How to soften the image of a video without being blurry?

[Andrew Lenczycki]

I played with the Soft Contrast FX from the Sony Vegas Pro 10 Video FX tab. I was able to soften the image I had by playing with the slider controls. Sometimes experimentation is your friend.

Andrew Lenczycki

[Raf Erosa]

I’m experiementing now. Have u watched Access hollywood or similar shows, and noticed that soft image? Im trying to do produce that look.

[Dave Haynie]

Technically speaking, a soft focus IS a blur — just a very visually pleasing blur.

There are really two options. You can soften a visual a bit by lowering contrast, but there’s only so far that can go.

Beyond that, you are blurring. In signal processing terms, the “hardness” is due to high frequency information: sharp, high contract lines, etc. To fix this, you run a low-pass filter over the visual — you’re filtering out some aspect of the higher frequency information. That’s exactly the same thing as a blur.

But you’ll problem notice, there are hundreds of different ways to blur something… that actually makes a ton of sense when you think in signal processing terms. Your “blur” is going to remove some high frequency information. It can be totally brute force, or very specific, about what’s being removed, and just how that’s done. You can blur simply by filtering high frequency data, by spreading just the high frequency stuff, even by lowering the amplitude of the high frequency data versus the lower frequency information, rather than removing it completely. Ok, I’m a geeky engineer.

In practical terms, your eyes tell the truth. Vegas has several blurring options… try them, but apply the blur sparingly. Most blurring algorithms have a number of degrees of freedom — different parameters that can be adjusted. For example, averaging adjacent pixels is the simplest form of a low-pass filter. If you totally average all adjacent pixels, you cut the resolution in half. But most algorithms let you tweak things, like the radius over which averaging takes place (the circle around each pixel that’s factored in to the new value of that pixel), and the strength of that averaging (do I add in 5%, 10%, 50%.. of the surrounding information). More sophisticated algorithms may only function on line detection, contrast detection, etc.

-Dave