What happens to the sensor when image size is set to medium or small?

[cfd]

I've just googled myself silly on this question ... is less of the image sensor used when I change the image size from large to medium / small?

If so what's the difference to crop factors?

Examples:

1. 24 - 36MP DSLR's

That's crazy in term of storage in the long run (how long are 4TB going to last me?), crazy in terms of Wi-Fi transmittion of images, crazy in terms of computer hard drive storage.

Still, I obviously want the new technology (higher ISO, faster processors, no OLPF, Wi-Fi, etc) ...

Question:
By setting the image size to medium (or small), do you only use a central part of the sensor - or - do pixels of the sensor get "merged" to produce larger individual pixels?

The 1.3 crop factor that Nikon advertise for the Nikon D7200 - in principle - isn't that just the same as changing the image size down from large?


2. 16MP COMPACT underwater camera

Considering how few underwater images of compact cameras actually turn out below 10m depth, 16MP is total overkill too (Nikon AW 130, 30m depth rated). Medium reduces the file size to 8MP - which probably isn't enough - but on balance, probably better than 16.

Question:
By setting the camera to medium 8MP image size, do I reduce a tiny 1/2.3-in image sensor (28mm2) to a miniscule image sensor (14mm2) - similar to that of a mobile phone???

In that case, wouldn't it be better to go for a smaller 12MP camera (i.e. Canon D30, 25m depth rated) and use the entire sensor area?

[mrinmoyvk]

Does actual size of image (in terms of pixels wide by pixel high) change? I never tried though. I always thought only the compression algo would become more complex to reduce the image quality and image size in bytes keeping the width and height same in terms of pixels.

Also I dont think the crop factor has any relation to width & height of image, it completely depends on Mega pixels of sensor. If crop factor had affected the width & height then it would have been digital zoom rather than optical zoom.
Image from 21Mp from mobile would be bigger than image from 16Mp from DSLR in terms of pixels but might be DSLR image is bigger in terms of bytes.

I might be completely wrong but this is what I think.

Lets see what experts say.

[Dave Humphries]

Hi Christina,

I'm in a rush at the moment, but quickly - you're confusing/combining two different things here.

The 1.3x mode on the D7100 and D7200 does reduce the number of pixels of the RAW (and jpg) images, it deliberately does so to give more 'reach' with Telephoto lenses. So you get 24MP or 15MP at 1.3x and these are 6000 x 4000 or 4800 x 3200 (from memory), by not using the entire sensor area.

I believe the "L, M, S" settings relate to jpg only and yes, I'm sure they do interpolate/average pixels to achieve this. The important point is, they will use the entire sensor area unless (I guess), you're in the 1.3 x mode and "S" or "M" in jpg, as well.

Hope that's right (as I'm rushing).

Welcome to the CiC forums from me, Dave

[Dave Humphries]

Hi again Christina,

OK, shopping and exercise done, not to mention a meal consumed, I now have more time to answer your plethora of questions properly.

I've just googled myself silly on this question ... is less of the image sensor used when I change the image size from large to medium / small?

No

If so what's the difference to crop factors?

Examples:

1. 24 - 36MP DSLR's
~
Question:
By setting the image size to medium (or small), do you only use a central part of the sensor - or - do pixels of the sensor get "merged" to produce larger individual pixels?

No - and - Yes - in that order

The 1.3 crop factor that Nikon advertise for the Nikon D7200 - in principle - isn't that just the same as changing the image size down from large?

No, because in this specific case it does only use part of the sensor area, for reasons outlined in my earlier reply.

2. 16MP COMPACT underwater camera
~
Question:
By setting the camera to medium 8MP image size, do I reduce a tiny 1/2.3-in image sensor (28mm2) to a miniscule image sensor (14mm2) - similar to that of a mobile phone???

No

In that case, wouldn't it be better to go for a smaller 12MP camera (i.e. Canon D30, 25m depth rated) and use the entire sensor area?

Not really relevant, because it doesn't work like that.

+++++++++++++++++++++++++++++++++

I hope that clears everything up and now you know we're here, you won't need to "google yourself silly" with photographic questions in future.

All the best, Dave

[rpcrowe]

Christina... If I shot JPEG (which I don't) I would always use the largest and least compressed JPEG. IMO, memory - both camera and computer - is really cheap these days and I see no reason for limiting my capabilities if I even want a large print.

