[mythtv-users] Live TV: Too fast now ?

Jean-Yves Avenard jyavenard at gmail.com
Mon Aug 18 02:30:21 UTC 2014


Hi

On 18 August 2014 01:11, Roger Siddons <dizygotheca at ntlworld.com> wrote:

> However, overall the WAF factor is worse because;
>
> 1. When starting Live TV it fails 60% of the time - "Could not open decoder"

this can't be due to the new live TV code. The new liveTV code is in
addition to the previous method.

That is:

it attempts to quickly detect what the stream is like. If it can't
find it quickly, it reverts to the old method, one that has been used
for years.

As such, it doesn't only try for 1.5s. It tries for 1.5s to see if a
very quick way to detect what you are playing can be found. And after
that start working on the file itself....

It will attempt to read the file up to 15s...

The one thing I can think of that would cause the issue, isn't related
to the liveTV changes found in 0.27.1, but in how the mythplayer
attempts to open the file

In mythplayer.cpp , int MythPlayer::OpenFile(uint retries)
line 938, change peekTimer.elapsed() > 1500 into:
peekTimer.elapsed() > 2500

and see how you go.

we have already up that value several times over the past few months
to cater for slow adapter...
looks like another one is needed.


>
> So, questions:
>
> 1. Are there any other settings that would have an impact ?

see above

>
> 2. Live TV always appears to check RAM for 1500 ms before checking the file.
> Are there any settings/ways to increase this ?

there's no need to increase that one for the reason describe above

>
> 3. Is it feasible/desirable for Live TV to re-examine the file periodically
> before giving up ? If the file exists/is growing then isn't there a chance
> that ffmpeg might eventually succeed. A long pause is preferable to a
> failure. And if you're trying to watch Live TV with a duff tuner then a
> delay of 10s before raising an error doesn't seem unreasonable to me.

it does so already.. for 15s


More information about the mythtv-users mailing list