Sensor Electronics

[kruegerk54]

I was in a discussion with someone where I'm not sure you can trust everything he says. I understand the major difference between crop and full frame sensors. Clearly the full frame sensor will cost more to make. But he also commented that it is not just sensor size but also the quality of the electronics that goes into sensors that drives the price higher. In other words he was claiming full frame sensor cameras generally have better electronics than crop sensors. For example will the current Canon 5D MK III camera (full frame at 24 mp) have better electronics than the 7D Mk II or 70D crop sensor at 20mp?

It seems the sensors are getting better as they evolve with time. I notice pictures from older cameras seem to have lower dynamic range than newer cameras, and others have told me they see less noise at higher ISOs when they have gone to a newer camera. I assume this means manufacturers keep improving on the electronics or the interpretation of the signals hitting the photosites of sensors. But my question is "Is there a clear difference in the electronics used for full frame vs crop sensors from cameras of the same generation to further account for the higher cost of full frame sensors?"

[xpatUSA]

But my question is "Is there a clear difference in the electronics used for full frame vs crop sensors from cameras of the same generation to further account for the higher cost of full frame sensors?"

Welcome, @kruegerk54, you got a name?

Both "better electronics" and "quality of electronics" are quite vague but you've certainly mentioned some of the usual variables. I'd say there is a difference but certainly not a clear one.

I'll stick my neck and draw on my past experience as an Electronic Controls Engineer . . .

FF is more expensive than 'crop' (horrible word) because there are less sensors per factory die and bigger chips on a die have a higher reject rate than smaller ones. An FF sensor theoretically runs a little hotter in the camera, making it less reliable than a 'crop' job, IMHO.

[dem]

The short answer is "no". The sensor quality per unit area, if you will, is pretty much the same across all manufactures and all camera formats from P&S to full frame made in a given year. Low light sensitivity of a 1.5 crop sensor increased only by about 1/2 stop in last five years (google for Tony Northrup videos for more analysis). So "the sensor electronics" is getting better with time but very slowly. The sensor size, not the quality, is the main difference here.

[Manfred M]

Sensor size is not the only reason that crop frames are generally less expensive than full-frame cameras. There are a number of other factors in play as well.

Full frame cameras tend to be used by commercial photographers and advanced amateur photographers, so they are designed for faster adjustments and shooting, generally without having to take one's eyes off the viewfinder. It's far more expensive to add extra buttons to a camera body than to slap things into a menu item. One of the advantages of the full frame camera is that they are physically bigger than crop frame cameras, so there is a lot more "real estate" on the body for camera controls.

Part of the reason for the larger size is that not only is the sensor larger, but that means the mirror, pentaprism (rather than the less expensive pentamirror used in enthusiast DSLRs). This means the cameras also need a larger mirror chamber, etc. To get the high speed burst mode operation, these mechanical components need to operate at higher speeds = larger forces in play which calls for a more robust design. Even the shutters in these cameras are much higher quality and built to last much longer than the consumer level cameras.

The other key cost driver in a camera is assembly costs. More expensive cameras sell less units than the ones aimed at the consumer market. This tends to mean that there is limited opportunity to use more cost effective production techniques. Full frame cameras tend to use die cast titanium bodies rather than injection moulded resin, found in consumer models. The same story can be found throughout the production cycle; higher costs due to small production quantities.

I remember that when Nikon announced the D4 and D800/D800E a few years ago, they had a monthly peak production capacity of 8000 units for the D4 and 20000 units for the D800 series. Assuming 20 working days a month and just under 7.5 productive hours a day, this means Nikon had a peak capacity of around 50 D4 cameras an hour and 130 D800 units per hour. These are not production volumes that justify any significant automation.

Like Ted, I'll stick my neck out. I'm a mechanical design engineer with extensive experience in product launches and production systems optimization.

[chauncey]

I was a nurse...production volume never entered the equation.

[Manfred M]

I was a nurse...production volume never entered the equation.

I suspect that this is not 100% correct. There is always some staffing model that "optimizes" the staff to patient ratio.

[inkista]

I was a nurse...production volume never entered the equation.

Isn't triage all about throughput efficiency? [ducks]

[William W]

I was in a discussion with someone where I'm not sure you can trust everything he says . . . he also commented that it is not just sensor size but also the quality of the electronics that goes into sensors that drives the price higher. . . .

Hi, Welcome.

That's an interesting statement that the 'someone' made.

It's understood that you came here seeking opinions on that statement, BUT, did you ask this 'someone' for first source documentation to back up the statement: and if you did not, can you do so now?

I'd appreciate you sharing if you receive an answer, thanks.

WW

[pnodrog]

As die size increases and transistor size decreases (45nm > 32nm etc) the production yield drops markedly. The yield is probably the most significant factor in the cost of microprocessors, memory and sensors etc.

