Downside of using FF lenses on crop bodies

wallstreetoneil wrote in post #17924841
Two things:

1) don't confuse 'look so much better', which has a lot to do with rendering and colours, with resolving power / potential mega pixels
2) you did not mention your apsc lens and Tony did not say it is true of every lens - what he was pointing out was that you have to test actual resolving power, as DXO has done, to know the resolving power of each lens on each camera

I am going to re-watch the video again and do some testing, but part of what he said I think came over in a slightly confusing manner. For example, let's take the 5D3, the 5DSR, the 7D2 and the 35mm L II and look at what DXO says about Sharpness and P-Mpix

5D3 - Sharpness = 18 P-Mpix
5DSR - Sharpness = 37 P-Mpix
7D2 - Sharpness = 14 P-Mpix

What Tony said, and may be confusing, is that a FF Lens on a Crop body might not be the best idea. I think he maybe made a mistake or has possibly misinterpreted DXO data - or I am incorrect. But if you look at the above data, it is interesting if you do a calculation on the 5DSR 24 P-Mpix and convert it into the size of a Crop sensor.

If you take a Canon Full Frame sensor resolving ability / sharpness and divide it 1.6 and then divide it by 1.6 again, this will tell you the resolving ability of that full frame sensor if it was converted into an APSC sized sensor. So, given that we know that the 5DSR is a glorified Full frame sized 7D2 sensor, if we divide the 5DSR's 37 P-Mpix by 1.6 and by 1.6 again (37 / (1.6^2) ) = 14.4 which is exactly what the 7D2 Sensor is resolving (coincidence? - don't think so). Tony talks about this in another video and mentions it in this video and that is if you are going to crop the image from a full frame sensor, to make it close to a Crop size, you very likely should have used a Crop to get better resolving ability. In the case that I just showed, we already know that a 5DSR basically has a 7D2 sensor imbedded in it and DXO's number seem to show this. If you take the same 5D3 sensor and divide by 1.6^2 you get approx 7 P-Mpix - so this would say, that you should use the 7D2 with the 35mm L II if you need to crop the 5D3 to a crop size (which everyone already knows).

Interesting that the crop lens that Tony mentions, the Sigma 18-38 F1.8, that he mentions as basically the sharpest crop sensor lens you can buy, that DXO states the P-Mpix = 13 P-Mpix on the 7D2 - which is slightly less that Canon's new latest and greatest, sharpness out of this world, 35L II at 14 P-Mpix on the 7D2 as mentioned above.

I think there maybe something in what Tony is saying but I have to figure it out because I think it could be partly wrong.

Quite right, on reviewing my post I found that I had deleted the bit about which lenses I used when reviewing it - sorry, my bad. I tried 4 EFS lenses the Canon 10-22 and 55-250 and an 18-55 (discount this one as it has mechanical problems) and the 60mm EFS Macro.
As you would expect the 60mm is superb on all formats (with an extension tube for APSH and FF) and exceeds my Canon 100 F2.8 macro in IQ - just! But that is quite a compliment to the 60mm! Perhaps it is not fair to expect too much from the 55-250 IS II as it is a very cheap lens though it does deliver remarkably good results for it's price. But in the 55 - 70mm range it did not fare well in detail resolution (or any other aspect) compared to my 24-70 - but then I wouldn't expect it to. The 10-22 was the highest quality lens of the bunch and I have tried a number of examples over the years. Back when I shot only APSC I bought the Canon 17-40 F4L in preference to the 10-22 due to it's better IQ (though not detail resolution) and the focal range was fine for my uses back then. Now I have replaced the 17-40 with the 16-35 F4 L IS (1 year ago) and really the 10-22 is not in the running anymore.
Try them for yourself and see what you think. To me the later L series lenses have significant advantages regardless of sensor size however they are very expensive! It is also remarkable how good the latest crop of EFS lenses are for the price - note for the price - not in absolute terms.
I don't pretend to have done exhaustive testing (because I haven't!) just relating my observations.

Boy this video proves how urban myths are spread, which have mistruths embedded in them! And viewers hear and propogate the BS heard in the video, even when it is wrong.

He says, "The aperture become a higher f/stop number effectively". UNTRUE...

• While the APS-C body sees a smaller FOV because it is only about 15 x 22.5mm, compared to FF 24 x 36mm sensor, that does not alter the intensity of light striking same AREA.
  
• The amount of light striking the sensor is UNCHANGED simply when mounting the lens on a smaller physical sensor, it does NOT alter the intensity of the light falling on a given area of the sensor (let's restate "...the intensity of light falling on one square millimeter is identical for both APS-C body and for FF body")
  
• The f/stop number behavior changes only in one sense, the DOF captured in an identical sized print from both FF and APS-C body...
  
