Rockbox mail archive

• Next message: Benoit Lubek: "RE: Yes/No Prompt Quirk"
• Previous message: Nielsen Linus (ext): "RE: Faster skip forward"
• Maybe in reply to: Mike Soultanian: "Faster skip forward"
• Next in thread: Nielsen Linus (ext): "RE: Faster skip forward"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]

I have some ideas on this...
 
> 1) Having to load 5 secs of the next song just in case the
> user wants to skip it wastes buffer space.

Yes, but if we extend the original idea a bit this could work: after the
first 3 seconds of playback (while the disk is still spinning), and if the
user hasn't pressed next-track, then re-use that 5 seconds of buffer space
for the current song instead.

> 2) The mp3 buffer algorithm would be more complex, as it
> would have to skip the prefetched data if the user doesn't press next.

Not if we implement it as above - it would be buisness as usual in this
case.

> 3) The extra directory lookup, seeking, ID3 parsing and bit
> swapping would probably take extra time

Yes, that is one drawback. But the extra time would occur in an already
time consuming process anyway - so it probably won't be noticable to the
user.

> 4) The track change detection algorithm would be more complex as well.

Yes, this is another drawback. Any takers?

I don't mind waiting for the noraml load time if I choose to miss the 3
second next-track window, it is just annoying at the moment having to wait
when you instantly press next-track. I could argue that it wastes more
battery loading the entire file - only to have to skip to the next one
imediately afterwards.

Stu.

• Next message: Benoit Lubek: "RE: Yes/No Prompt Quirk"
• Previous message: Nielsen Linus (ext): "RE: Faster skip forward"
• Maybe in reply to: Mike Soultanian: "Faster skip forward"
• Next in thread: Nielsen Linus (ext): "RE: Faster skip forward"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]