p.2 #1 · Crop sensor vs. full - reverse calculations?
This has been a most informative thread; some great posts. I think curious80 summed it up very nicely.
abqnmusa, the reach advantage is very real and the APS-C sensor noise is not an issue until the ISO is pushed beyond 1600. Under 'normal' light conditions I can get great shots with a $4000 combo (7d/100-400) rather than having to have a $18000 or $19000 combo (1dx and either a 300 2.8 with 2x TC or a 400 2.8 with 1.4x TC). Now before anybody gets too wound up, I would agree that the IQ on the latter would be better, but not markedly below 1600 ISO. This is the first significant occasion where the ISO got pushed too much and IQ dropped. It would also be nice to have the better dynamic range of the more expensive cameras, but it just isn't worth another $14k or $15k to me. I am an avid enthusiest, not a pro.
I suspect my path will involve patience; waiting for a 7d Mark II or whatever they call it to come out and waiting for a reasonable price on a used 300 2.8 (to go with a 1.4x and 2.0x TC) or a used 400 2.8 (to go with a 1.4x TC). This would normally give me 1 to 2 stops improvement on the camera and 1 to 2 stops improvement on the glass, or overally two to three stops improvement. As AJSJones said, with a 2.8, I would have had 1/1250 (1/800 to 1/1000 would have been fine) and with the next gen of APS-C, 3200 ISO will be pretty good (I could probably shoot 2500 based on the 1/800 to 1/1000).
So basically, start saving, watch for the next gen camera and a good deal on a lens. Patience. Patience. Patience.
p.2 #2 · Crop sensor vs. full - reverse calculations?
EverLearning wrote:
I feel I need to decide what I am doing with the body before I seriously consider what I want to do regarding longer lenses (which is another post). Input greatly appreciated!
Consider the reverse, 7D is very capable from what many have shown sharing the photos @ ISO 1600 and even 3200 assuming good coverage. Couple of folks in wildlife forum did a comparison and accounting for shooting conditions for wildlife (generally lower light and high shutter needed), 7D Still had a slight edge in effective resolution compared with 5DIII.
F5.6 lens is a bit too slow in tough conditions, consider a 300 f2.8 IS V1/V2 (+1.4x II/III & 2x II as options) before spending big $$s on body.
Dont get me wrong, I love my 5DIII, but I purchased the 300 2.8 before upgrading the body and if I was to have only 1 for wildlife, it would be lens.
p.2 #3 · Crop sensor vs. full - reverse calculations?
AJSJones wrote:
At 1/320 with f/5.6 you would have had 1/1250 ish with a f/2.8 lens. At that speed, with modest PP NR, you would have been happy - it's the glass you need, not the new camera, ... well ... you could go for both
Exactly. It's the glass.
I use the 100-400 for my dog sport stuff, and unless it is a sunny day, that lens is... well... nearly useless.
The problem is the 200-400 is just coming out of vapourware and also insanely expensive.
Maybe the solution is a 300/2.8 plus a couple of TCs (1.4 and a 2) if you're on a budget?
p.2 #4 · Crop sensor vs. full - reverse calculations?
300 f/2.8 + 1.4x buys you 1 stop of speed over 100-400L, but that's might not be enough. Either you'll want the 400 f/2.8 or a camera you're happy to use at higher ISO than you would on the 7D. Trouble is those cameras are reach limited FF or 1.3x crop. No cheap way out of the dilemma if you need reach and speed.
p.2 #5 · Crop sensor vs. full - reverse calculations?
Pixel Perfect wrote:
Hmm mixing unit's of pixels/inch and mm is not usually wise. You meant 232pixels/mm as your numbers are correct for the pixel sizes.
The calculation is even more trivial. 7D is a 1.61x crop sensor, so to get FF equivalent resolution it's 18*(1.61)^2 = 46.7MP.
