2
\$\begingroup\$

I have a camera which has the Sony IMX267LQR sensor. From the sensor documentation, its diagonal is 16.1 mm and pixel dimensions are 4096 (H) × 2160 (V). The camera manufacturer listed it as 1 inch sensor format. I therefore got a 1 inch compatible lens (model: Navitar NMV-8M1).

I couldn't find the horizontal and vertical dimensions of the sensor so I made the assumption that I can treat the image pixel dimensions and sensor size dimensions as similar triangles. Therefore I have one triangle with pixel dimensions 4096 (H) x 2160 (V) x 4630.63 (D) and the sensor with millimeter dimensions 14.25 (H) x 7.5 (V) x 16.1 (D). My Navitar's lens specification states the picture size as 12.8 x 9.6 mm. My questions are:

  1. Did I make a correct assumption regarding finding the sensor dimensions using similar triangles? I think I did because I tried the same method on the Sony IMX253 with known dimensions from Wikipedia and my numbers came very close.

  2. Did I buy the wrong lens then for my IMX267LQR sensor? My lens's picture size is listed as 12.8 mm but I'm getting the sensor size dimension as 14.25. Should I get a lens for 1.1" sensors then? For Navitar, their 1.1" lens lists picture size as 14.1 x 10.6 mm, which seems closer to 14.25.

  3. If I do get a larger lens, would it have any negative impact on my image quality? I know that the improper mating of a smaller lens format for a larger sensor will result in optical vignetting, but I don't know if there are any negative consequences of a larger lens format on a smaller sensor (other than price).

Improve this question
\$\endgroup\$

2 Answers 2

Reset to default
2
\$\begingroup\$
  1. Did I make a correct assumption regarding finding the sensor dimensions using similar triangles?

Yes. But you don't need to worry about them. All that matters is the sensor diagonal (16.1 mm in your case), because the lens projects a cone of light that lands on your sensor as an image circle. As long as the sensor diagonal fits within the lens's image circle, you're set.

In your case, it's very close. A 1" sensor is considered to have a 16.0 mm diagonal. So a lens that is intended for use with a 1" sensor would project that size of image circle, at a minimum. It probably projects just a little larger, to minimize the sharp vignetting at the corners of a standard 1" sensor.

  1. Did I buy the wrong lens then for my IMX267LQR sensor?

Personally, I don't think the extra 0.1 mm your sensor extends will be a problem. However, if you went with Navitar's 1.1" format 8mm lens, it states it projects an image circle of "⌀17.6", meaning the circle's diameter (⌀) is 17.6 mm. That would certainly cover your sensor, while also slightly reducing corner vignetting.

  1. If I do get a larger lens, would it have any negative impact on my image quality?

The 1.1" format lens states a resolution of 160 lp/mm (center) / 100 lp/mm (corner), whereas the 1" format lens states a resolution of 120 lp/mm / 80 lp/mm (center / corner, respectively). That means the 1.1" lens is a higher-resolving lens than the 1" format lens.

Generally speaking, there is no problem putting a smaller sensor behind a lens designed for a larger one. That is exactly what the whole "crop factor" discussion is about in the DSLR and mirrorless world. Lenses designed for so-called "full frame" (43.3 mm film/sensor diagonal) are mounted all the time on "crop" bodies with sensors that are 1.5–1.6 times smaller than the full frame sensor. Those crop bodies are just looking at a smaller portion of the image circle projected by the lens.

Improve this answer
\$\endgroup\$
7
  • \$\begingroup\$ Thank you @scottbb for the explanations, it was very helpful. I'm still a little confused on the answer for Q3. So if I take an image of a scene with the 1" lens and lets say the image is completely white with five black squares of 20x20 pixels each (4 squares on each corner, 1 square in the center) and I replace the lens with a 1.1" and take the picture again, would the five black squares still be in the same location and still be 20x20 pixels? \$\endgroup\$
    – tinker102
    Mar 24, 2021 at 7:31
  • \$\begingroup\$ @tinker102: Yes. The 1.1" lens is optically better: less vignetting than the 1" lens (if any), and in this case according to the specs a sharper image. Disadvantages of the 1.1" lens are possibly higher price, slightly more weight, slightly larger physical size. \$\endgroup\$
    – Roel Schroeven
    Mar 24, 2021 at 9:33
  • \$\begingroup\$ @tinker102 No. The lens projects its image on circle of a given size. The 1" sensor can't see the same extent as the 1.1" sensor. The image produced by the 1" sensor is literally a crop of the image produced by the 1.1" sensor. Ignore pixels on the sensor, I think it makes things more confusing (because you have to account for pixel pitch). If you captured an image with the 1.1" sensor, and then print it, and then crop about 9% around the edges (4.5% from each of the top, bottom, and both sides), what's left is what you would have captured with the 1" sensor. \$\endgroup\$
    – scottbb ♦
    Mar 24, 2021 at 17:04
  • \$\begingroup\$ @scottbb sorry I think I may have not worded my comment correctly. I meant keeping the sensor the same (1" sensor, or 61mm), but changing the lens. First image would be 1" sensor, 1" format lens. Second image would be 1" sensor, 1.1" format lens. I was wondering what would then happen to the black squares in the second image. \$\endgroup\$
    – tinker102
    Mar 24, 2021 at 18:47
  • \$\begingroup\$ @tinker102 I understood you, but I think I addressed the wrong thing. Sorry about that. =) You won't have any issue going to a lens with a larger image circle. Things should nominally be the same, as long as the lens's focal lengths are identical. The kicker there is that focal lengths are often nominal (i.e., roughly 8mm, but perhaps more like 7.9mm, or 8.2mm, or even further off). Quality-wise, you won't have an issue. But in terms of magnification, that depends on how precise the focal length is, and how similar the lenses are in optical design, and how similar they are at... \$\endgroup\$
    – scottbb ♦
    Mar 24, 2021 at 20:38
0
\$\begingroup\$

I recommend first calculating the active area of the sensor using the active pixels and pixel pitch. This is because CMOS sensor size format types are categories, not exact values.

Did I make a correct assumption regarding finding the sensor dimensions using similar triangles? I think I did because I tried the same method on the Sony IMX253 with known dimensions from Wikipedia and my numbers came very close.

My calculation shows that IMX267, running in 4096x2160 output mode with 3.45 micron pixel pitch, has an active area of 14.131mm x 7.452mm x 15.972mm. This is a true '1"' - if you are running in a larger output resolution, you may have a full diagonal of 16.1mm.

Did I buy the wrong lens then for my IMX267LQR sensor?

A general rule of thumb is that you will need a lens that has relative illumination of >50% at the full diagonal of the sensor, or a 16.0mm image circle. Otherwise, you may have black/dim corners when using the IMX267.

Unfortunately, Navitar has not provided the relative illumination chart for the NVM-8M1 on their website. So, you'll have to take their word for that the lens is designed for a 1" sensor. enter image description here

If I do get a larger lens, would it have any negative impact on my image quality?

The previous post about crop factor, using a larger format lens on the smaller format image sensor, is correct. The MTF / sharpness is dependent on the specific optical design of each lens - the relationship between image sensor format and contrast sharpness (geometrical square wave MTF) is not very strong.

Improve this answer
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.