The performance expectation is greater with FF cameras so the speed and resolution of the A/D converters is also higher and as a result the processor needs to cope with more data at a higher speed.

So better performance in one area forces better performance in other areas and performance costs. A bit like a car - up the engine size/power and you may need to upgrade the gearbox, provide better brakes etc etc. So it is not just a dearer engine that increases the cost.

If the benefits of the cost of development did not eventually spill down to the lower product ranges it is almost certain that the cost of a FF camera would be even higher.

[kruegerk54]

First I want to thank everyone for your input. You all have confirmed what I suspected. It is clear that a larger sensor will increase the cost of the camera for various reasons expounded here. But no one has suggested that the electronics around the sensor will give "better signal interpretation" in a ff vs crop sensor unlike what my contact was claiming. I take this all to mean my suppositions were correct - if there is any improvement in picture quality with a ff sensor it is due mainly to the increased size of the photosites (hence better signal/noise ratio) and not to the electronic translation of that signal.

Hi, Welcome.

That's an interesting statement that the 'someone' made.

It's understood that you came here seeking opinions on that statement, BUT, did you ask this 'someone' for first source documentation to back up the statement: and if you did not, can you do so now?

I'd appreciate you sharing if you receive an answer, thanks.

WW

Well Bill (William) if I see him again I can bring up this topic. He has said other things that I don't agree with and something about him rubs me wrong so I won't go out and search for him. I can tell also he is not very fond of me although we are courteous to each other at the second time we met (which was when he brought up this sensor technology topic).

[Manfred M]

I take this all to mean my suppositions were correct - if there is any improvement in picture quality with a ff sensor it is due mainly to the increased size of the photosites (hence better signal/noise ratio) and not to the electronic translation of that signal.

No, not necessarily true at all. Larger sensors can have the same size photosites as smaller ones. As an example so far as I understand, both the 1.5 x crop frame Nikon D7200 (24MP) and full-frame D810 (36 MP)have approximately the same pixel pitch. The full frame D4s (16MP) has far larger pixels than either of these other cameras, hence the better low light performance.

Part of the reason for higher image quality is that a larger sensor requires less magnification of the image to give you the same size output as a crop frame. Sensor design can impact image quality; look at the image quality improvements between two subsequent generations of cameras (not so much now, but in older cameras).

[inkista]

Because we are photographers and graphically oriented...

Here's different sensor sizes on the same wafers with the same flaws in the same places would look like (i.e., dust or something landing on the wafer during manufacturing), and the resulting yields.



The geometry of actual sensor sizes/yields without flaws is like this:

[rpcrowe]

I totally agree with Manfred when he stated:

Full frame cameras tend to be used by commercial photographers and advanced amateur photographers, so they are designed for faster adjustments and shooting, generally without having to take one's eyes off the viewfinder. It's far more expensive to add extra buttons to a camera body than to slap things into a menu item. One of the advantages of the full frame camera is that they are physically bigger than crop frame cameras, so there is a lot more "real estate" on the body for camera controls.

However, in the case of Canon Crop Cameras - there are two species of these animals...

The 7D and 7DII and ancestors such as 50D, and 40D have a pretty full set of button controls with information on shoting options appearing in the viewfinder. They also have the handy dual dial system to adjust your parameters. The 100% viewfinders also make it easier to view both the image you are shooting as well as your camera parameters...

The Rebel family (and perhaps some others) are primarily menu driven in selection of camera parameters.

I absolutely dislike the menu driven options and find it far easier to use button (or in the case of the higher end Canon cameras: button and dual dial driven selection).

[george013]

I totally agree with Manfred when he stated:

Full frame cameras tend to be used by commercial photographers and advanced amateur photographers, so they are designed for faster adjustments and shooting, generally without having to take one's eyes off the viewfinder. It's far more expensive to add extra buttons to a camera body than to slap things into a menu item. One of the advantages of the full frame camera is that they are physically bigger than crop frame cameras, so there is a lot more "real estate" on the body for camera controls.

However, in the case of Canon Crop Cameras - there are two species of these animals...

The 7D and 7DII and ancestors such as 50D, and 40D have a pretty full set of button controls with information on shoting options appearing in the viewfinder. They also have the handy dual dial system to adjust your parameters. The 100% viewfinders also make it easier to view both the image you are shooting as well as your camera parameters...

The Rebel family (and perhaps some others) are primarily menu driven in selection of camera parameters.

I absolutely dislike the menu driven options and find it far easier to use button (or in the case of the higher end Canon cameras: button and dual dial driven selection).

The Nikon D80 and D300 where also button driven. Both DX.

George

[William W]

. . . You all have confirmed what I suspected.

