you can't change it on dual-graphics laptops if this's what you're talking about

In any case conversion between YV12 and NV12 color formats is lossless so you don't need to worry about it.
They different only in order of color planes in memory. Why ffdshow outputs in NV12 by default? Probably because of

NV12 is the preferred 4:2:0 pixel format for DirectX VA.


> but why the conversion between 1280*720 to 2048*720 occurs in ffdshow

for example, if you have 2048-width screen and "outer lighting" in SVP is on

3,828

(10 replies, posted in Using SVP)

this error is about SVP 4/plugins/libsvpflow*.dll are missing or can't be loaded for some reason

Read this first

The last SVP version supporting 10.10 is 4.2.146!

It's impossible now to install Vapoursynth libs (and mpv with Vapoursynth support obviously) via Homebrew due to missing zimg dependency - Homebrew says

zimg: macOS El Capitan or newer is required.

What we can do? We can build zimg, Vapoursynth and mpv manually and place them into Homebrew's "cellar".

So, considering a clean OS X install, the following should work:

1. Install Homebrew

/usr/bin/ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"

2. Install all the mpv dependecies:

brew install libass tesseract python3 mpv

this can take a while, especially on 10.9
3. Remove installed mpv cause we gonna replace it with our own build:

brew unlink mpv
rm -fR /usr/local/Cellar/mpv

4. Unpack .tgz with pre-built zimg, vapoursynth and mpv into /usr/local/Cellar
5. Link our binaries into /usr/local:

brew link --overwrite zimg vapoursynth mpv

6. Restart SVP

> when i first installed SVP right before the performance test it said i was using OpenCL 1.2 and that it was known to cause freezes and crashing...?

it's a warning message about that
probably a little bit outdated now hmm

3,831

(1 replies, posted in Using SVP)

> i was wondering if there's a way to adjust the frame buffer ahead on SVP4

why?
see All settings -> frc.prefetch section.

>  have 32GB RAM so it would be nice to use at least more than 4GB on 64bit MPC-HC

you can try 8K videos from Youtube big_smile

> if there's a way to utilize more GPU resource using OpenCL 1.2

nope

> the GPU ID is 11 so what does this mean?

1st GPU device on 1st OpenCL platform

3,832

(15 replies, posted in Using SVP)

Still I don't know any video conversion manuals for macOS.
If you find one - let us know wink

3,833

(5 replies, posted in Using SVP)

10-bit is only for mpv and VLC

neither ffdshow nor PotPlayer doesn't support 10-bit Avisynth processing


> Do I set the output of ffdshow raw video filter to 10- or 8-bit?

8-bit

3,834

(22 replies, posted in Using SVP)

as I said before, may be (!) the major difference comes from

prefiltered = RemoveGrain(C, 22)

nothing can stop you from pre-process video before svpflow

=======

prefiltered = RemoveGrain(input, 22)
super = SVSuper(prefilter,"{gpu:1}")
vectors = SVAnalyse(super, "{refine:[{thsad:100}]}", src=prefilter)
smooth = SVSmoothFps(input, super, vectors, "{rate:{num:2,den:1}, algo:21, mask:{area:50}}", mt=threads)

=======

 BlkSize    = Default(BlkSize, C.Width>1200||C.Height>900 ? 32 : C.Width>720||C.Height>480 ? 16 : 8)

you know you'll get significantly different results for the same video scaled to different sizes, right?

3,835

(13 replies, posted in Using SVP)

> doesn't SVP process and output at 4:4:4?

nope

you can get 4:4:4 from 4K YV12 only if you'll scale down to 1080p after SVP and the scaler is smart enough to not scale chroma planes (i.e. - madVR, not sure about EVR)

3,836

(7 replies, posted in Using SVP)

Use 64-bit MPC-HC.
Otherwise you may need to reduce number of processing threads in SVP down to 9 or even more (I haven't tried it with 10-bit 4K in 32-bit MPC-HC)

LoadPlugin("...\svpflow1_vs.dll")

3,838

(15 replies, posted in Using SVP)

you probably mean "how can I convert video to 60 fps on Mac" otherwise your question doesn't make any sense hmm

I'll just leave it here big_smile

http://i.imgur.com/69AOjNS.png

3,840

(22 replies, posted in Using SVP)

yep

the most common reason is a major Windows updates, after that you'll have to re-install graphics drivers manually

3,842

(15 replies, posted in Using SVP)

what do you want to do?
download what video from where?

3,843

(22 replies, posted in Using SVP)

> Is this the same method you use?

this's the idea
mask resulting frame with two source frames in "bad" areas, where "bad area" means "area with high SAD value of the motion vectors"

"external" implementation requires 10 lines of avisynth code and is slow as hell
"internal" implementation requires just another multiply-add operation for each frame pixel

3,844

(22 replies, posted in Using SVP)

have you read that "YFRC" sources? hmm

3,845

(15 replies, posted in Using SVP)

if you're asking about SVPtube - drop-down menu on the "Play" button -> "Download"
ffmpeg binary must be installed

3,846

(3 replies, posted in Using SVP)

Nothing can stop anyone to take .dll/.so/.dylib from the SVP installation wink

But I agree that it should be updated. We'll do it soon enough...
However there's nothing useful in the Avisynth part since rev.128

3,847

(1 replies, posted in Using SVP)

As long as it allows processing through ffdshow raw filter and provides stable source frame rate - why not?
ProgDVB used to work at least some time ago...

> so 1800x beats 7700k for svp right?

~1.5 times

I'm not sure if 7700k can do highest SVP settings with 4K, while mine 1700 @3500 @1.1V @60W can smile

Indeed.
However I personally prefer 1700 smile

And 1070 is an overkill "if u dont need it for games"

> i5 7600 or ryzen 1700?

this's not even a question
ryzen is 2 times faster