I indeed meant 232 pixels/mm, but the original calculation (5568 x 8532) was indeed correct (in spite of the wrongly stated units)
BTW, my calculation method is more precise. If you multiply 14.9mm by 1.6, you fall short of getting same pixel density count across 24mm ...but 23.84mm; and 22.3 * 1.6 = 35.68, not 36mm!
p.2 #6 · Crop sensor vs. full - reverse calculations?
Lance Couture, I get what you are saying, but as somebody who has used the 100-400 for several years now, I would respectfully disagree with your statement that the lens is nearly useless unless it is sunny. I have taken many excellent shots with it in other conditions. But yes, grey skys along with non-slow moving object or a moving platform is definitely a challenge with a crop sensor camera and the associated ISO limitations.
After reading the various posts and upon further consideration of what I shoot, I am fairly confident that I need to stay with a crop sensor for now. Nothing is perfect and I will 'pay' for it in select situations, but those situations would be less (for me) than the suboptimal situations I would have with a FF sensor. I will just have to wait for Canon to come out with the next gen 1.6x APS-C camera.
p.2 #7 · Crop sensor vs. full - reverse calculations?
EverLearning wrote:
Lance Couture, I get what you are saying, but as somebody who has used the 100-400 for several years now, I would respectfully disagree with your statement that the lens is nearly useless unless it is sunny. I have taken many excellent shots with it in other conditions. But yes, grey skys along with non-slow moving object or a moving platform is definitely a challenge with a crop sensor camera and the associated ISO limitations.
I need a nominal shutter speed of 1/1250th for freezing my dog sport stuff. On a rainy day, mid-day, that means I'm usually at 1600-3200 ISO, depending on my location since the 100-400 is f5.6 at the long end.
It doenst matter if I'm using my 1D3 or my 7D, that lens under those conditions is kind of a dog (no pun intended). I both love that lens and am frustrated with it b/c it gives me the reach I need, but its range of applications due to its slow speed make it a very specific-use lens for me, hence my statement of "nearly useless".
I have looked at the Siggy 120-300/2.8 OS, but I just cant bring myself to buy one for both reasons of money and Sigma's auto-focus track record, despite people saying good things about it.
p.2 #8 · Crop sensor vs. full - reverse calculations?
wilt wrote:
BTW, my calculation method is more precise. If you multiply 14.9mm by 1.6, you fall short of getting same pixel density count across 24mm ...but 23.84mm; and 22.3 * 1.6 = 35.68, not 36mm!
p.2 #11 · Crop sensor vs. full - reverse calculations?
Pixel Perfect wrote:
But the 7D is a 1.61x crop camera not a 1.6x.
Monito wrote:
Actually, 1.61 is 1.6, but not 1.60.
curious80 wrote:
huh? 1.61 is 1.61. It is neither 1.60 nor 1.6
For scientific calculations (indeed, for all calculations), 1.61 (three significant figures) is 1.6 to two significant figures, but of course 1.61 is different from 1.60 (both have three significant figures).
1.6 is not 1.60 since the second decimal place is unknown when only one decimal place is stated. You can't start from 1.6 and go to 1.60, but you can start from 1.60 or 1.61 and go to 1.6.
Schools need to do a better job of fundamental education.
p.2 #12 · Crop sensor vs. full - reverse calculations?
In general going full frame will not help with low light when you need reach. This is because the fastest lens for a given angle of view will be about a stop slower with full frame than with APS-C. For example: say you're currently shooting with a 7D and 400mm f/2.8 if you upgrade to a 1DX you'll need a 600mm lens to provide the same field of view, the fastest one available is f/4, so the better noise performance of the 1DX is offset by the need to raise the ISO.
What you really need is a faster lens for your 7D. You might want to look at a used mark I 500 f/4.0L IS, a 500 f/4.5L or a 400 f/2.8L II (non IS) which should be available for about the cost of a 5D mkIII
In the meantime there are techniques to better manage noise. When you're shooting with a long lens the dynamic range you're trying to capture is often much lower than with a wideangle (as you have less background in shot, probably wont be capturing the sky etc), if your subject is in it's natural habitat where it is well camouflaged, the dynamic range can be very small indeed!
