Same effect is also called pixellation.
I am often asked how to remove mosaic from videos using Video Enhancer. Frequently people ask what filter to use to remove mosaic. Most of these people are Chinese lovers of adult content which they often have with most interesting parts pixellated. ;) Some sites and forums even advertise Video Enhancer as a tool for removing mosaic. I think they are mislead by pictures from Video Enhancer's main page, where we see low resolution video frame enlarged 4 times, so it looks like it's a big mosaiced frame. That's not the case actually, it's just zoomed small picture.
In general, Video Enhancer is not for removing mosaic and there are no filters capable of doing it. There is a way to improve mosaiced parts, but this will also degrade heavily quality of unmosaiced part of image. This is what one should do if he or she wants to get rid of mosaic in video:
- Get VirtualDub and Video Enhancer.
- Open your video in VirtualDub, zoom in your video and measure the size of squares in mosaiced part. Let's call this number N. In example above each square is 4x4 pixels, so N=4.
- Apply resize filter, make new size N times smaller (so 360x288 video from example would become 90x72) and choose bilinear method.
- Choose a lossless codec (like Huffyuv, Lagarith or MSU Lossless Codec)
and save your video in AVI file. You'll have a small video with no
squares in it:
- Open it in Video Enhancer, go to Advanced mode and add SR (Super Resolution) filter several times, each time doubling the size of video, each time in high quality mode. If N=4 you need 2 SR filters, if N=8 you need 3 SRs and so on.
- After you made this chain of SRs, choose output compression for video and optionally for audio and press Start.
If you successfully performed all steps and you're lucky enough, you'll get your video with the same dimensions as original mosaiced video, but processed with Super Resolution. You'll lose details of unmosaiced parts, but in many cases increase quality of mosaiced part. Here is an example of what you'll have:
One may ask: why downsize video in VirtualDub if Video Enhancer also can downsize video (just set small size in SR filter)? The answer is if you downsize video in Video Enhancer it uses Lanczos3 method which makes a sharper image. This image is harder for super-resolution to restore, you'll get more artifacts, result will look worse.