Canon 1.4x tele vs. Genuine Fractals

[brett_cole]

I'd be interested to see a comparison test of shooting a scene with the 1.4x

tele, versus shooting it without and then using GF to enlarge it 140%. A real

test, not opinion as to what might be better.

 

Thanks,

 

Brett

[rainer_t]

Very likely the answer is 'it depends' ... in

 

http://www.bobatkins.com/photography/tutorials/tc3.html

 

... you find examples where simple bicubic upsizing (not even something fancy like GF) gives better results than a TC does.

 

So, likely upsizing isn't bad ... if the TC can beat upsizing depends on how well the TC works together with the lens.

 

Sorry, cant help you with a real test ... so eventually ignore this

if you find this is just an opinion.

[brett_cole]

Thanks, this is what I was looking for

[andy_jones2]

Quick answer : the better the lens, the better the effect of the 1.4x. Canon's were optically

designed to complement the Canon super tele series (300, 400 500 and 600 primes), so

they don't usually work as well with zooms (as Bob showed) When I had a 300 2.8 I was

curious and found the 2x was better than digitally upscaling (can't find the images now)

but with that and the 500/4 my experience is similar to Bob's : 1.4x better than upscaling

while the 2x was marginally better. The extent to which all this matters is obviously

dependent on the final image and how it's viewed. Original pixels count for more when

you plan to print "to the limit" rather than downscaling for e.g. webdisplay.

 

Andy

[brett_cole]

Should have stated I'm using a 400 5.6

[mark u]

Using a 1.4x on a f/5.6 lens is likely to leave you with focus problems, as discussed in your other question. The quality of the software upsized shot is going to depend on your camera sensor and on the nature of the subject. Some subjects upsize better with different algorithms. You may also need to take some account of shutter speed and camera shake - at a maximum of f/8 you will be forced to use a higher ISO than at f/5.6 in all probability, which will also affect the comparative results.

[lester_wareham]

If have found both the 1.4X and 2X (as well as 2X vs 1.4X image) to be better that bicubic upsizing with both the 300/4 IS and 200/2.8.

 

It does indeed depend on the lens quality, but even the lesser primes work well.

 

Remember, if the information is not there all Genuine Fractals can do is create information, ie extrapolate, ie guess.

 

I have seen comparisons with GF and the more recent bicubic methods and the difference did not seem to be much.

[lester_wareham]

PS. I have a write-up of that here http://www.zen20934.zen.co.uk/photography/LensTests/tc_comp/index.htm

[bobatkins]

On a good lens, the 1.4x will always be better. You can't add information that's not there. Genuine Fractals can make a good guess, but that's all it is, a calculated guess at the missing information.

 

With a bad lens, it could be a close call, and GF might even look better.

 

See http://www.bobatkins.com/photography/tutorials/tc3.html and http://www.bobatkins.com/photography/reviews/canon_EF_70-300_II.html for some examples (using bicubic interpolation rather than GF though).