LAV vs ffdshow (1 Viewer)

[Rbentley100]

Hi, I am just taking a new look at my post processing and seem to have forgotton everything.

I currently use LAV splitter and LAV audio to bitstream and handle my subtitles etc. It is my video I am now curious about.

LAV video offers hardware acceleration but I dont know how much is offered under the hood. i.e. oes it offer upscaling SD content?

Or would I be better to decode video with ffdshow as described in the mediaportal wiki? Or use both? If so which order should I use them in, LAV 1st the ffd?

Thanks in advance for any clarification.

(P.S. If I am missing a 3rd option please let me know )

 

[kiwijunglist]

3rd option: you can use separate decoders for mpeg2 (usually SD) and h264 (usually HD)

 

[Owlsroost]

LAV video does not do upscaling etc (it's just a decoder) - you'll need to use FFDShow for that.

If you use LAV for decoding, you can't use 'DXVA2 Native' decoding mode with post-processing filters (e.g. FFDShow) because it needs to connect directly to the renderer.

Tony

 

[Rbentley100]

ok, thanks guys,

so if I add both ffdshow raw video and LAV video decoder to the post processing, then setup LAV just to work on h264 (via its properties, there is an option on the right hand side of the 1st tab, then deselect most options on the 2nd tab)

then setup ffdshow as described in the mediaportal wiki (blur/sharpen/resize etc) and just stop it from doing anything on h264 video then this will work?

on my post prossesing I now have LAV splitter LAV audio, LAV video and now ffdshow raw video.

What do you guys tend to run with?

Cheers again for the help

 

[Owlsroost]

Don't put filters into the post-processing section in MP unless you actually want to use them for that - it can mess up the graph building. MP 1.3.0 Beta gives you more control over which decoders are used for which format (so you can use different decoders for MPEG2, H264 and VC1, for example). Any post-processing filters are added after the decode filter (in the filter graph), so the input to the post-processing filter is decoded raw video (in one of the RGB or YUV formats).

Tony

 

[Rbentley100]

ok, cheers. Sorry about the noobish questions. all the processing available always confuses me

So I should look to 1.3.0 beta. and changing the decode filters for different video formats? As by the time it reaches post processing I have to make a choice between ffdshow and LAV?

This leads to my next question, I have LAV splitter as a post process. So before it hits post processing everything is just an mkv file isnt it? So how can the decoder filters differentiate from HD and SD, and therefore work out which filter to use?

Surely everyone has some setup to get the best video qulity from mediaportal? ie upscale SD content and use hardware acceleration for HD content? I am not tied to LAV or ffdshow tbh, I just have used LAV for a while now and like how it handles forced subtitles and bitstreaming audio.

I have tried multiple google searches etc and havent managed to find a comprehensive guide. They often just point to codec packs like k-lite and shark, now I am sure these work very well but they add a lot of stuff I am not interested in and all they usually do is apply everything else in a nice big bundle, I prefer applying just specifically what I want/need. If there is a guide or something to read somehwere that I am missing that would also be great.

Sorry about the ignorant questions, thank you for yor patience

 

[kiwijunglist]

I suggest you do one of these options

1. Everything set to LAV except mpeg2 which is set to ffdshow. Don't use any post processing. Setup whatever upscaling you want within the ffdshow settings

2. Check SAF, i think it has built in upscaling profiles for SD.

 

[Rbentley100]

Cheers for the help, I am now running mediaportal 1.3 like this (mpeg 2 as ffdshow, rest using LAV) and it is all looking good for now. havent tested forced subtitles etc yet though

 

[SciDoctor]

You can also try LAV set to DXVA (copy back) as this is supposed to allow connection to post processing. Will increase cpu usage as data is 'is copied back' from GPU memory to CPU memory. ATI being more cpu intensive so than NV in this respect.