Understanding Full Frame

[Dave Throgmartin]

As prices become more affordable I've been interested in leaning more about full frame.

a) IQ

When I compare the Kodak Color charts in the 7D I'm taken back by how much better the 5D2 image quality is than the 7D +1/3 S=2 crops. The difference in resolution is astounding. The full frame ISO looks at least 1 stop better as well.

b) DOF

I came across one of Daniel Browning's posts on another forum and this is what he said:

100mm f/4 on APS-C has the same angle of view as 160mm f/6.4 on FF35.
100mm f/4 on APS-C has the same depth of field as 160mm f/6.4 on FF35.
100mm f/4 on APS-C has the same diffraction as 160mm f/6.4 on FF35.
100mm f/4 on APS-C has the same total amount of light as 160mm f/6.4 on FF35.
100mm f/4 on APS-C has the same noise as 160mm f/6.4 on FF35.

If I apply this logic and look at the equivalent of 24-105 f/4 crop equivalence when shot on full frame I get a 15-66 f/2.5. That not very fast on a crop lens is now surpassing the very highly regarded 17-55 f/2.8 in both speed and versatility while also getting better image quality! Also, a great prime like the 35mm f/1.4 L is now an equivalent 56mm f/2.2 lens. Still nice, but not nearly as nice as the real thing and probably equivalent to the much cheaper 50mm 1/4 now.

c) Conclusion

For those of you who are experienced with full frame cameras, especially those of you who have used both formats, are my thoughts above right?

Thanks
Dave

[andnowimbroke]

F4 might act like something around 2.8 in terms of shallow DoF, but the sensors for focusing still sees f4.

[HDNitehawk]

If I apply this logic and look at the equivalent of 24-105 f/4 crop equivalence when shot on full frame I get a 15-66 f/2.5. That not very fast on a crop lens is now surpassing the very highly regarded 17-55 f/2.8 in both speed and versatility while also getting better image quality! Also, a great prime like the 35mm f/1.4 L is now an equivalent 56mm f/2.2 lens. Still nice, but not nearly as nice as the real thing and probably equivalent to the much cheaper 50mm 1/4 now.

Most of what you said is correct. Daniel's comparisons were for a crop camera with a 100mm lens. I am not sure of your logic in your statement, are you saying that the 35mm F/1.4 on a crop frame vs a 56mm on a full frame?

The 35mm F/1.4 shot at F/1.4 on both cameras is going to have low IQ no matter if it is shot on a FF or a Crop Camera. It may have a deeper DOF and narrower FOV but the quality will still be lacking, until you tighten the aperture a bit.

[Dave Throgmartin]

I am not sure of your logic in your statement, are you saying that the 35mm F/1.4 on a crop frame vs a 56mm on a full frame?

Sorry, my intent was that the awesome 35L when used on a crop doesn't appear like it would be quite so awesome any more, given Daniel's work up it would now be:

35 x 1.6 = 56mm focal length
56mm focal length / 25mm pupil = f/2.2 aperture

compared to full frame. It appears the 50mm f/1.4 that is much cheaper would be able to provide the similar (maybe better) results on a full frame camera than what could be achieved on the crop with the 35L.

Dave

[neuroanatomist]

Basically right, except that as noted, the effect of sensor size on aperture applies to depth of field for the same framing, and *not* to exposure (since that's determined by light per unit area, independent of sensor size).

ISO noise in an image is dependent on total light gathered, which is determined by sensor size. Independent of sensor circuitry, FF has a 1.3-stop ISO advantage, meaning in terms of exposure you can more than make up for the stop between f/2.8 and f/4.

[ahab1372]

Sorry, my intent was that the awesome 35L when used on a crop doesn't appear like it would be quite so awesome any more, given Daniel's work up it would now be:

35 x 1.6 = 56mm focal length
56mm focal length / 25mm pupil = f/2.2 aperture

compared to full frame. It appears the 50mm f/1.4 that is much cheaper would be able to provide the similar (maybe better) results on a full frame camera than what could be achieved on the crop with the 35L.

Dave

I think your math is correct for focal length and aperture, but other attributes that affect IQ of a lens won't change (bokeh, CA, etc.)
I'm sure I will be corrected if I'm wrong

Arnt

[neuroanatomist]

I think your math is correct for focal length and aperture, but other attributes that affect IQ of a lens won't change (bokeh, CA, etc.)
I'm sure I will be corrected if I'm wrong

Some of those do change, generally for the worse on FF when comparing an EF lens on crop vs. FF. A crop sensor samples the 'sweet spot' of the FF image circle, meaning an EF lens on APS-C shows sharper corners, less lateral CA, less vignetting, and often better bokeh (remember - that's quality, not quantity; vignetting results in a cats-eye effect in bokeh, and less vignetting means that effect is mitigated).

Despite that, I think the 'overall IQ' (subjective) is better on FF for an EF lens.

[Dave Throgmartin]

The other thing I'm having difficulty figuring out with the 5D2 is how do you compose off-center photos. For example, let's say you want to take a rule of 3rds photo, how do you do it?

With my 60D it is easy, the AF points in the upper left / right and lower left / right are basically there. Change your focus spot, shoot at f/2.8 or whatever you want, and take the (usually) in focus picture. If you have to focus then recompose on the 5D2 at wide apertures that would seem to lead to a low keeper rate.

Dave

[btaylor]

You can still use the outer focus points at wide apertures on the 5D2. It's just that the cross type focus sensors are only active at f/2.8 and wider, but the outer focus points aren't cross type anyway.

If I'm shooting at 85mm and f/1.4 on my 5D2 for example, I just use the focus point closest to the area I want the subject in and adjust from there, rather than using a centre focus point and recomposing. Gets me pretty close.

[Colin500]

Basically right, except that as noted, the effect of sensor size on aperture applies to depth of field for the same framing, and *not* to exposure (since that's determined by light per unit area, independent of sensor size).

ISO noise in an image is dependent on total light gathered, which is determined by sensor size. Independent of sensor circuitry, FF has a 1.3-stop ISO advantage, meaning in terms of exposure you can more than make up for the stop between f/2.8 and f/4.

The way I usually combine these two issues into a single explanation: Exposure doesn't change, but ISO noise does, because "more amplification" is applied with the crop sensor to make up for the loss in light, only that this fact is hidden by ISO being normalised to give the same exposure regardless of the actual "amplification".

This is of course simplifying a lot of things, most of which somehow "average each other away" so that it describes the overall effect sufficiently well, at least for my purposes. And it can be done without even explicilty looking at the effective aperture change and explaining how and why applying the crop-factor to the aperture depends on which effect you're examining.