Sensor Format Blind Test

TL;DR

Rank the four shots below from best to worst. Mention the position as top-right, bottom-left, etc. Submit your ranking as a comment.

Sensor Format Blind Test
Rank the above images from best to worst. Mention positions as top-right, bottom-left, etc.

The Fuss

These are interesting times for camera enthusiasts. Digital cameras have advanced to a level where most cameras are capable of  publication quality snaps and you don’t even need to carry a camera to take good pictures, thanks to smartphones.

One big debate this has fuelled is regarding the importance of sensor size now. Some people believe, as they always have, that bigger is better and a full-frame sensor will always win.

Some people believe that beyond a certain threshold (1/1.7″? 1″? 4/3?) sensor size becomes irrelevant because the rules of equivalence (same DoF, same shutter speed, same FoV) dictate that all sensors will be operating in a similar fashion since they’re seeing the exact same amount of light, for the exact same amount of time.

Furthermore, it is argued that most pictures are never going to be printed and hung on walls. Instead, they will be viewed on phone, computer or TV screens at a nominal resolution of, currently, 2MP (HDTV resolution) or in future, 8MP (4k TV resolution). At such resolutions, minor differences in image quality become immaterial.

To find out how deep the shade of grey is between the black and white of FF supremacy vs. FF irrelevance, I’ve devised a simple test. The image you see above/below is a PNG screenshot of DPReview’s Studio Comparison Tool, loaded with camera test shots from cameras with different sensor sizes. I have chosen the web resolution from the tool to match the “common use” scenario, i.e., sharing pictures on the web.

Now, can you rank these shots from best to worst? Mention the position as top-right, bottom-left, etc. Submit your ranking as a comment below.

Sensor Format Blind Test
Rank the above images from best to worst. Mention positions as top-right, bottom-left, etc.

Once I have obtained enough responses, I’ll publish a summary and also reveal the link that will load this very comparison in DPReview’s Studio Comparison Tool, thereby revealing the cameras chosen.

Thanks in advance for participating!

Update: The poll is closed and the results are out!

  • Neo

    TR BL TL BR

  • Guy Kapuscinsky

    top left, bottom left top right bottom right

  • clive

    best -> top-left, bottom-left, top-right, bottom-right <- worst

  • damaru

    Decreasing order of perceived quality on a smartphone

    Bottom left – top left – top right – bottom right

  • Adnan

    Top right then top left then bottom left and finally bottom right

  • nazia

    top right top left bottom right botton left

  • Suhas

    While it is hard to find differences, i can say i don’t like bottom right (Suhas)

  • Sumeet

    1) Top-left
    2) Top-right
    3) Bottom-left
    4) Bottom-right

  • damaru

    With 25% web sized sample of playing cards?

    You don’t need a poll to tell that sensor size does not matter. But to answer a greater question, which is, is a larger sized sensor better than smaller sized sensor given certian other paramters, in situations that you shoot?

    you need to specify those paramters and no poll is going to answer that to your satisfaction.

    consider following scenario. I am shooting a family gathering in my living room. There are a large number of people in a small area. here are my requirements:

    a. To get a good subject isolation, i need a pretty shallow DOF.
    b. Since it is a small room, i need a large FOV.

    It is easy to satisfy the above requirements with a larger sensor system than a small sensor system. i say easy and because, on a smaller sensor system, it is theoretically possible get the desired effect with a very very fast glass but that would be very costly or too bulky.

    While the math you easily glossed over with “rules of equivalence” mey all work out, practicality dictates your choice. it is not just about pixels, noise, sharpness or DR.

    • With 25% web sized sample of playing cards?

      You don’t need a poll to tell that sensor size does not matter.

      If that’s what I believed, I wouldn’t have run the poll.

  • Mike

    Best to worst:
    upper-left
    upper-right
    lower-left
    lower-right
    (same for both images, I assume it was merely posted twice for convenience)

  • calvinscorner

    At web resolution most of these images look identical. If only one pays attention to minute details, the difference is visible. I’d rank images (best to worst) as follows
    1. Top left
    2. Bottom left
    3. Top right
    4. Bottom right