[cfd]

Hi Christina,

I'm in a rush at the moment, but quickly - you're confusing/combining two different things here.

The 1.3x mode on the D7100 and D7200 does reduce the number of pixels of the RAW (and jpg) images, it deliberately does so to give more 'reach' with Telephoto lenses. So you get 24MP or 15MP at 1.3x and these are 6000 x 4000 or 4800 x 3200 (from memory), by not using the entire sensor area.

I believe the "L, M, S" settings relate to jpg only and yes, I'm sure they do interpolate/average pixels to achieve this. The important point is, they will use the entire sensor area unless (I guess), you're in the 1.3 x mode and "S" or "M" in jpg, as well.

Hope that's right (as I'm rushing).

Welcome to the CiC forums from me, Dave

Hi Dave,

Thank you so much for your clear answer on this (sometimes it's good to rush!).

Interpolation of pixels ... that's interesting - and very clever from a technical point of view. Shame it only applies to jpg. Nikon do have the new 'small RAW' option (D810, from memory), so it may be applied there too.

I shoot RAW whenever I can (cause I like the security of having post-shooting options) - so I'll just have to use the RAW compression option to reduce file sizes a bit - but that's fine, always worked well for me.

As far as the diving camera is concerned, I think I'll have to return it anyway - interpolation or not - as the Altimeter reads & records up to -59m on dry land ... in a town that's 20m above sea level. They had it for repair - and returned it just as bonkers (-41m now). You can imagine what they will tell me when it breaks under water (at a recorded depth of "-80m" or similar) as far as the warranty is concerned. Shame, the spec looked so good.

Thank you also for welcoming me to the forums (so kind) ... I've still a lot to learn on how to use it properly!

Christina

[lukaswerth]

Christina,
I will only give an answer the first part of your question, as I don't know about underwater photography at all:
From the top of my head, and this is what I always took for granted, there is a difference between cropping and saving small image sizes: with cropping only a part of the sensor should be used, and with smaller image sizes you have a quality loss, but the same image area. You can try this out in a few minutes by setting your camera to cropping, and then to some compression/jpeg quality. If you don't have the camera yet, enter a store, ask them to show you a camera, and try it out there, with the help of the salesman.

With regard to storage size on the computer, transmission rates and so on: I'd say another key to all this is to see what one wants from one's files, and if one needs the big files, then one has to weed out the directories, and keep only, well, keepers.
Easier to write than to do, I know, and I am myself often unable to follow my own advices, but this is the solution I see - and I want the full RAW files, jpegs just don't make sense to me for my expressive efforts.

Lukas

[Manfred M]

Interpolation of pixels ... that's interesting - and very clever from a technical point of view. Shame it only applies to jpg.

As raw files are technically manipulated output of what the camera has recorded, interpolating would remove the essence of RAW to begin with. As jpeg files are based on going from 12-bit to 14-bit data down to 8-bit jpeg, there is a lot of calculation work done on the image; interpolation for downsampling is actually a fairly straight forward process.

Nikon do have the new 'small RAW' option (D810, from memory), so it may be applied there too.

I don't know about the D810, but the D800 has three raw options; uncompressed (full-size); lossless compressed (slightly smaller and what I use) and lossy compressed (even smaller, but not that much versus lossless compressed). The other (small) advantage of compresses raw is that the download time from camera to computer is slightly faster than with the uncompressed version. This is what I use when shooting raw.

As far as the diving camera is concerned, I think I'll have to return it anyway - interpolation or not - as the Altimeter reads & records up to -59m on dry land ... in a town that's 20m above sea level. They had it for repair - and returned it just as bonkers (-41m now). You can imagine what they will tell me when it breaks under water (at a recorded depth of "-80m" or similar) as far as the warranty is concerned. Shame, the spec looked so good.

Perhaps the pressure calibration been messed up a touch at the camera repair and it should go back; but then I rather doubt that this function is going to work all that well on land, given that it is designed to read water pressure.

I've never used a camera when diving, so have no direct experience, but in the scheme of things the reading is "close". Standard air pressure at sea level is 101.325 kPa (1 atm), and you have to get to around 5500m for that to around 50 kPa (1/2 atm), assuming constant temperature. While diving, (as water is a lot more dense than air) the same kind of pressure differential would be at around -5m; as for every approximately -10m / -33ft you double the air pressure (2 atm), so I'd be quite surprised to see such a small difference on land give you such a high level of reading error (unless you are a tech diver, you're not going to get below -40m).

