That's not what he's saying. You can make a sensor as oversized as you like. It will not help you at all with vignetting issues if your lenses are standard full-frame lenses.
Ok I don't understand what photonut is Trying to say. What has vignetting to do with square sensors? They wouldn't behave differently WRT vignetting than rectangular ones. Unless you override the default aspect ratios and really want the whole thing with black corners and all. My idea is that the square version should accommodate the present formats thus be at least long side squared. Even I would admit that nobody wants a 30x30 sensor for FF.
But I don't. I want to record 30x30 in this example. So I can use my old lenses.
True that some lenses have baffles. For this discussion I consider that a side issue. It's more about the principle. A more serious problem is that some mounts (such as Sony (as demonstrated by one poster) have the contacts intruding into the IC so no way they could use a square sensor.
How? My point is that the sensor should be enlarged to 6000x6000 in which case there's no penalty in terms of resolution in any aspect ratio. Photonut's post was rather confusing so might have misunderstood. Particularly the part about vignetting.
A square sensor that is covered by the current image circle of a FF lens would be just over 30mm x 30mm, not 36mm x 36mm. To be covered by an image circle, the diagonal of the sensor must be no greater than the diameter of the image circle. The diagonal of a 36mm x 24mm sensor is not quite 44mm, so we assume the image circle diameter to be about 44mm. A square sensor with the same diagonal is just over 30mm x 30mm. So if you made a 36mm x 36mm sensor, which has a diagonal of about 51mm, the corners of the sensor would extend 3,5mm beyond the outside of the image circle, and this wouls give you strong vignetting.
Well that's about the largest square sensor a current FF lens will reliably cover.
But there are penalties in terms of performance, contrast, veiling flare, complexity and cost.
Ok I guess I need to spell it out once more: the only square sensor that makes any sense is one that allows present aspect ratios without compromising resolution in addition to the benefits for more square aspects, thus must be at least 36x36 for FF. Suitable sizes for APS and M43 respectively would be 24x24 and 18x18 (rounded upwards). The parts of the square that fall outside the IC are discarded by default. It's called an oversized sensor. I believe Panasonic has an LX something with one though it's not square.
I use a m4/3 camera that can be set to sq format, 16x8, 2x3 or 3x4. When you frame the shot in the viewfinder, the viewfinder shows only that format, but the camera still records the whole sensor image of 4x3, it is presented for review as the whole 4x3 but framed with a white line depicting the chosen format. You can revert back to to 4x3 in PP, no need for a mask.
The issue is that if you were shooting square, and you masked the sensor to square, the camera would be a little less flare prone.
Yes but APS cameras are no more expensive than M43 cameras so can hardly be a problem. FF may be different as the price curve turns upward just beyond that. Which is why I advocate square sensors for M43 and APS, not FF. This thread just got off with FF and for many of us it's still some sort of reference so here we are. As a theoretical example it's as good as any other. It's a scalable thing, applicable to anything from 1/2.5" to ME.
That's a neat idea. However I think you exaggerate the issue. Canon lenses may have baffles but haven't seen any on my M43 lenses. Haven't noticed any particular problems with flare though a scientific study could show an advantage for Canon. Of course the same goes for hoods. I use mainly round ones because of stepup rings and sometimes it's even what the lens maker provides. Anyway, zooms render hoods rather ineffective much of the time anyway.
That is true in theory but again: compare M43 and APS. Can't see much of a problem. And consider this: a square sensor lets you utilize more of the lens' image circle every time you go from 3:2 or 4:3 toward squared ratios. And you gain AoV. So how I see it is that the advantages would heavily outweigh the costs.
Shouldn't be too hard. On Instagram square is still popular in spite of them allowing rectangular nowadays. So actually the real surprise is that phones don't get square sensors either. Weird.
So you want to put something like a 36MP 36mm x 36mm oversize sensor into a camera and use it to take 26MP square pictures, 25MP 4:3 pictures, 24MP 3:2 pictures and 22.5MP 16:9 pictures, using current FF lenses.
Precisely, except scaled down to M43 as that's what I have and can't afford a system upgrade.
Those are the restrictions + some mounts block the "view" with the contacts. Not M43 or Pentax as far as I can see.
The oversized sensor would probably cost more than a system upgrade. Certainly adding a NEX6 and a few adapters to give me APSC cost less than £200, even FF mirrorless systems can be had for under £500.
Looking at a MFT lens you might thing that but I've just had a quick look at one of my MFT bodies. The contacts would be in the way of stretching the sensor to be the same height as it's width. There's very little space between the lower corners of the sensor & the end contacts.
It becomes more of an issue when there are light sources in the part of the field of view that will be cropped out. You don't want the image of the light source bouncing off the mirror-like sensor if you aren't going to have it in your final image.
A square sensor only lets you utilize more of the lens' image circle if your final image will be square. If your final image will be in a different aspect ratio, you would have been better off with a sensor of the matching aspect ratio that was sized for the image circle.
Probably the best sensor size would be a compromise. Not square, but not as wide as the wider aspect ratios. Something in the middle, biased towards the more common aspect ratios. That would be the shape that makes the most efficient use of the image circle across a wide range of typical images.
Square peg … round hole. They won't fit perfectly no matter how you size them.
Obviously. Not a good idea.
Of course.
Now you got the idea.
Wouldn't make any sense.
IMHO wasting part of the sensor is mild poison compared to how much of my lenses' IC I lose because camera makers insist on sticking to old film formats.
That's an interesting idea. There are already 'wasteful' 'multi aspect' sensors (Panasonic GH2 and LX?). Wasting 33% of your sensor area may sound like much but they are different parts in different aspects. The always lost parts are small, just a bit off the corners. and the sensor size is still much smaller than the next format up which in the case of M43>APS doesn't seem to increase costs or size much if at all.
Here's the point you miss -- every aspect ratio except 1:1 loses with a 30x30mm sensor. Right now the 3:2 aspect ratio absolutely wins in terms of optimally utilizing sensor area for the biggest variety of alternative crops over 1:1 aspect ratio within the available image circle -- that's what my chart showed. Take some time to grasp that so it doesn't leave you confused.
Then the idea here is that everyone pays for unused sensor area, and the sensor is the most expensive component in the camera, so that you can get more area for a couple of barely used aspect ratios.
You should take a look at the edge performance of your lenses beyond what is captured on the sensor you are currently using them on. You can do this with an adapter and by putting them on a larger format sensor. I'm pretty sure you will be rather disappointed by what you see, I'm reasonably certain I would be.
Volume is a big part of cost, so if it's a low volume product it's going to cost more. Frankly, given the relative volume of m4/3 compared to APS-C/DX, and most of us don't know what the underlying costs or resulting profit margins are, I would be surprised if the smaller sensors don't actually cost more.
You could possibly do this for m4/3, this from wikipedia "The usual size of the sensor is 18 mm × 13.5 mm (22.5 mm diagonal), with an imaging area of 17.3 mm × 13.0 mm (21.63 mm diagonal)." if you use the imaging area you would reduce the frame to 15.3 X 15.3mm to cover the image circle of a m4/3 lens, You gain a small amount on the height and loose a small amount on the width of a m4/3 format, is it really worth it over simply using the square format already available in m4/3 cameras.
