APS-C vs FX

[dfhaskin]

Differences in image quality with enlargements made from an APS-C sensor with 24 megapixels (Sony Nex-7) and an FX sensor with 24 megapixels (Nikon D3x).  Will the smaller sensor (APS-C) produce better results?

[ErikKaffehr]

Hi,

Here are two studio shots with APS-C (Sony Alpha 55 SLT Sony 16-80/3.5-4.5 ZA) and "FX", Sony Alpha 900 24-70/2.8 ZA, the link has both JPEG and RAW (DNG) images:

http://echophoto.dnsalias.net/ekr/images/Demos/ApsVsFX/

An actual pixels crop (APS-C blown up to 6036 width) is shown below (left Alpha 900 right Alpha 55SLT)

Your mileage may vary. A smaller format (APS-C) will make larger demands on the lenses than a larger format. So a very good lens on an APS-C camera may outperform a mediocre lens on FX. Regarding the Sony NEX-7, the E-mount lenses are generally not very impressive.

I have both APS-C (Sony Alpha 700 and Alpha 55 SLT) and full frame (Sony Alpha 900), I can make A2-prints of excellent quality with both systems. On a recent shoot I had problems with subject motion caused by wind, so I shot with both my Sony Alpha 900 with it's 70-300/4.5-5.6G and my Sony Alpha 55 SLT with my Sony 24-70/2.8 ZS. On the APS-C camera I used larger aperture and higher ISO and actually got sharper pictures. See the enclosed screen dump. Sony Alpha 55 SLT on top, both images scaled to about 6000 pixels height.

Nikon is expected to announce an 36MP full frame camera (D800?) any time now.

To sum up:

- FX is probably better all other factors kept similar
- Sony NEX lenses may be less sharp
- On the other hand you can put anything on a Sony NEX
- Mirrorless will have no vibration from flipping mirror

Best regards
Erik

Differences in image quality with enlargements made from an APS-C sensor with 24 megapixels (Sony Nex-7) and an FX sensor with 24 megapixels (Nikon D3x).  Will the smaller sensor (APS-C) produce better results?

[allegretto]

Eric,

huge difference in moire patterns. What do you attribute that to?

[ErikKaffehr]

Hi,

Two factors. First is combination of pattern in the cloth with the pixel pitch. For some combinations moiré may arise, for others it may not. The main reason is probably that the Alpha 55 SLT has a weak or even non existent optical low pass (OLP) filter, also known as AA-filter. If the lens resolves higher than twice the pixel pitch aliasing will result. In most digital cameras a "Bayer  matrix" arrangement is used, having pixels representing different colors, so the aliasing will show up as color Moiré. The AA-filter would cut resolution so aliasing would be reduced or eliminated.

Stopping down the lens to f/11 would probably reduce aliasing, but also loose some sharpness due to diffraction.

Best regards
Erik

Eric,

huge difference in moire patterns. What do you attribute that to?

[allegretto]

Hi,

Two factors. First is combination of pattern in the cloth with the pixel pitch. For some combinations moiré may arise, for others it may not. The main reason is probably that the Alpha 55 SLT has a weak or even non existent optical low pass (OLP) filter, also known as AA-filter. If the lens resolves higher than twice the pixel pitch aliasing will result. In most digital cameras a "Bayer  matrix" arrangement is used, having pixels representing different colors, so the aliasing will show up as color Moiré. The AA-filter would cut resolution so aliasing would be reduced or eliminated.

Stopping down the lens to f/11 would probably reduce aliasing, but also loose some sharpness due to diffraction.

Best regards
Erik

Hey,

Yes. I just wondered if you felt it was more lens or camera. I thought the 900 had a very weak AA filter, but perhaps the A55 has near-none?

I can tell you that I just acquired the A77 and used exactly that lens (24-70Z) this weekend at a pumpkin patch shoot with the kids and was VERY impressed with the resolution and low moire. The kids were wearing somewhat heavier clothing that had a near corduroy texture. A perfect set up. Even at pixel-peeking enlargements the lines were straight and clear. I also like Sony's color balance right out of the camera with little need for LR color tuning.

My Leicas are a bit surreal, and my old Canon stuff... I'd rather not say what I thought of the color balance. But Sony seems very nicely rendered. For my taste anyway.

[bjanes]

Differences in image quality with enlargements made from an APS-C sensor with 24 megapixels (Sony Nex-7) and an FX sensor with 24 megapixels (Nikon D3x).  Will the smaller sensor (APS-C) produce better results?

The above scenario contains too many variables for proper analysis. A more meaningful comparison would involve the same brand of camera where the APS and full frame camera have the same megapixel count, as with the Nikon D3 and D300. If you frame the images so that the field of view is the same with both cameras and use the same lens at the same aperture, the D3 should give better results since it is less taxing on the lens.

Ken Rockwell gives an unusually lucid analysis (for him) using MTF plots. See the section "MTF Explanation: Diffraction and Pixel Pitch".

Regards,

Bill

[ErikKaffehr]

Bill,

Thanks for the link to the Ken Rockwell article. I wanted to write something similar on my own.

On the other hand, I have shot a table top image comparing my full frame and APS-C cameras and I'd say that I had a much smaller difference than what Ken Rockwell has seen.

I absolutely agree with your way to put it: "If you frame the images so that the field of view is the same with both cameras and use the same lens at the same aperture, the D3 should give better results since it is less taxing on the lens."

Best regards
Erik

The above scenario contains too many variables for proper analysis. A more meaningful comparison would involve the same brand of camera where the APS and full frame camera have the same megapixel count, as with the Nikon D3 and D300. If you frame the images so that the field of view is the same with both cameras and use the same lens at the same aperture, the D3 should give better results since it is less taxing on the lens.

Ken Rockwell gives an unusually lucid analysis (for him) using MTF plots. See the section "MTF Explanation: Diffraction and Pixel Pitch".

Regards,

Bill