Reasons for wanting higher mega pixel count

Started 3 months ago | Discussions thread
Scottelly Forum Pro • Posts: 13,592
Re: Reasons for wanting higher mega pixel count

JimKasson wrote:

Scottelly wrote:

JimKasson wrote:

Scottelly wrote:

JimKasson wrote:

Scottelly wrote:

JimKasson wrote:

canonpfs wrote:

Much is said about the Sony A7R IV's "too high" mega pixel count. It seems the only reason that people can think of for needing a high mega pixel count is cropping. Besides cropping, here are two additional reasons for wanting more mega pixels:

1. Sharper images. Have you ever looked at an image on your rear LCD and thought that they were sharp only to find them blurry on your home monitor? Yes, I know we should look at our images on the LCD at higher magnifications but please bear with me for a moment. The reason why the blurry images looked sharp on the LCD is because a large image is displayed very small. So you shoot in low light and there is a breeze. You are already at a high ISO to get a higher shutter speed but the flowers in the foreground are still slightly out of focus. With lots of mega pixels available this image can be down sampled. At a smaller size the flowers are now sharp. Taking an image from 61MP down to say 18MP sharpens slightly out of focus areas.

That is true, but it's a second order effect if the CoC of the blurry flowers is much bigger than the pixel pitch. It occurs because the sharpening algorithms work better with more samples. By the time the CoC gets to 4 or 5 pixel pitches, the improvement is virtually gone.

2. Lower noise. Provided the 61MP sensor provides the same high ISO performance as the 42MP sensor the larger sensor will result in lower noise when down sampled. Noise just seems to magically disappear with down sampling. So if you have noise in the 61MP image just down sample it and the perceptible noise will be less.

This is true if you use nonlinear noise reduction. If you use linear noise reduction, it's pretty close to a wash, since the full well capacity for a given sensor technology is mainly proportional to the area of the pixel.

This is something I've been reading for years and years, yet the new 61 MP Sony sensor offers more dynamic range than a similar sized (full-frame) 42 MP Sony sensor. Since the technology is mature, and since both are back-lit sensors that have basically maxed out their possible fill-factor, shouldn't the 42 MP sensor offer more dynamic range, rather than less, if what you're saying is really true?

You are ignoring the words "for a given sensor technology" in the above.

I see, but what's the point in talking about full-well capacity limits, as they relate to are of the sensor and density, if the technology is constantly changing, and the ratio doesn't apply, since the sensor is new vs the old one you're comparing it to? This bothers me, because if things are not limited by physics, but only by the technology used, and the technology is not static, there seems to be no point in mentioning the full-well capacity and the number of pixels vs sensor area.

That is meant only to apply within a sensor generation, not across generations.

Then there's no point in mentioning this stuff, because the new sensor is not from the same generation, right?

I was offering general information. The FWC and the RN numbers for the a7RIV are as far as I know unavailable, so it is impossible for me to offer specific information. It appears that I have offended you. If you find the information that I've posted useless, please ignore it.

Sorry, but it just seems like a form of misinformation to me.

Misinformation: "false or inaccurate information, especially that which is deliberately intended to deceive"

In what sense is anything that I have written false?

Are you saying that I intend to deceive?

I've been mislead in the past to think things that weren't true, and as a result I ended up spending more money than I probably should have at the time. It didn't hurt me per say, but it bothers me when I see things that I think are misleading.

This concept of photo-sites with larger area being better for dynamic range bothered me when the Nikon D7100 came out, offering more dynamic range than any previous full-frame camera from Nikon or Canon, despite the photo-sites having a much smaller surface area. Nobody was ever able to explain to me why the little, APS-C sensor could capture so much more dynamic range than the older full-frame sensors. Sure, they said there was newer technology in the Nikon D7100 sensor, but that only goes so far, when people continue to make claims like this one you've just made. Given the fact that technology continues to improve, I don't see how what you're saying can be applicable, yet people keep saying stuff like, "They can't get much more out of a sensor, because the limits of the laws of physics have nearly been reached."

Have you ever seen me make that claim, except in the case of QE?

Well, you seem to be alluding to the same old claims that it's physics, but it's not, because it's technology, and that's constantly changing, isn't it?

Are you suggesting that QE's in excess of unity are on the horizon?

I'm saying that we are not reaching the limits of physics, but only the limits of today's technology, and that will change tomorrow or next week or next year, and for people to think there is some physical limit, because of the thing people suggest are happening, is a mistake.

I think a QE of greater than 1 is a limit, unless you have electron cascades, which doesn't really help.

Why is it not a limit?

