APS-C vs 1.4TC

which is better?

A. Full frame camera (5D MarkII 21MP) with Canon 100-400mm lens and 1.4 teleconverter, for a max effective 560mm reach.

B. APS-C camera (7D MarkII 20MP) with same lens and no teleconverter, for a max 620mm effective reach [correction: 640mm reach]

Better for what?

First of all, what you are comparing is 560mm FOV on FF vs. FOV on APS-C which is equivalent to 640mm on FF (400 * 1.6 = 640mm) not 620mm.

Assuming we start with superb lens (not the 100-400) with 1.4 teleconvertor, we know we can expect to lose about 10% of its max MTF due to teleconvertor.
Both measurements were done in the past by photozone.de tests of several Canon 70-200 lenses, although they have discontinued the practice now.

So if we started at 3400 line-pairs/picture width (LP/PW) we can expect to end up with a net 3060 LP/PW using 70-200mm f/2.8 L IS II. Using same lens on APS-C, photozone.de measured 2400 LP/PW. Both FF and APS-C cameras in the tests were 15Mpixel cameras, BTW.

So here we compare:

1. 24 x 36mm image on sensor, framing an 8.5' x 12.7' subject area at subject distance of 100', with 3060 LP/PW of detail resolution
   
2. 14.9 x 22 image on sensor, framing an 4.6' x 6.9' subject area at subject distance of 100', with 2400 LP/PW of detail resolution
   

1. we crop off 46% (by length) of the FF image (to equal APS-C image), then blow up that cropped area by magnification of 31.5x,
   netting 191 Line-pairs/Inch (3060 / 16 = 191)
   
2. we blow up the full APS-C image by magnification of 27x, netting 150 Line-pairs/Inch on print (2400 / 16= 150)
   

So let me get this straight....I'm better off--for resolution sake-- using a 1.4x converter on a FF camera than I would be using an APS-C camera (assuming both cameras have the same MP count)? That's a stunner.

"B" is better. It's cheaper and lighter (no 1.4TC). I actually have both cameras but don't have the 100-400 nor TC. I can't speak to the IQ but 7Dmk2 is better for long tele.

texshooter wrote in post #17488910
So let me get this straight....I'm better off--for resolution sake-- using a 1.4x converter on a FF camera than I would be using an APS-C camera (assuming both cameras have the same MP count)? That's a stunner.

Oops...significant f/up on my part!
1. we crop off 46% (by length) of the FF image (to equal APS-C image), then blow up that cropped area by magnification of 31.5x, netting 103 Line-pairs/Inch ((3060 * 0.54) / 16 = 103)

#2 is better than #1

Interesting, but I got lost in the story problem somewhere and would have to parallel diagram the two options to see where the train ended up and if I liked the view from the station when I got there.

Interesting question.

Edit to ask:
Wilt, does your 46% factor in the 1.4x TC? My tired head gives around 84% or so. (400×1.4)÷(400x1.6) to take the FF to the same zoom as the crop, but then you probably factor in frame size somehow. Tired and time for sleep, so subscribed to look at this tomorrow.

The 4.6' framed height for APS-C was determined by putting in 640mm FL, while 8.5' framed height for FF was determined by putting in 560mm for FF, so it did factor in the 1.4x teleconvertor when using the FOV calculator with the respective FL.

The 8.5' frame height needed to be trimmed to equal 4.6' frame height captured by the APS-C camera, so both 16x24 prints would have same content...
4.6 / 8.5 = 54%, so trim 46% off the FF image.

Thanks for the clarification. So, in this example, the crop camera with the 100-400L would potentially produce better results than a FF camera with the 100-400L and a 1.4x TC, disregarding the various differences in how each camera internally captures and processes the same scene and image, correct?

There must be a catch.

In the example above, the 7DII wins.

treebound wrote in post #17489556
Thanks for the clarification. So, in this example, the crop camera with the 100-400L would potentially produce better results than a FF camera with the 100-400L and a 1.4x TC, disregarding the various differences in how each camera internally captures and processes the same scene and image, correct?

Unfortunately the only near-real conclusion one can draw is IF we had access to MTF data from the two bodies specifically in question, as we have DP Review comparisons only from two 15MPixel bodies from the two format sizes. 5DII has about 2% higher pixel count in a single axis than 7DII, and we can figuratively think of them both as 20-21MPixel cameras, so the comparative results ought to be similar to the two 15Mpixel cameras. So let's do a theoretical projection of 5DII vs. 7DII from a slightly different angle...

• The Nyquist Limit predicts a similar outcome for line-pairs of resolution from the two cameras, where the entire FF full image is magnified by 1.6x less than the magnification applied to the APS-C full image, to produce the same 16x24 print.
  
• Since, as demonstrated by the earlier analysis, we are aware of the fact that the FF image has to be cropped by 46% before magnification to 16x24 coverage to achieve the same final subject size on print, we can predict that it ends up with 46% fewer line-pairs on the final print.
  

Wilt wrote in post #17489853

Since, as demonstrated by the earlier analysis, we are aware of the fact that the FF image has to be cropped by 46% before magnification to 16x24 coverage to achieve the same final subject size on print, we can predict that it ends up with 46% fewer line-pairs on the final print.
[/LIST]

I think I have steered my question in the wrong direction. I understand (at least I hope I do) about what you said about the need to crop the FF image to approximate the APS-C image. But the scope of my inquiry assumes nobody is doing any cropping. I am predicating my A or B option choice on the assumption that the image shot with the FF camera was made with "foot zooming", not Photoshop cropping. That is, each RAW file from either camera has the same corner to corner subject size. It is under these conditions that I ask which solution is best. In other words...

Does the loss of IQ caused by the 1.4 teleconverter exceed the IQ loss caused by the smaller APS-C sensor?
The whole point is I don't want to spend 12G on this bad boy...
http://www.usa.canon.c​om …p/ef_600mm_f_4l​_is_ii_usm
if I can get just as good results with a second smaller sensor camera body and this little guy...
http://www.usa.canon.c​om …00mm_f_4_5_5_6l​_is_ii_usm

texshooter wrote in post #17489985
I think I have steered my question in the wrong direction. I understand (at least I hope I do) about what you said about the need to crop the FF image to approximate the APS-C image. But the scope of my inquiry assumes nobody is doing any cropping. I am predicating my A or B option choice on the assumption that the image shot with the FF camera was made with "foot zooming", not Photoshop cropping. That is, each RAW file from either camera has the same corner to corner subject size. It is under these conditions that I ask which solution is best. In other words...

Does the loss of IQ caused by the 1.4 teleconverter exceed the IQ loss caused by the smaller APS-C sensor?
The whole point is I don't want to spend 12G on this bad boy...
http://www.usa.canon.c​om …p/ef_600mm_f_4l​_is_ii_usm
if I can get just as good results with a second smaller sensor camera body and this little guy...
http://www.usa.canon.c​om …00mm_f_4_5_5_6l​_is_ii_usm

IMO, no. I think having a 2x on FF would start to degrade as much as having crop factor, 1.4 is nothing although I havent shot that much with 1.4. I wouldnt mind doing a test some time, but not surprised if 1.4 FF beats out aps-c even when equalized.

texshooter wrote in post #17488726
which is better?

A. Full frame camera (5D MarkII 21MP) with Canon 100-400mm lens and 1.4 teleconverter, for a max effective 560mm reach.

B. APS-C camera (7D MarkII 20MP) with same lens and no teleconverter, for a max 620mm effective reach [correction: 640mm reach]

There is only one situation in which you will get better IQ (in one aspect) from the 5D2+1.4x, and that is when you have so much light that you can hand-hold the cameras and get get a good exposure without blur at base ISO, or have the lens on a tripod and your subject is still, and you can again shoot at ISO 100. That allows to collect more total light, and therefore, get better IQ if the exposure is all above the zone where the 5D2's legendary banding noise is present.

The 7D2 has much higher quality noise than the 5D2, almost devoid of color chunks and banding patterns.

Also, the TC always deteriorates the analog projection of the lens (even if ever so slightly with a high quality one), and the TC always compromises the AF ability of the system - and you're considering using it on an f/5.6 camera compared to not using it on an f/8 camera - the former won't AF at all unless you use a tricky (DGX) or non-reporting TC, or tape the pins, and even then, its attempts to focus will likely fail.

So, unless you are shooting stationary things at ISO 100 and plan to keep shadows dark, if present, and are willing to manually focus, the 7D2 and no TC is a no-brainer. You can use the TC with it, too, although you will lose AF speed and again, compromise the analog projection slightly (but still get more detail.

Here's my history since the 50D for shooting birds:

50D and 100-400. Used TCs often with manual focus. IS had a habit of jumping with this body, blurring about half of the shots with a several-pixel blur in the same direction every time.

Bought a 5D2, and the jumping went away, and high-ISO noise was a little better.

Bought a 7D, and found that for focal-length-limited work there was no longer any benefit to the 5D2 (5D2 was only useful for shooting wide angles and at wide apertures with shallower DOF).

Bought a 6D; found that despite the lower number of pixels-on-subject, the 6D gave less noise in focal-length-limited situations, so I often carried both the 7D and 6D and used the 6D when the light became scarce.

Bought a Tamron 150-600, and the 100-400 became pretty much a paperweight.

Bought a 7D2, and resigned the 6D to wide angle and aperture shooting, like I did previously with the 5D2. The 7D2 outperforms the 6D in IQ in both focal-length-limited situations, and in equivalency (same DOF and FOV), mostly in resolution, with similar noise.

As you can see, I improved on the 5D2 in steps, and still, in the end, the 7D2 won over the cameras that already bested the 5D2.