What this means is you can expose a long way to the right, you might be able to dial in +3 exposure compensation from what the camera's meter is saying and still not blow your highlights. Obviously you will need to shoot RAW and adjust all your exposures back down in post.
Expose to the right still works even when you're at your maximum shutter speed/aperture, as by raising ISO the signal is amplified before readout/digitization giving you a better signal to noise ratio.
p.2 #13 · Crop sensor vs. full - reverse calculations?
And now for something completely different:
From what I have seen, the Sony NEX series (and presumably the newer A series with the same APS-C sensors) has significantly better image quality than the Canon 18 MP APS-C sensor. I have borrowed both a 7D and a NEX-5N at various times, and the image quality of the NEX blew me away, much cleaner and more DR at comparable ISOs. This is not to say that you should switch to Sony, as pretty much every aspect (eronomics, AF performace, etc.) of the NEX except image quality was not up to the same standard as the 7D, so kind of a non-starter for action work (maybe some others, like the NEX-7 or higher end A series, are better?). Supossedly the new Olympus ILC has excellent IQ as well, and in a M4/3 (2x crop) format. In any case, there is some hope: better image quality in a high density sensor is out there, and we can have some hope that Canon will catch up at some point.
p.2 #14 · Crop sensor vs. full - reverse calculations?
Maybe a good combo is a Canon 1D IV with the mentioned Sigma 120-300mm. I have the Sigma 300mm with 2x and find IQ and focus both excellent.
The above combination with a 1.4x gives about 218-546mm f/4 equiv.
Plus you'll have the f2.8 option, retain zoom flexibility, and the higher ISO's capability of the 1-series. Also not as expensive as other options if bought second hand.
p.2 #15 · Crop sensor vs. full - reverse calculations?
Matt Grum, there are some interesting points in your post; some I think I may be misunderstanding and some I will disagree with. You do give up a stop by having to go to the bigger glass, but generally gain two stops regarding comparable levels of noise. So if I am shooting 1DX/600 4.0/6400 ISO I am still better off than shooting 7D/400 2.8/1600 ISO. Plus, if you step out of the confinements of only thinking Canon, I now have another option (which I have started doing serious analysis on) of the Nikon D800e. At 36MP, I can crop it to the same as the 1.6 APS-C 7D and have almost as many MP while still benefiting from better noise management, 2 stops better DR and better AF. And of course, if I don't need to crop the picture, I am WAY ahead of the 7D. If I was just starting to put a kit together, it would be a pretty easy decision. Canon needs to get its act together in this regard!
Dialing EV up +3 will push the shutter speed way down; demanding that the ISO be pushed way up. So perhaps I am misunderstanding your point, but this approach would seem to make a bad situation much worse. Feel free to elaborate if I have misunderstood.
Coffeyboy, the challenge in selecting lens is that Canon and Nikon have done a beautiful job of controlling or at least influencing 'credible' reviews out there and thereby creating FUD on 3rd party providers such as Sigma. Qualitative statements about Canon (or Nikon) lens x being better than Sigma lens y (or Tokina or ...) may be valid but how about quantitatively? 99% IS better than 98.9%, but how much should one pay for that difference? More importantly, good luck trying to determine what the real difference is!
p.2 #16 · Crop sensor vs. full - reverse calculations?
Monito wrote:
For scientific calculations (indeed, for all calculations), 1.61 (three significant figures) is 1.6 to two significant figures, but of course 1.61 is different from 1.60 (both have three significant figures).
1.6 is not 1.60 since the second decimal place is unknown when only one decimal place is stated. You can't start from 1.6 and go to 1.60, but you can start from 1.60 or 1.61 and go to 1.6.
Schools need to do a better job of fundamental education.
