Sony Clear Image Zoom vs. Digital Extender

Which is better, and how do they each work? In this rather crude example, I'm comparing images shot on the Z90 with Clear Image Zoom and Digital Extender (center scan).

This video is split into 3 main parts.

1. Introduction
2. Comparison of the two images
3. A fully illustrated look at how I believe these two technologies work and differ. As I don't have full Sony details, some of this is educated guesswork. You have been warned! This is YouTube, after all.

What have your experiences been? Did I get this right and does the explanation make sense? Let me know in the comments if you can add more detail to this.

Thanks for watching.

[TedTheFiddler]

Very Interesting. I think though what you end up going with is always going to depend on what your content and audience will be. You use such an extreme zoom in the first place either because you have to, doing say... Wildlife or Sports, or News Gathering, or for Pure Artistic Expression.

[KrunoslavStifter]

In case you are not familiar how Sony Clear Image Zoom works. Here is the explanation. It took a bit of research but basically Sony has a database backed into camera software somewhere in the background where they camera tries to identify what kind of image are you pointing at. It is not A.I. powered but it does use a clever algorithm. Basically lest say you point at glass buildings or something where there is not a lot of high frequency details, the camera will compare the image it sees and its database and based on comparison it will choose a different interpolation or enlargement method. What makes it clever is that this algorithm will work dynamically. It will constantly compare what the camera is pointing out with its database or frame by frame basis and chose algorithm it thinks its most appropriate for the enlarging.

If you point your camera at the cat, because it has fur (high frequency detail) it will choose different interpolation method than when you point the camera at for example blank wall or sky where there is low frequency details. It will adjust in real time, frame by frame automatically, and since you are feeding the camera with new data all the time, it will try to always pick the best method for enlarging.

When you use most methods in post production they would typically let you choose one interpolation method for all frames so the results tend to be less optimal. There are some new algorithms powered by AI like GigaPixel from Topaz etc that can try to invent details based on its own database, but they are very slow to process your videos.

Sony is simply allowing us to use little processing power and no AI technology to do more than what we would normally be able to do. Its actually older technology Sony developed but its very clever solution. Off course because it does not crop anything it can be used on any shooting mode camera supports as long as the processing is not too demanding. 24 fps for example in 4K or something, but if you had to shoot 120 fps or something like that it might be too much for some CPU's on older cameras to handle.

Now my question: What is "Digital Extender" I have not heard about that term. Is that your way of saying what everyone calls cropping in? Or is this some kind of legit mode called this way?

Cheers!

[chrisw]

so you take the 5k sensor and double it, your at 2.5k, thats why you can clear image zoom, plus digital extend to just around 1080p, and get a near 30 times zoom with very little image degradation. I think.

[HotCarTrack]

Is there a massive quality drop when Digital Extender is engaged? Or would I be better off shooting 4K and then cropping in post. Cheers