Ok..this is a long- worth it though:

When Canon announced that the new T3i would have movie digital zoom, I initially believed it would be a ‘crop zoom’ type thing (where the image is just magnified as it is being recorded, hence, great loss of resolution), but in their announcement, they were very specific that the new ‘Movie Digital Zoom’ would use a smaller part of the sensor to record video, and video users could record video 3-10x ‘while still maintaining full HD quality’.

Here is the clip from the announcement:
Ok..this is a long- worth it though:

When Canon announced that the new T3i would have movie digital zoom, I initially believed it would be a ‘crop zoom’ type thing (where the image is just magnified as it is being recorded, hence, great loss of resolution), but in their announcement, they were very specific that the new ‘Movie Digital Zoom’ would use a smaller part of the sensor to record video, and video users could record video 3-10x ‘while still maintaining full HD quality’.

Here is the clip from the announcement:
I love Canon. When I first read this- I was excited and in utter awe of how in the world they could make this work. I figured the only way they could pull off 10x Zoom was in the wording of the definition of the total pixel surface area, . 5184 x 3456= ~17.9 MP, or about 18 MP. 1920×1080 = ~2MP, or just under 10 times the total number of pixels used. (You can read the full article here): How the T3i’s Digital Zoom (Probably) Workscreating new pixels to fill in the gaps.

When I first got my T3i, the Movie Digital Zoom feature was the first thing I checked out. The 3x looked pretty good on my little screen and I was impressed with it, 5-10x not so much.

So my question became “If the Movie Digital Zoom feature is using a (much) smaller part of the sensor to zoom in, just really how good is the resolution of the image?

or in other words:

Putting aside all this technical non-sense…how GOOD does it look? How practical is it? Methodology-

I decided to do some tests on my own

1. I created a compression measurement tool composing of a 45 degree black poster board cut with a razor and glued onto a white square poster board creating a diagonal line. Due to the square nature of pixels, true diagonal lines, especially when zoomed in can tell you a lot about what is happening compression wise with an image. The compression area on either side of the line can even be measured if needed. If Canon has figured a way to preserve True HD Resolution (at least what the definition of it is in my mind) these compression zones should appear similar, all things being equal.

2. I also needed something a little more organic, like a beautiful model, to see what the compression effects would be on something say, “human”.

3. I would shoot the control at 1080, 30fps, capturing both the model and the board. Because shooting 3x-10x greatly changes the size of the subject, I needed a way to recompose and maintain their approximate perspectives- after zooming 3x, 5x, 7x, and 10x so they were relatively the same size in each image. I decided to have my model/board stay put, and after changing the movie digital zoom amount, would walk back (and forth) on a very, very long beach, until the measuring board filled the perfectly square focusing guide for Movie Digital Zoom. (Please give me some clever points for figuring this out). The focal length of the lens itself would stay fixed.

4. All movies were recorded with a Canon T3i affixed with a 70-200mm 2.8 L lens, shooting at f4.5, 1/1250 and an ISO of 100.

I shot about 20-30 seconds of video at each setting:

1080 (Control)
MDZ @ 3x
MDZ @ 5z
MDZ @ 7x
MDZ @ 10x

Videos were uploaded into Adobe Premiere Pro, and snapshots were taken from each distance. These snapshots were then imported as layers into Photoshop CS5 and aligned using guide lines. Stacked images of both the diagonal and model’s face were then zoomed in and compared. I have included the results for you here.

Notes: To be consistent in focusing, I used the “Focus” feature built into the camera and recorded only after achieving focus lock indicated by the box turning green. Manual focusing would have proven to be too inconsistent and subjective.

— The greater the distance, the less practical panning and finding my subject became. At the 10x distance, I was lining up my shots by gently tapping on the tripod.
— Wind probably effected the results. While there was an occasional gentle breeze, for the most part I blocked this with my body. Snapshots taken from the video were only the best, and clearest.
— Water Evaporation (ie the Oasis effect) may have also come into play creating some distortion.
— Vibrations in the ground could have also effected the results.Results: Frame Grabs:

(Anything with an “x” is referring to magnification in Movie Digital Zoom)Diagonal Comparison- Magnified at 1200%Face of Model Comparison @ 200% Magnification Summary:

While the Canon T3i Digital Zoom is an extremely cool feature and idea, the results indicate that the images do not appear to be Full HD appearance when compared with the control. It is possible the video images as recorded have HD Pixel dimensions,- my results show that image quality degrades as the Movie Digital Zoom feature is zoomed in. Granted there were other factors such as wind and vibrations which may have played into this, my opinion is that Movie Digital Zoom is not high quality in appearance.

That said, Movie Digital Zoom represents a very interesting wrinkle in Canon’s video technology. While it isn’t high quality HD video, and very difficult to track moving subjects, it could have some interesting and usable applications for something like surveillance.

There will be a little Youtube Video coming soon showing the video clips.

If you are the proud owner of the Canon T3i, I have a fantastic training video on it ready for immediate download:

Canon T3i Crash Course Video Training