  If you look at 100mm lens on APS-C body, and 160mm lens on FF body, assuming you have f/4 aperture on APS-C you need to use f/4 * 1.6 on the FF body to have effectively the 'same DOF'... APS-C 100mm f/4 DOF = FF 160mm f/6.4 DOF. For example, at 100' distance FOV is 14.8' x 22.4' for both and DOF zone is 39' deep (84' - 123')
  

• Using photozone.de tests of 35mm f/1.4 lens on both 20MPixel APS-C and 20MPixel FF bodies,
  
• the MTF value at f/2.8 is 3869 lp/ph (line pairs per picture height) on 20MPixel FF while it is 3380 lp/ph on 20MPixel APS-C body. In other words, with a 1.6x smaller FOV image from APS-C has 13% less total detail on the same size print.
  
  
• Using photozone.de tests, this time testing of 24-70mm f/4 USM L IS lens,
  with 38mm f/6.4 on 20 MPixel FF vs. 24mm f/4 on 15MPixel APS-C so that area of view is identical on both and DOF is identical on both...(although, due to lack of comparable MPixel sensor testing with same zoom lens)
  ...38mm f/6.4 on 20 MPixel FF has MTF about 3555 vs. 24mm f/4 on 20 Mpixel APS-C MTF of 2369. So using same zoom on both, at same FOV and same DOF on both cameras the MTF rating is 33% less with APS-C on the same size print, in spite of the fact that there is a 14% lower pixel count in a single axis with the tested cameras.
  

• From above we already know that 24-70mm at 38mm f/6.4 on 20Mpixel FF has MTF about 3555 and comparing it to 18-55mm at 24mm f/4 on APS-C MTF of 2499. About 30% less detail from APS-C and the kit lens on the same size print, in spite of the fact that there is a 14% lower pixel count in a single axis with the FF camera.

Wilt, you could have saved yourself a lot of time by just paying attention to the video. He didn't compare the 24-70 on FF to the 18-55 on crop. He compared the 24-70 V1 on the 7D to the 18-55 on the 7D. And he used DXO's tests not his own

FEChariot wrote in post #17925601
Wilt, you could have saved yourself a lot of time by just paying attention to the video. He didn't compare the 24-70 on FF to the 18-55 on crop. He compared the 24-70 V1 on the 7D to the 18-55 on the 7D. And he used DXO's tests not his own

If I have to use specifically the lenses that he choses to mention, in order for the 'truth' to be validated, then it is ONE CASE and not a generalization that can be drawn about 'all FF lenses' vs. 'all APS-C lenses' to guide our purchases, does it?!

I did not cherry pick lenses for test that specifically disprove what he said. I chose based upon what was available from photozone.de and if the results corroborated his statements, great...the generalization seemed to hold true. But if the photozone.de results seemed to contradict his statements, the generalizations seem to be false, even if specific lens discussions mentioned by him were indeed valid.

Wilt wrote in post #17925611
If I have to use specifically the lenses that he choses to mention, in order for the 'truth' to be validated, then it is ONE CASE and not a generalization that can be drawn about 'all FF lenses' vs. 'all APS-C lenses' to guide our purchases, does it?!

He said in the video that the choice depends on the lens and in no way made a generalization about every lens.

FEChariot wrote in post #17925619
He said in the video that the choice depends on the lens and in no way made a generalization about every lens.

OK, that is fine. He did state, as a generality, "Lenses designed for smaller sensors will give you sharper results because they're focusing all that light that is coming into the lens onto the sensor itself." not true.

And what about the absolute untruth in the statement that you have to multiply the f/stop as well, as it transmits less light?!

Wilt wrote in post #17925622
OK, that is fine. He did state, as a generality, "Lenses designed for smaller sensors will give you sharper results because they're focusing all that light that is coming into the lens onto the sensor itself." not true.

And what about the absolute untruth in the statement that you have to multiply the f/stop as well, as it transmits less light?!

I believe that to be absolutely true.
I fail to see how changing the surface area of the receiver makes any difference to the lumens/mm squared transmitted through the aperture.

I'm with Wilt on that point.

Part of the reason for confusion on the amount of exposure vs. lens type and aperture is due ot the fact that (as was noted above) the f/ratio is intended to provide an easy method for comparing exposure across multiple lenses and bodies. The measurement which actually determines the amount of light reaching the pixels is not intuitive, but relates more to the actual aperture diameter and pixel size/pitch than the f/ratio.

Roger Clark does a very technical summation HERE.

GregDunn wrote in post #17926181
Part of the reason for confusion on the amount of exposure vs. lens type and aperture is due ot the fact that (as was noted above) the f/ratio is intended to provide an easy method for comparing exposure across multiple lenses and bodies. The measurement which actually determines the amount of light reaching the pixels is not intuitive, but relates more to the actual aperture diameter and pixel size/pitch than the f/ratio.

Roger Clark does a very technical summation HERE.

Yes, that may explain the absolute amount of light reaching an individual pixel, but all (and he recognizes this) that affects is the S/N ratio. When he starts talking about IQ, though, he introduces the angle of coverage of the individual pixels, which re-introduces the focal length into the question.