[george013]

Been reading this thread several times and had the feeling I didn't understand something. So downloaded the manual of the D7200. The areas are new to me. Honnestly, I don't see the advantage for it, except for selling.

Summarize. I'll stay with JPG for this.

The sensor is 24x16mm. Resolution 6000x4000.
1.
Sensor area.
You can select different areas you want to use. You can select full sensor or a 1.3 crop. This crop factor is not FF based. If FF based it would be 1.95. .

2. Diskfile.
To save the JPG you can change the resolution. What I understood is that the whole frame is recalculated for a resolution of 4496x300 or 2992x200 in DX, 3600x2400 or 2400x1600 in 1.3 crop.
The JPG can be saved in 3 qualities: fine, normal and basic given a compression ratio of 1:4, 1:8 or 1:16. The more compression, the less quality.

George

[Saorsa]

We seem to be getting into a lot of questions which are how processing is performed between the digital and analog worlds lately.

Once the physical sensor is manufactured the number of pixels it can resolve is fixed in a real physical device. Most of the formats and aspect ratios offered are there to accomodate post processing and are accomplished by only storing the desired pixels. In other words, pre-cropping.

If you are shooting video for a 16x9 screen, why keep irrelevant pixels that will need to be removed later? If you know you are going to have to crop your FX image because of FL length limitations of your kit why not use the DX mode?

[Manfred M]

If you know you are going to have to crop your FX image because of FL length limitations of your kit why not use the DX mode?

One word; flexibility, especially if I am in an unfamiliar shooting situation. If I have all the data, I can make an informed decision when I post-process the image. If I've already thrown away the data, I'm stuck and can't change my mind because the camera made those decisions for me.

[chauncey]

flexibility

IMHO, 'tis indeed a rare bird that can 100% properly compose that image in the viewfinder...
take all the information into your PP software and do it there.

[Manfred M]

IMHO, 'tis indeed a rare bird that can 100% properly compose that image in the viewfinder...
take all the information into your PP software and do it there.

That's a skill a lot of us old SLR film shooters had. It's a skill that transferred quite nicely to digital.

When shooting transparencies (slide film) you got full frame and no opportunity to crop, so you made darn sure that you had a level camera and cropped the scene in the viewfinder. You couldn't correct any of this in the wet darkroom, so you had to be 100% bang on with your shot. With the cost of slide film and processing, you had a minimal margin of error.

I do agree, it's a skill that few digital shooters seem to have, which is really too bad..

[Saorsa]

Nikon used to have a type T focusing screen for their SLRs that had an outline in TV screen format. Even with rounded corners.

[chauncey]

Aah...transparencies/film...what is that?

[xpatUSA]

That's a skill a lot of us old SLR film shooters had. It's a skill that transferred quite nicely to digital.

When shooting transparencies (slide film) you got full frame and no opportunity to crop, so you made darn sure that you had a level camera and cropped the scene in the viewfinder. You couldn't correct any of this in the wet darkroom, so you had to be 100% bang on with your shot. With the cost of slide film and processing, you had a minimal margin of error.

I do agree, it's a skill that few digital shooters seem to have, which is really too bad..

I find that the said skill is unfortunately forced on my good-self. I explain:

I like Sigma DSLRs. The first ones came with a low actual "resolution" of 3.4 MP i.e. 2268x1512px images. The next one along was 4.7MP, not a huge increase in image size, 2640x1760px. Since I view only on my 1280x1024px monitor, there is some cropping leg-room but not much. So, correct framing is quite important for me.

With the advent of the "Merrill" models, both the sensor size and the "resolution" increased significantly to true APS-C and 15MP. This resulted, not unnaturally, in huge (by my standards) file sizes of 50MB+ and a new need to down-sample every image to monitor size, coupled with waits of a minute or so between edits in Sigma's Photo Pro proprietary converter - too long for this impatient old man. I tried two "Merrill" cameras (SD1M and DP2M) and sold almost immediately for that very reason.

