[mythtv] Understanding the myth source code

[Stuart Morgan]

On Tuesday 25 May 2010 21:38:54 James Courtier-Dutton wrote:
> Hi,
> 
> I am trying to understand the myth source code.
> I have a question.
> When recording a program, why is the mythbackend trying to decode
> audio and video streams?
> Why does it need to decode them?

In addition to what Raymond has already said, we also inspect the frame 
headers to determine information such as resolution, or the audio headers to 
establish the number of audio channels. This is used as metadata to describe 
the recording in the user interface. As Raymond said it's not full decoding, 
in fact the video and audio are not decoded at all, just the mpeg headers and 
only as much of those as we need to.

Then there is preview generation, this happens in a separate spawned backend 
instance but it writes to the main backend log so it might appear that the 
backend is decoding video frames but in reality it's another process entirely 
which seeks to the desired frame and decodes only that frame, saves the image 
data then exists.

> Is this why myth takes a long time to switch channels?

No.

-- 
Stuart Morgan