Some of his other assumptions are tailored to his argument, in particular that moving from one lens (300 f/4 in his example) a shorter focal length lens with a larger physical aperture (in his example the 300 f/2.8) is a better choice than a longer focal lens with nominally the same physical aperture (400 f/5.6). This would only be true if you shot the shorter lens wide open and were able to get enough pixels on target that the reduced cropping of the longer lens didn't negate the perceived advantages of the wider aperture. Is it better to fill the frame with your subject using a 400 mm f/5.6 lens or to have to crop out 44% of the image to get the same coverage using a 300 f/2.8, all other considerations being equal? If you can't move your camera position to match angle of view, then for any given sensor, the focal length that best suits your desired framing(neither too short nor too long) will give you the best image quality. Comparing framing between the two sensor sizes introduces the issue of sensor noise, which for comparable technologies will be better with the larger sensor (don't compare a 20D with a 5D3, for instance - they're several generations apart).

If you follow his advice, and only use lenses that are optimized for the aps-c sensor, you have a problem of availability. To go beyond the kit lens Holy Trinity of 10-18 stm, 18-55 stm, and 55-250 stm, you have the Sigma 18-35mm and 50-100mm F 1.8 ( which potentially may have AF issues, mfa or no mfa, dock or no dock), the EF-S 60mm macro, the EF-S 24mm pancake, and......nothing else Samyang does not make their specialized 21 F 1.4, 12mm F 2.0, and 50mm F 1.2 for EF-S, only micro 4/3 and the Canon M Mount (the do make the 16mm F 2.0 for EF-S though).

To summarize, you just don't have much choice of lenses available to follow his "use lenses designed for aps-c on aps-c bodies" advice.

Addendum-I forgot about the Tamron 60mm F 2.0 Macro optimized for APS-C-does not seem like that has been a very popular lens though, due to its reputation of being so so when shot at F 2.0

The Canon 17-55 f/2.8 IS is missing from your list, too.

...then there are the third party APS-C image circle lenses (with usual EF mount) made by Tokina, Tamron, etc.

• Sigma AF 12-24mm f/4.5-5.6 HSM DG II
  
• Tokina 11-16mm f/2.8 Pro DX
  
• Tamron AF 60mm f/2 SP Di II LD [IF] macro (which you listed)
  
• Tamron AF 17-50mm f/2.8 SP XR Di II LD Asph. [IF] VC
  
• Tamron AF 18-270mm f/3.5-6.3 Di II LD Asph. VC macro
  
• Tamron AF 11-18mm f/4.5-5.6 SP Di II LD Aspherical (IF)
  
• Tamron AF 17-50mm f/2.8 SP XR Di II LD Aspherical [IF]
  
• Tamron AF 18-200mm f/3.5-6.3 Di II LD Asph.(IF) XR macro
  
• Tamron AF 18-250mm f/3.5-6.3 Di II LD Asph.[IF] macro
  

Hi Milt,

I was mostly considering primes in the list-there is definitely a dearth of high quality primes designed for the APS-C sensored bodies. I guess the message is clear-if you shoot APS-C, we want you framing away with zoom lenses. And many in the list you supplied are suspect in terms of image quality-Tony was talking about high image quality APS-C specific lenses in his video, to avoid using lenses designed for full frame.

anscochrome wrote in post #17928867
Hi Milt,

I was mostly considering primes in the list-there is definitely a dearth of high quality primes designed for the APS-C sensored bodies. I guess the message is clear-if you shoot APS-C, we want you framing away with zoom lenses. And many in the list you supplied are suspect in terms of image quality-Tony was talking about high image quality APS-C specific lenses in his video, to avoid using lenses designed for full frame.

No such message exists. Ef-s lenses exist to provide a wide angle option on crop cameras and a few inexpensive lightweight zooms. There is no reason to make ef-s primes lomger than 28 and there seems little demand for any wider.

anscochrome wrote in post #17928867
Hi Milt,

I was mostly considering primes in the list-there is definitely a dearth of high quality primes designed for the APS-C sensored bodies. I guess the message is clear-if you shoot APS-C, we want you framing away with zoom lenses. And many in the list you supplied are suspect in terms of image quality-Tony was talking about high image quality APS-C specific lenses in his video, to avoid using lenses designed for full frame.

Well if you are going to limit to fast primes, there just is the 30/1.4 in AF land, but Tokina now has the 14-20/2.0 in addition to the Sigma 18-35/1.8 and 50-100/1.8.

One thing about Tony's video that would have made more sense is if he had compared similar focal lengths. An 18-35 on crop is a different lens than a 24-70/2.8 on crop. If I had to choose between the two and I was going to be shooting portraits, I would pick the 24-70 because the focal length would be more useful for that even if it wasn't as sharp. Now that there is the 50-100, I would pick that but that wasn't an option when the video was made. I think he should have compared FF UWA zooms to crop standard zooms, IE 16-35/4 or 2.8 and 17-40/4 to the 18-35 or 17-50(55) type of lenses

Wilt wrote in post #17925622
OK, that is fine. He did state, as a generality, "Lenses designed for smaller sensors will give you sharper results because they're focusing all that light that is coming into the lens onto the sensor itself." not true.

And what about the absolute untruth in the statement that you have to multiply the f/stop as well, as it transmits less light?!

I don't agree with those statements at all and thus why I didn't mention that point when I responded to you. Still I wonder if Tony really meant to say those thing as he did. He seems to have a better grasp on these concepts if you watch a bunch of his videos. Makes me think he misspoke.