My theory of Video Enhancing.

[Arkamond]

I love your program. I really do. But I have theories of how to use other filters with SR. Such as not to sharpen after a Deblock. And to put multiple SR filters such as 1280x720 then to 1920x1080 and use a sharpen between those two and a deblock afterwards. To get rid of borders on something I use Any Video Converter because I can play the video and crop it with a box and get the rez of the cropped part. so then I set the video output conversion size to the size of the crop and use a FFV1 lossless codec to lose (nearly) no quality. I also put that into a MKV becuase for some reason my codecs dont like AVI for use in VE. I have tried really really long processes but my comp gets to 70C (Intel I5 quad 760 2.8 ghz.) (overpriced dell...) So I dont do long processes. Maybe less complicated processes are better? I am not sure, but I would like your take on my theories and maybe provide me with some of your own? Thanks.

[yourmom]

Hey Arkamond! Howareya?

In regards to the overheating, can I ask what type of pc you use? (Desktop? Laptop? Well ventilated? etc).

I'm not too sure about your other methods though. I mean, I was always under the impression that one should not attempt 'any' sharpening on their videos until the video itself is at it's final resolution.

In regards to your resolutions stated, I dont think I can comment too much on this as I do not know what resolution your original videos are using. For optimum usage of this super resolution technology, one would need to set video enhancer to twice the resolution of the original (so if your videos are 640x480 for example, the most optimum output would be 1280x960. One would repeat the super resolution process if the results are too small, or downscale the video if the results are too big). It has also been mentioned on these forums that you can obtain a 'marginal' improvement if you were to reverse the etire video before upscaling it a second time (See http://forum.infognition.com/index.php/topic,277.0.html).

Between edits, I personally prefer to save the results as an uncompressed avi. When I aproach the point in my project where I have a final uncompressed video which is ready for mkv compression, I prefer to use an external x264 encoder. The particular means that I use to perform this procedure at the moment involve the software 'MeGUI'.

As for the majority of methods that I personally use, I have made a post about it at http://forum.infognition.com/index.php/topic,566.0.html if you would like to give that a read later

[Arkamond]

I use a Desktop Studio XPS 8800 by Dell. (Atleast I think its that number!) But anyways I have seen better output if you put your filters right after the largest SR and before you put it down to final rez in VE. I usually do: Double Rez (Default size) then depending on how large the frame size is already do a second double rez with default size. Then I use A MSU Smart deblocker or other filter. (Works best at larger sizes!)  Then I use another SR to 1920x1080. theoreticly when the frames are larger the filters can notice more in the frames and fix more that may be hiding in smaller frames. So when its downsized it looks really good. I have actually seen this with most video's I enhance.

[Manoman455]

Yeah I was referenced lol.
SR to 1280x720 then sharpening it would mean you sharpen the actual blocking as well.
SR again to 1920x1080 then deblocking will be quite CPU intensive for mid-range PCs.
Deblocking might not be as effecient and accurate, since the deblocker might think that the sharpened blocks are actual lines in the video and will leave them alone. I might be wrong, I never used it like that.