No, the commentaries in this thread have not so confirmed and that is a flawed conclusion. (Manfred beat me to it, but expanding on what Manfred wrote):

It is clear that a larger sensor will increase the cost of the camera for various reasons expounded here.

Correct .

*

But no one has suggested that the electronics around the sensor will give "better signal interpretation" in a ff vs crop sensor unlike what my contact was claiming.

Correct, no one has suggest that, but no one has proven or shown it to be NOT so in every case, either.

*

I take this all to mean my suppositions were correct - if there is any improvement in picture quality with a ff sensor it is due mainly to the increased size of the photosites (hence better signal/noise ratio) and not to the electronic translation of that signal.

Incorrect conclusion: as per Manfred’s comments and the expansion, above.

*

. . . if I see him again I can bring up this topic. He has said other things that I don't agree with and something about him rubs me wrong so I won't go out and search for him. I can tell also he is not very fond of me although we are courteous to each other at the second time we met (which was when he brought up this sensor technology topic).

If you do not like the company of this man then don’t seek him out on my account.

WW

[Manfred M]

The Nikon D80 and D300 where also button driven. Both DX.

George

D300 was a "prosumer" camera (I know a number of wedding photographers that shot them) and the D80 (and the D90 which replaced it) were considered "enthusiast" cameras and had the multiple button feature as well. So far as I know; this continues with the D7x00 series. The D3x00 and D5x00 are more menu driver.

[billtils]

OK, so now I am totally confused. If someone could please give me a (simple) answer to these 2 questions it should help:

1) I have a Nikon D7100 DX (APS-C) camera with a sensor that is 24.1 MP; I am contemplating buying a D610 FX (full frame). It has a 24.3 MP sensor - are these numbers so close by coincidence or is there a reason for it?

2) The D610 has an autocrop mode that will let me use my FX lenses (well, lens singular actually since my prime and longer telephoto zoom are FX ones) but crops to about 10MP; how will that 10MP on the FF sensor compare to the 24 on the FX when it comes to editing the image (including cropping it)?

Thanks!

[Manfred M]

Hi Bill - welcome to CiC

OK, so now I am totally confused. If someone could please give me a (simple) answer to these 2 questions it should help:

1) I have a Nikon D7100 DX (APS-C) camera with a sensor that is 24.1 MP; I am contemplating buying a D610 FX (full frame). It has a 24.3 MP sensor - are these numbers so close by coincidence or is there a reason for it?

That happens to be the size of sensor selected by the manufacturer to put into the cameras, based on what was available from sensor manufacturers. The actual decision was likely based on marketing studies on what would sell.

2) The D610 has an autocrop mode that will let me use my FX lenses (well, lens singular actually since my prime and longer telephoto zoom are FX ones) but crops to about 10MP; how will that 10MP on the FF sensor compare to the 24 on the FX when it comes to editing the image (including cropping it)?

Both will give your the same size of image, one will happen to have 24MP to work with (the D7100) and the other will have 10MP to work with. If you display images on your computer screen, you likely won't notice any difference as most screens today display at roughly 2MP, so both images would have to be downsampled. Even with some modest cropping, the differences should be negligible.

On the other hand if you go for printed output and do large prints; A3 or larger, you will be able to see a difference in print quality; especially when pixel peeping. The 24MP image will have higher resolution than the 10MP one.

[billtils]

Thanks Manfred

[Hallen]

I'm not and expert nor and engineer. I did remember reading some materials on the DX vs FX (crop vs full frame) sensors and I seemed to remember something about overall quality.

From the Nikon site:
"The FX sensor, with more "light gathering" area, offers higher sensitivity and, generally, lower noise."
Link to the Nikon site
Now obviously, that's a very vague statement and could mean many things.

From another site:
"Think of those pixels as buckets that attract light particles – the larger the bucket, the more light particles it can store in a given amount of time. These buckets are known as “photosites” and their size plays a huge role in sensor sensitivity and ability to accurately gather light in various lighting conditions. Bigger buckets are always better than smaller ones, because more light particles can be stored in those, without getting over-filled."
Photography Life
Again, this is an over-simplified statement.
The basic premise seems to be that the bigger the individual sensor (not the full sensor board, the individual sensors on that board) is for a given pixel produced, the more light will hit that individual sensor producing a higher quality result.

I'm not saying any of the above is right or wrong, especially from a pure technical aspect, but I think it is saying that bigger sensors have the capability of producing a better result in general. For example a 24MP DX vs a 24MP FX will have the same number of sensors on the board, but each sensor on the FX format will be larger and therefore will have more photons strike that area in a period of time as compared to the DX format. This should produce higher dynamic range and lower noise. At least, that's my very much a layman interpretation.