People should not be mislead by things that people say that seem to be about physical limits, because it's not the physical limits that we are approaching. It's just the limits of today's technology, and next year or the year after there will be a better sensor that lets photographers do things they can't do with today's best sensor.

You qualify your statement above by mentioning the "given sensor technology" right? What's the point in mentioning all the other stuff, if you're going to basically negate what you said by "qualifying" it with an out phrase?

Again, please ignore what I've said if you find it inappropriate.

I'd rather not ignore things that I find inappropriate. "The only thing that evil needs to succeed is for good men to do nothing." Not that I'm saying what you wrote is evil, but if I don't agree with something, I think it's my duty to say so, and either find out why I am wrong, or convince others that what I think is right. There are limits, of course, but as I read through forums here, if I see something that bothers me, and I think I have something helpful or useful to say, then I normally say it, as I'm sure you do too.

I've read comments like that for more than ten years, yet camera/sensor companies keep on making sensors that have smaller photo-sites, yet can somehow capture more dynamic range.

What are your thoughts on this Jim?

It's unclear to me at this point how much Claff PDR improvement the a7RIV offers over the a7RIII.

Of course. Sony's claiming 15 stops. It will be interesting to see what Bill's calculations show.

It is true that full well capacities (FWCs) per square um tend to (fractionally) increase with each generation, and that read noise (RN) tends to decrease,

And that seems almost impossible,

Why do you say that? There is no physical limit to FWC that I know.

I was just commenting about the fact that full well capacities per square um tend to increase with each generation.

You said it was "almost impossible". Why should it be so?

Well, I said "that seems almost impossible." I mean how can we keep seeing big improvements in something that was near its maximum possible limits ten years ago. I'm thinking of an article in a well-respected website with the word Landscape in the name. Maybe I'm just remembering it wrong, but from the various things I've read over the years, I was lead to believe that back when full-frame sensors hit 12 MP and APS-C sensors were at 6 MP and 8 MP we were at about the limits of design, with regards to fill factor and the physics involved. That has turned out to be totally wrong.

Trees don't grow to the sky, but FWCs aren't like QE, where we're within a binary order of magnitude of theoretical.

What does it mean to be within a binary order of magnitude? Are you talking about 10% of theoretical limits? Are you suggesting we can go many times farther than we have come so far? If that's so, then why has it taken Sony so long to step up from the 24 MP sensor in the A900 to the 61 MP sensor in the A7r IV? The cameras are about the same price, but it's been about 15 years, and they've barely doubled the number of pixels their flagship sensor captures. To me that seems pathetic, but I realize there are a lot of hurdles to get over, and I also realize that image quality has improved in ways other than just resolution.

I'd think at some point they'd reach a maximum, but that just doesn't seem to happen. The sensor manufacturers just keep finding ways to make smaller photo-sites catch more energy. I guess it's like the efficiency of solar panels though. They've got a long way to go, though we were mislead to believe they were about at their maximum possible efficiency years ago.

I never saw any such claim about solar panels, and I have been reading about them in IEEE publications for years.

Well, since I don't read IEEE publications, maybe that's why I saw stuff that was inaccurate or just lies. It could be that I read something written by an anti-solar-energy Big Oil shill. It's possible, right?

when we read some of the things we read about limits of physics in sensor noise,

What things?

but improvemets keep on happening. It's amazing.

Within limits, it's predictable. Look at the curves for disk areal density vs time. Or transistors per unit area. There are gross effects that change those curves over a long enough time, but you can go a long way towards making many prections with a ruler and a piece of log paper.

so I would expect some PDR improvement. I would expect EDR -- unnormalized EDR, which is the only form I think makes sense -- to get somewhat better because of RN decreases, but again, at this point the is unknown, as Sony has not provided the protocol for their EDR claims.

Over time, as the pitches become finer, Claff PDR becomes more a measure of RN and less one of FWC, but we are a long way from the RN being the only important component, so, within a generation of sensor technology, I'd expect PDR to be about the same regardless of pitch.

O.K, but it's almost never within the same generation, right? This new Sony sensor is going to be compared to sensors from older generations, like the Nikon 45 MP sensor and the Sony 42 MP sensor and even the really old Canon 50 MP sensor. It might be compared to the new Fuji sensor too, but then the massive difference in size comes into play, and we'd be comparing apples with oranges.

All I can think is that this sensor is new and in a camera that's in the same price range, so it's probably significantly better. That seems to be the way it goes pretty much every time.

-- hide signature --

Scott Barton Kennelly

 Scottelly's gear list:Scottelly's gear list
Sony SLT-A65 Sigma SD1 Merrill Nikon D810 Sigma sd Quattro H Sony DT 18-55mm F3.5-5.6 SAM +20 more
Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow