MP 1.3.x dshowhelper development (1 Viewer)

[SciDoctor]

Testing now thanks.

Shame there is such a dislike of ATI drivers; over on the doom lav thread its endemic, I shy from posting over there incase I am burnt at the stake.

 

[SciDoctor]

681 has been as stable as 677 so far, lots of testing, quick format changing etc haven't caused any lockups crashes. SD, HD, TS, DVD, BluRay no problems.

Were the changes you made subtle or substantial compared to the code for your initial version I presume tested on NV hardware?

Are the ATI ways so much different to NV ?

 

[SciDoctor]

Now stability isn't a problem with 681, critical viewing has shown it not to be a smooth as 677 in the visual appearance.

Bringing up the stats shows render time has increased and the grasphs aren't as smooth; obvious with scrolling text and horizontal panning of scene movement.

Is there anything in the changes since 677 that could have caused this. Scren caps less than acouple of minutes apart.

 

[red5goahead]

@SciDoctor

50 fps on 60 Hz refresh rate? (it doesn't help in this case but you don't use MP Audio renderer?)
Are you joking us?

 

[SciDoctor]

@SciDoctor

50 fps on 60 Hz refresh rate? (it doesn't help in this case but you don't use MP Audio renderer?)
Are you joking us?

Not an option on the monitor I use, fixed at 60Hz. I know more recent equipment has a better range and could improve matters slightly but within the context of the test and observations has little impact .

Don't use MPAR, found it to be unstable as it is vey much beta.

 

[doveman]

Not an option on the monitor I use, fixed at 60Hz. I know more recent equipment has a better range and could improve matters slightly but within the context of the test and observations has little impact .

Don't use MPAR, found it to be unstable as it is vey much beta.

Same here. My TV only does 60hz. Works well enough most of the time.

 

[Owlsroost]

Now stability isn't a problem with 681, critical viewing has shown it not to be a smooth as 677 in the visual appearance.

Bringing up the stats shows render time has increased and the grasphs aren't as smooth; obvious with scrolling text and horizontal panning of scene movement.

Is there anything in the changes since 677 that could have caused this. Scren caps less than acouple of minutes apart.

There aren't many functional changes between v677 and v681, the main one is the position in the startup sequence when the DWM queue changes are performed. The actual implementation is different, but I can't see how this would produce the performance differences you are reporting - I think it's more likely to be the timing of the DWM changes affecting things (in some weird way....)

For comparison, attached is v683 with the DWM queue changes performed as early as possible, the same as in v677.

Tony

 

[SciDoctor]

Thanks again for your hard work on this. Testing now for this evening.



That is much better .5ms drop in render time and much smoother.

 

[Owlsroost]

Thanks again for your hard work on this. Testing now for this evening.

View attachment 117969

That is much better .5ms drop in render time and much smoother.

OK - I'll make the early/late initialisation a registry option in the next version (with a default of 'early')

Tony

 

[SciDoctor]

Thanks again for your hard work on this. Testing now for this evening.

View attachment 117969

That is much better .5ms drop in render time and much smoother.

OK - I'll make the early/late initialisation a registry option in the next version (with a default of 'early')

Tony

"DWM queued mode initialisation moved later in the start-of-play sequence, to minimise the temporary black screen effect."

Using the 683 and the black screen is almost non-existant compared to 677, something else must have changed for the possitive in this respect compared to 677 surely ?

683 is now the default choice for my set up after testing although 677 didn't fail/cause problems during four weeks of use.