So it is that I stick with the 4.7MP, 1.7 crop, Sigma SD14 and am still paying due attention to framing. And, if I am very very careful, I can shoot in the Foveon low-res binned mode and and actually get slightly lower-noise 0.8MP raws to play with

[cfd]

Hi GrumyDiver,

Thank you for your comments. I've just written a reply to you and lost it all (grrr) so let's hope this one works:

I don't know about the D810, but the D800 has three raw options

The D810 has an additional 12 bit RAW file option "NEF ...small size available (12-bit uncompressed only)"
See http://www.europe-nikon.com/en_GB/pr...fessional/d810 >Tech Specs > Storage File Formats

Nikon were making a song and a dance about this option at a recent photography show I went to; it's ~ 10MP (from memory).

I've never used a camera when diving, so have no direct experience, but in the scheme of things the reading is "close". Standard air pressure at sea level is 101.325 kPa (1 atm), and you have to get to around 5500m for that to around 50 kPa (1/2 atm), assuming constant temperature. While diving, (as water is a lot more dense than air) the same kind of pressure differential would be at around -5m; as for every approximately -10m / -33ft you double the air pressure (2 atm), so I'd be quite surprised to see such a small difference on land give you such a high level of reading error (unless you are a tech diver, you're not going to get below -40m).

No, not a tech diver (no fun!) - just Advanced, so not supposed to go below 30m (oops).

Sure, all true; as you probably know, you can't qualify to dive without learning extensively about the pressure differences between air and water (and what it does to your system).

However, altitude is part of the GPS data recorded - and it works perfectly accurately in my mobile phone (iPhone) and in a Canon Compact camera I have (S100). It also worked perfectly fine in another camera I accidentally flooded at 10m depth (housing leaked), a Panasonic TZ10.

So if these 3 manufacturers (Apple, Canon & Panasonic) can get the altimeter to function correctly, I don't see why it shouldn't function properly in a Nikon AW130. Even adjusting from GPS (GLONASS, a fairly good Russian system; the Panasonic I had employed that too), my particular AW130 fluctuates wildly - and is out by up to 100m (!), including negative altitudes.

There are 2 possible explanations: either it's just my model that's crazy - or Nikon haven't got the GPS thing figured out in general.
Kind of makes you wonder why they haven't launched a DSLR with a built-in GPS since the D5300 - and there have been many launches since.

I don't know, maybe for some illogical reason, I just prefer Nikon DSLRs and Canon Compacts ... at least, they work, year in, year out, recognise a Macro in front of the lens, record accurate GPS readings (if they find the satellite, that is) and don't flood on you under water (my personal experience of 3 Canon compacts + housings since 2007).

Anyway, thank you for your comments - and btw, I have no idea how you manage to do your dive log without a camera ... by dive 3 of the day, I've forgotten what I saw on dive 1 (unless it was a Manta or a shark, haha).

Best wishes - and hope the quoting thing worked (and the site will process this reply) ...

Christina

[Manfred M]

Christina - altitude / depth measurement is done two different ways; above water, the GPS signal is used. Pressure guages would be susceptible to error from air pressure changes (listen to a weather report and you'll notice that this can change rapidly over time).

Below water (I have been told that GPS signals do not penetrate water particularly well), the camera has to use a pressure gauge, much like what you have on your regulator or dive computer.

[george013]

Christina - altitude / depth measurement is done two different ways; above water, the GPS signal is used. Pressure guages would be susceptible to error from air pressure changes (listen to a weather report and you'll notice that this can change rapidly over time).

Below water (I have been told that GPS signals do not penetrate water particularly well), the camera has to use a pressure gauge, much like what you have on your regulator or dive computer.

When I read that GPS has an accuracy of 20 meter for civilian purposes, then even if you can use GPS under water there will be a fault of 20m you have to calculate with. Quite a lot for diving.

George

[Manfred M]

When I read that GPS has an accuracy of 20 meter for civilian purposes, then even if you can use GPS under water there will be a fault of 20m you have to calculate with. Quite a lot for diving.

George

I believe the 20m is a "worst case" scenario and I find that in the places I use it accuracy / repeatability tends to be in the ±5m range. Altitude readings are remarkably accurate too; I expect that there must be some kind of link to topographic mapping data.

A former dive buddy is an electrical engineer who did a lot of work in electronic sensor development. He was working on an underwater navigation tracking system and is the one who told me that GPS signals did not penetrate water far. He also mentioned that GPS accuracy could be improved greatly if there was a ground station, at a known location, that transmitted a signal that could be used to correct for GPS satellite positioning errors.

Unfortunately, we lost touch so I have no idea if he ever got his prototype off the ground.