Wrong comparisons in FF vs. Crop questions?

It seems like I see a lot of irrelevant comparisons in discussions trying to answer a poster's question about upgrading from crop to FF. It seems to me that the sensor size is just a massive feature difference. Very few people would deny that a BIG part of the difference between the image quality of say a G12 and 50D is the sensor size. So why so much focus on ISO noise (I guess partially related to sensor size), FPS, etc. The sensor size gap is significant!

imahawki wrote in post #11388278
It seems like I see a lot of irrelevant comparisons in discussions trying to answer a poster's question about upgrading from crop to FF. It seems to me that the sensor size is just a massive feature difference. Very few people would deny that a BIG part of the difference between the image quality of say a G12 and 50D is the sensor size. So why so much focus on ISO noise (I guess partially related to sensor size), FPS, etc. The sensor size gap is significant!

Sensor size is a very complex issue and many times is discussed without full detail of the true difference.

I think pixel density is an underdiscussed topic when it comes to sensor size discussion. My favorite example is the 1dmk2n vs 5d. Both have a pixel density of 122, which means they are essentially the same exact sensor, with the 1d being chopped to 1.3x crop. So at full size these two cameras should take images with the exact same IQ and noise characteristics. However, if you compress the two images to the same size, the larger 5d image will appear to have less noise. So ultimately the real advantage of FF is the ability to have higher resolution images (more megapixels) without increasing the pixel density.

When pixel density goes up, your resolving power and ability to capture detail goes up. However, higher pixel density means worse high ISO performance due to less signal (read: smaller pixels = less light) causing an decrease in signal/noise ratio.

Anyway, my point is that FF has some select advantages. But understanding those advantages involves a much deeper understanding of whats really going on with your sensor.

Good points. I'd like to dig into pixel density a bit more. I usually view pixel density from the negative aspect that you point out. But you also appear to be pointing out some benefits of higher pixel density. Can you elaborate more?

Here's some interesting stuff:
https://photography-on-the.net/forum/showthre​ad.php?t=706255
https://photography-on-the.net/forum/showthre​ad.php?t=747749

The advantage of higher pixel density? Are you seriously asking?

What;s the advantage of a 1D MkIV over a 1D classic?

Same sensor size, but increased Pixel Density gives the MkIV 18MP vs. the 1D 4.5 MP

I realize the feeling today is that more MP is getting to the point of "over crowding" but at the same time I think we all understand the advantage of 18MP vs. 4.5.. no?

Higher pixel density = more pixels on target = higher resolution, more detail, and believe it or not, better image quality. More detail even means lower perceived noise in your final image out put (PRint)

Also jd8817's diagram showing a photo-site 5X larger does not apply to a 1.6x vs. ff comparison. We are talking about fractions larger, not whole numbers. The diagram shows the advantage of moving from a P&S to any DSLR.

imahawki wrote in post #11388461
Good points. I'd like to dig into pixel density a bit more. I usually view pixel density from the negative aspect that you point out. But you also appear to be pointing out some benefits of higher pixel density. Can you elaborate more?

More pixels in the same area mean more detail, which I suspect is why the 1D line has an APS-H as its a good combination of high iso of full frame and tightly packed APS-C.

Think about it like a game of connect the dots. The more dots you connect, the more detailed the drawing.

CyberDyneSystems wrote in post #11388774
Higher pixel density = more pixels on target = higher resolution, more detail, and believe it or not, better image quality. More detail even means lower perceived noise in your final image out put (PRint)

I think this is the point Daniel Browning was making is his posts (linked above).

You can improve noise performance by either having bigger pixels or by having such high resolution that when compressed the noise approaches zero.

jd8817 wrote in post #11388801
I think this is the point Daniel Browning was making is his posts (linked above).

You can improve noise performance by either having bigger pixels or by having such high resolution that when compressed the noise approaches zero.

Indeed, we first started to talk about this phenomena a lot when the 1Ds MkII was released.
With it's denser pixels, yet similar era technology for noise processing, measurements showed the 8.5MP 1D MkII to MEASURE a given cropped area @ 100% with lower noise,.
Yet, in practice, the 1Ds II final total image, with it's additional resolution and detail always looked to have lower noise when viewed at real world viewing methods.

CyberDyneSystems wrote in post #11388820
Yet, in practice, the 1Ds II final total image, with it's additional resolution and detail always looked to have lower noise when viewed at real world viewing methods.

Aye. There's the rub. Many of these discussions have little or no interest in real-world viewing methods. Only in 100% crops.

pknight wrote in post #11389067
Aye. There's the rub. Many of these discussions have little or no interest in real-world viewing methods. Only in 100% crops.

True enough. But when I edit a shot I'm often looking at it at 100% or more and the differences are apparent there. Kind of hard to ask someone to settle for less once they've seen the difference. If you go only on real-world viewing we could probably all get G12's

gonzogolf wrote in post #11389097
True enough. But when I edit a shot I'm often looking at it at 100% or more and the differences are apparent there. Kind of hard to ask someone to settle for less once they've seen the difference. If you go only on real-world viewing we could probably all get G12's

Kinda like how you go to different restaurants, some better than others, and sit at the tables. They all look clean and you happily lay your hands and silverware on those tables. However one may be cleaned by conscientious attendees that use cleaning solutions and clean rags, the other just wiped down by the same water-dampened rag for the last 4 tables. Get a microscope out and see what you think. Even though tables at both places have the appearance of being really clean, bet you never frequent the one place again.

Well, if my photos were dining tables, and IQ differences at 100% were germs, I could see that argument. My photos are for looking at on the wall or screen. This is why I am involved in photography. Not to pixel peep. If I can't see the difference on the wall or on the screen, I'm not too eager to pay extra to produce the image.

'Pixel' density and sensor size are related by coincidence of design only. There is no physical requirement that a cropped sensor have a higher 'pixel' density, or that a FF body have bigger 'pixels'. These are entirely separate concepts and should not be confused.

The advantages offered by FF over crop are few, but important. To me, by far the greatest is the larger and brighter viewfinder possible.

Conversely, smaller sensors offer one major advantage: that lenses can be cheaper since a smaller image circle is required.

That's it. Decide if you want a brighter viewfinder or cheaper lenses.

Once that is done, the decision can be made about the trade off between 'pixel' size and density.

Bells and whistles aside, there may need to be just 4 classes of body: FF vs cropped x big 'pixel' vs high 'pixel' density.

The more I make prints, the more I can see that what I see on the screen at 100% does not relate to what's on paper very well at all. I've only been printing up to 19" and realize that if and when I want to go larger, it might be a big deal but for now, I cannot find a reason to change from the 1.3 crop of the MKIII with only 10MP. Intellectually I keep telling myself that I must go to a MKIV or IDSMKIII but in real life it has yet to be obvious with anything I've done. Doesn't stop me from wanting those cameras cause I do, and I wouldn't mind having a 5dII also because I think it should be better and is, but maybe not for any of what I am doing.

I recently had a shot amongst a whole bunch of photos that someone wanted a print, I thought at 800 ISO, 1/20th f2.8 on the MKIII that it probably wouldn't look that great in an 18" print, this was the result of looking at it on my computer at 100%. But I tried anyway, and did my best to process it well and sent it off to MPIX. I was pretty startled by how good the results were. My own printer (I was low on ink otherwise I would have printed it here) has also been showing me this a lot but usually I don't print larger than 8 1/2 x 11" so it's not as big an issue. My G11 prints looked really great at that size.