and expect) will support your positive findings from tests
#2 and #4.
------------------------------------------------------------------------
mrw's Profile: http://forums.slimdevices.com/member.php?userid=38299
View this thread: http://forums.slimdevices.com/showthread.php?t=104141
_______________________________________________
Radio mailing list
[email protected]
http://lists.slimdevices.com/mailman/listinfo/radio
