T3i macro video at 3x digi-crop?

[skitron]

Has anyone shot any "macro" videos with the T3i 3x digital crop with a 100mm (or so) macro lens? I'd be interested in seeing clips if anyone has done it, plus of course hearing your opinions.

 

[NormanBates]

there's this:
http://cheesycam.com/testing-3x-w-100mm-f2-8l-is-macro/

 

[skitron]

Perfect! Thanks for the link, I gotta say I'm impressed with the 3x. From what I understand there is no loss in video resolution at 3x and from the looks of the video sample it appears the 3x may be very useful.

 

[gene_can_sing]

I seriously hope they have the 3x crop on the new 5d3. It has some serious artistic possibilities with that. Plus, it just makes the camera just that much more flexible.

 

[pgabor]

Perfect! Thanks for the link, I gotta say I'm impressed with the 3x. From what I understand there is no loss in video resolution at 3x and from the looks of the video sample it appears the 3x may be very useful.

The 3x looks like is a true crop of the sensor, and the higher settings are digital zoom. I made some calculations about this here on the forum:
http://www.canonrumors.com/forum/index.php/topic,657.msg8804.html#msg8804
http://www.canonrumors.com/forum/index.php/topic,657.msg8813.html#msg8813
But i recommend you to read the whole tread!

 

[NormanBates]

the 3x crop mode is decidedly soft compared to the standard video; it has less artifacts (no line-skipping-induced aliasing/moire), but if you see tests of 3x crop vs zooming with the lens, the 3x crop is comparatively soft:
http://vimeo.com/20964077

that's because, even if it is 2.7x instead of 3x, as pgabor says, it's too much: you need at least a 2.5K bayer pattern to create a sharp 2K image

still, if you need it, it's really useful; but it has limitations

 

[WarStreet]

the 3x crop mode is decidedly soft compared to the standard video; it has less artifacts (no line-skipping-induced aliasing/moire), but if you see tests of 3x crop vs zooming with the lens, the 3x crop is comparatively soft:
http://vimeo.com/20964077

that's because, even if it is 2.7x instead of 3x, as pgabor says, it's too much: you need at least a 2.5K bayer pattern to create a sharp 2K image

still, if you need it, it's really useful; but it has limitations

in 3x mode only a fraction of the lens glass is being used. The less glass used, the less the resolving power of the camera/lens combo. I think this is another factor for the softer result.

BTW. I like your video and your 'cheating' setup ;D

 

[NormanBates]

it's not my video, I just linked to it

on the softness issue: true, you're just looking at the center of the image, never at the corners, and you're magnifying your image, so resolving power issues become more important; still, having a 1.7K or 2K bayer pattern to resolve a 2K image will always lead to soft 2K images, no matter how sharp your glass is

 

[pgabor]

it's not my video, I just linked to it

on the softness issue: true, you're just looking at the center of the image, never at the corners, and you're magnifying your image, so resolving power issues become more important; still, having a 1.7K or 2K bayer pattern to resolve a 2K image will always lead to soft 2K images, no matter how sharp your glass is

The only problem with that comparison video is that he using a 18-135, a not too stellar lens, and between 18mm and 54mm, there is a significant difference between the resolving power. (for ex. look at the photozone test) (And consider that even the basic resolving power is relatively low with that lens, not speaking of the zoom mode) It would be a much clearer comparison with a 70-200 f2.8 mk2 or with some nasty telephoto lens (100 macro and a 300 f2.8, or 200 f2.0 and a 600mm, yeah it would be quite complicated to get those lens, but it would be the best possible comparison)

 

[skitron]

the 3x crop mode is decidedly soft compared to the standard video; it has less artifacts (no line-skipping-induced aliasing/moire), but if you see tests of 3x crop vs zooming with the lens, the 3x crop is comparatively soft:
http://vimeo.com/20964077

that's because, even if it is 2.7x instead of 3x, as pgabor says, it's too much: you need at least a 2.5K bayer pattern to create a sharp 2K image

still, if you need it, it's really useful; but it has limitations

Isn't the digi-crop (not the 'zoom' that goes beyond 2.7x or 3x or whatever the true number is) the same as cropping a still the same amount in post?

The first video example is thru a 100 L and it is in a different league than the second due to the lens, just like a 3x crop of a still shot thru a 100 L or some other hi-resolving lens is in a different league than any "short to short-tele" zoom.

Now I would agree that 3x on APS-C has it's limitations for the reasons stated, but using a stellar lens will make all the difference and it becomes quite useable as seen in the first video. And that is the same as I see when cropping a 100 L 3x vs cropping a 24-104 L at 100mm the same way (which is a better zoom than used in the second video).

So I'd have to agree that I'd love to see it on a 5d3 because that body hopefully will have enough more pixels to make the 3x fantastic when cropped in similar proportion. It would greatly expand the toolbox on the video side.

 

[NormanBates]

the ladybug video has been resized to 1280x720; you can get a sharp 1.3K image out of a 1.7K bayer pattern; I bet the original 1920x1080 footage looked like uprezzed 720p

 

[EYEONE]

it's not my video, I just linked to it

on the softness issue: true, you're just looking at the center of the image, never at the corners, and you're magnifying your image, so resolving power issues become more important; still, having a 1.7K or 2K bayer pattern to resolve a 2K image will always lead to soft 2K images, no matter how sharp your glass is

The only problem with that comparison video is that he using a 18-135, a not too stellar lens, and between 18mm and 54mm, there is a significant difference between the resolving power. (for ex. look at the photozone test) (And consider that even the basic resolving power is relatively low with that lens, not speaking of the zoom mode) It would be a much clearer comparison with a 70-200 f2.8 mk2 or with some nasty telephoto lens (100 macro and a 300 f2.8, or 200 f2.0 and a 600mm, yeah it would be quite complicated to get those lens, but it would be the best possible comparison)

I'm not really sure it would actually matter. The 70-200 can resolve more detail than the 18-135 yes, but it's only a 2mp image anyway. As long as both lenses resolve more tha 2mp does it matter?

 

[skitron]

I'm not really sure it would actually matter. The 70-200 can resolve more detail than the 18-135 yes, but it's only a 2mp image anyway. As long as both lenses resolve more tha 2mp does it matter?

I would think it matters because its a crop and the right question is can the 18-135 out resolve the 18 MP sensor. Why would cropping the center 2 MP for video be any different than cropping the center 2 MP of a still? In my experience there is no comparison between cropping shots this amount thru a 100 L and a 24-105 L and to me the difference in results look very much like the differences in the videos.