Moonlight Decoder doesn't remember Performance Settings

[3ngel]

Hi to all,
much compliments for the quality of the decoder. But why the decoder doesnt' remember the Performance Setting?
In other words, i open Graphedit, open Moolight Decoder properties, set up "Decode all frames" and "IDCT Double Precision". Close Graphedit. Reopen Graphedit, open Moolight Decoder properties, but only "IDCT Double Precision" is ON, the performance is remained to "Obey Quality Frame". There is a way to fix this?
Thank so much

[Vovs]

Hi,
The perfomance parameter cannot be written into registry now at all. Therefore you can get the default value only after graph reloading. Are you really want to use this feature?

[3ngel]

Yes, absolutely. Because in this way, after i construct my graph in Graphedit, i can save my graph in .GRF file, close graphedit, and use the resulting .GRF file in a program like avisynth, being sure that all the parameters of the decoder are set right. And this is of ENORMOUS usefulness and process easyness.

[SnoopyDog]

Hi

I'm a registered user of the MPEG Player. The Elecard / Moonlight video codecs are very good piece of software!

With regedit, you can manually create a DWORD value named "Quality" in the MPEG2 Decoder key. A value of 4 (as far as i remember) should be the setting you need. But i can not see any performance benefit with this setting.

[3ngel]

Well, what to say... fantastic! It works!! Thank you so much!
I see a difference in the quality of the output, with the "Obey quality" the frames have more background noise, while with "Decode All", frames are more clean.
So thank you again SnoopyDog!

PS : You know pheraps what is the function meaning of "Do postprocess" and "HQ Upsample"?

[SnoopyDog]

I'm happy that i could help you  

I only know, that "Do postprocess" means: "Removes MPEG compression artifacts using deblocking and deringing filters", but i don't know, what kind of data will be upsampled with "HQ upsample".

By the way: don't you have problems when you enable Double Precision IDCT? When i do this, i got a black picture when play MPEG2 videos with WinAmp. The Elecard mpeg player disables Double Precision IDCT by itself, when i start playing an mpeg2 video clip.

[3ngel]

Mmm, ok i've understood the "Do PostProcess", but regarding "HQ Upsample" there is some guy from the developer team that can explain this? Thanks

Regarding Double Precision IDCT, why you use winamp to play videos?
You can use Media Player Classic, that is the best for playing videos. With this you can set up the decoder filter parameters on the fly.

[peter]

Utter rubbish. There is no difference in video quality for 'decode all frames' and 'obey quality messages'. The quality setting affects decoder performance. When you set 'decode all frames', decoder ignores quality messages from video renderer, and doesn't skip frames if they are to be presented late. I.e., if system performance is insufficient, video playback will be slow, lip sync will be lost. You can manually specify what frame types to decoder (for instance, skip B frames and decode only I and P, or decode only I frames), or you can specify decoder to choose what frames to skip automatically to keep up with sync.

'Do Postprocess' option is useful at low bitrates. It performs deblocking an deringing, which doesn't have any effect on high quality video.

'HQ Upsampling' option improves chroma visual quality significantly. Unless you care about 3-5% CPU usage, you may want to turn this on always. This option works only for YUY2 and UYVY output formats, however. And it doesn't have effect when DXVA h/w acceleration is in use - upsampling is peformed by video card hardware in this case.

'Double precision IDCT' will make decoding really slow. Too much slow. You can't even imagine how slow. This is intended for video editing applications. You will never notice visual quality difference during playback. And even after reencoding video 5-7 times.

[3ngel]

Well Peter, thanks for the explanation, but don't pretend to know i don't know what i'm talking about . In the specific case i need "Decode all frame" 'cause i'm not viewing it in realtime but only decoding to a lossless format, then i use also "Double precision IDCT" that yes it's d**n slow on decoding but the final quality is terrific , and now that i know, i'll use even "HQ Upsampling" for sure
In any case thanks for the explanation.

[peter]

If you are transcoding content, it is a good idea to turn one 'Decode all frames' - just to be sure decoder will not skip anything because downstream filters for some reason decide to send it quality messages. All frames have timestamps, so synchoriisation is guaranted.