I understand the concept of significant digits very well and thats exactly why 1.61 is not 1.6. 1.6 is a number defined only up to 2 significant digits whereas 1.61 is a number defined to 3 significant digits. So they are not the same number. 1.61 is a more precisely known number and by calling it 1.6 you are throwing away the extra information that is known about the number.
If I take your logic forward then I can also say that 1.61 is also 2, because 1.61 is 2 up to one significant digits. So next time I could just say that 7D is a 2x camera
Anyways thats waaaay off topic. The Pixel Perfect's post just explained that if wilt uses the more precise figure for 7D's 'crop factor' then the calculations will match up, which is true.
Oct 18, 2012 at 02:20 PM
Steve Spencer Offline • • • • • • • Upload & Sell: Off
p.2 #17 · Crop sensor vs. full - reverse calculations?
Have you thought about the 1D MK IV. At 16 megapixels and APS-H size sensor, it give you a bit more reach than a 1DX or 5D MK3, but should give you a lot better high ISO than your 7D. It also shoots at 10 fps and has very good AF. It might split the difference nicely. It's cost is not much more than a 5D MK3 and substantially less than a 1DX.
p.2 #18 · Crop sensor vs. full - reverse calculations?
curious80 wrote:
1.61 is a more precisely known number and by calling it 1.6 you are throwing away the extra information that is known about the number.
That's my point.
curious80 wrote:
If I take your logic forward then I can also say that 1.61 is also 2, because 1.61 is 2 up to one significant digits. So next time I could just say that 7D is a 2x camera
That would be stupid to use a single digit of precision given the information widely available and because of the large error that would be introduced (MegaPixels). Nothing I wrote advised you to be stupid.
However, for the OP's purposes two digits are sufficient. Three (if you have them) provides a more accurate calculation and if somebody is worried about a few missing KiloPixels, then PixelPerfect's explanation is a good way to track them down.
When I wrote that you can go in one direction (by rounding) but not the other, the real illogic would be to think you must or should.
Monito wrote:
You can't start from 1.6 and go to 1.60, but you can start from 1.60 or 1.61 and go to 1.6.
p.2 #19 · Crop sensor vs. full - reverse calculations?
EverLearning, I was talking from a theoretical viewpoint with regards to format sizes, i.e. a lens 1.6x as long and just over one stop slower casts the same total light over a 35mm frame as the shorter one would over an APS-C frame. In other words the two stop advantage is more to do with particular sensor models than size alone.
Compare this with the wide end where max aperture gets worse as focal length decreases, here there are huge gains in light gathering to be had with full frame, compare the 24 f/1.4 with the fastest 15mm available.
Secondly, if you keep shutter speed and aperture constant and reduce the ISO you pictures will look worse (if you correct the brightness during raw conversion). Here is an experiment where I did exactly that:
So if reducing ISO for a fixed shutter speed and aperture makes the noise worse, then increasing ISO will make it better. A lot of people will disagree with this (increasing ISO has to make your images worse, right?) I won't argue the point, do your own experiment if you don't believe me (remember I'm talking about a fixed shutter speed and aperture, do you experiment in manual mode!)
p.2 #20 · Crop sensor vs. full - reverse calculations?
Matt Grum wrote:
Secondly, if you keep shutter speed and aperture constant and reduce the ISO you pictures will look worse (if you correct the brightness during raw conversion). Here is an experiment where I did exactly that:
So if reducing ISO for a fixed shutter speed and aperture makes the noise worse, then increasing ISO will make it better. A lot of people will disagree with this (increasing ISO has to make your images worse, right?) I won't argue the point, do your own experiment if you don't believe me (remember I'm talking about a fixed shutter speed and aperture, do you experiment in manual mode!)...Show more →
So what you've done is grossly underexposed the ISO 100 shot, then pushed it dramatically in PP to match the exposure of the ISO 1600 shot and it looks worse. Umm we already knew that. The ISO 100 shot will have so few tonal levels that pushing it will cause image to break down very quickly.
