- September 14, 2006
- 321
- 23
- Home Country
- United Kingdom
Whilst I now have my refresh rates working perfectly, still working on getting sync for 23.876 stuff. Been trying reclock with varying success. Sometimes its fine for entire movie; sometimes goes badly out of sync, only fixed by stop and start playback again.
Still have just cleaned up my filters and drivers and reinstalled; trying out Catalyst 8.12.
Reading on the Slysoft forum, seems to be that filters used have a big impact on success of Reclock. So looks like I need to try out powerDVD since that is seen to be most stable.
Still have just cleaned up my filters and drivers and reinstalled; trying out Catalyst 8.12.
Reading on the Slysoft forum, seems to be that filters used have a big impact on success of Reclock. So looks like I need to try out powerDVD since that is seen to be most stable.
What is odd and IMHO worth investigating is that when using PowerDVD Reclock is very very stable. The same DVD played in PowerDVD, as shown in an earlier log, barely varies +/- 3ms after it has settled down in the first couple of minutes. The same seems true of Blu-rays in PDVD and mkvs played in MPC-HC using Haali and the Cyberlink h.264 decoder.
It seems Reclock's accuracy in timing AV sync is very dependent on the precise filters. With sufficient buffer surely this should not be the case. Anyway something seems very wrong in the way Reclock is using its buffer. The logic looks flawed. A bigger buffer should enable it to better compensate for any variances, not make things worse, as at present. At the moment a bigger buffer is fine where there is no problem, such as with PDVD, but not where there is. This must be back to front, no????
Quite apart from making Reclock work with all filters instead of only some, fixing this issue may also benefit even "well behaved" filters and players. Even with PDVD there are a couple of moments in a movie when presumably the PC is doing something and the audio sync slips out to say 70ms, taking 10+ seconds to recover. It is possible, given the buffer does not seem to work right now that if we get to the root of the problem even this may be improved.