[Khukri]

Hello folks!

I have a question for those who are not only mathematically inclined but more so understand digital photography, sundry nuances of sensor efficacy etc as it relates to pixels and final picture quality out of a digital camera.

Background to question: Assume:
• You have a Canon 5D Mk II (full frame 21mp) and a Pentax K3 (APS-C 24mp).
• The sensor quality of both cameras is not the same. It remains what it is in real life.
• The 5D sensor size is 24x36 = 864 sq mm and the K3 sensor size is 16x24 = 432 sq mm.
• On both bodies you use a Sigma 50mm lens in respective mounts. The two lenses are electro-optically identical in terms of picture quality.
• No special in-camera processing or correction is applied.

Now, you shoot the a subject with both cameras with identical settings/lighting. What you get is:

• Canon – a 50mm perspective photo using 21mp on 864 sq mm surface area = 24,305 pixels per sq mm.
• Pentax

– The 1.5x crop factor gives you a 75mm perspective photo using 24mp on 432 sq mm surface area = 55,555 pixels per sq mm.

Next, you crop the Canon photo to equate the Pentax coverage of 75mm. In doing so, you delete portions of the Canon file from all sides to retain a portion such that the Canon photo matches the Pentax photo in angular coverage.
As I see it, you have now cropped the Canon photo from 864 mm size to 432mm size to achieve Pentax’s 75mm angle.
What you now end up with is effectively a Canon 432 sq mm size file of 10,500 pixels and a Pentax 432mm size file having 55,555 pixels.
You now make prints of both photos enlarging them to size 16x12 inches (400x300mm) using identical inputs for the printer.

The question – Keeping in mind difference in sensor quality of the two cameras in real life – pixel pitch/size favouring Canon and the file size favouring Pentax -

1. Which file will give a better quality image?
2. Is the Canon sensor so good as to make up the difference of 5x additional pixels in the Pentax file?

If anyone here has both cameras and can arrange to do a test and put up results, it would be revelatory!

Thanks in advance to all of you who can offer some informed inputs.

[Timd]

The question is why? Without looking into too much detail, this is an unfair comparison. The Canon is crippled by this comparison as you are not using the camera to its advantage. An alternative thought experiment would be to have a 50mm lens on the Pentax and a 75mm lens on the Canon and then look at the differences. You will, however, get the purists out to say while the FOV is the same, the DOF is different. To get it the same will need a faster lens on the Pentax...

My question is again: what is the purpose of this comparison?

[Rondec]

The only way the comparison works is if you are using equivalent lenses. That is to say, if you are shooting with a 55mm lens on your K3, that you assume you would choose an 85mm lens for your Canon. Obviously there are certain situations where you might not able to do that, for example if you have a 300mm lens for your K3, you may not be able to afford a 500mm lens for your Canon.

All of that said, the Canon will be a stop better with regard to noise at high iso. Dynamic range will be better (almost certainly) on the K3 file at lower isos. Everything else will probably be more dependent on the lenses you have/use.

[Alcazar]

Hi!

I guess you want to find out if you lose the longer reach a APSC camera gives you when you switch to a FF camera or if you would have the option of getting that reach by cropping your FF image by 1.5 in post?

To simplify, i only looked at sensor width.
Here's what i got:

Sensor width:
Canon: 5616 px on 35.8 mm
Pentax: 6016 px on 23.5 mm

Crop: When i use only 23.5mm of the Canon sensor, i have 3686 px left.

If i then print those pixels on 16 inches, i can print at:
Canon: 230 pixels per inch
Pentax: 376 pixels per inch

However, we're ignoring whether the lens actually delivers that resolution in both cases, and that the Canon has an AA-Filter
Greetings,
Micha

[rawr]

Hello folks!

I have a question for those who are not only mathematically inclined but more so understand digital photography, sundry nuances of sensor efficacy etc as it relates to pixels and final picture quality out of a digital camera.

I guess you haven't discovered DxOMark by DxO Labs - DxOMark.

[kenyee]

Is the Canon sensor so good as to make up the difference of 5x additional pixels in the Pentax file?

Depends on print size, but even at small print sizes, the Pentax file will have more detail IMHO.
This is the reason there are long argument threads on Nikon Cafe, etc. about whether APS-C or m4/3 gives you extra "reach". The answer is yes because of the pixel quantization you just realized. In the old film days (usually, it's the old timers who are arguing no), you stick it on an enlarger and shift it a bit and you have a magic enlargement w/ slightly less detail, but still a loss of detail.

That's why I always thought that most people would have APS-C (or m4/3) and FF....FF if they want the best night vision low light photography or narrow DOF, and APS-C/m4/3 for sports and the rest of the time when you want a bit more DOF and "reach" w/o dropping a lot of money on a 500mm lens...

[JinDesu]

The 5D Mk2 isn't the most amazing sensor out there. Given your scenario, the k-3 will win in detail, DR, noise performance, etc.

[normhead]

The 5D mk II isn't going to have the resolution (in lw/ph) of a k-3 on without cropping. Cropping the 5dMkII image simply makes it worse. Even given use of a 75mm lens on the Canon to match the 50 on the K-3, the Canon will probably be in second place for resolution. Noise,CA low light capability and other factors may apply. But for straight resolution, the K-3 should be better.

Although I'd really like someone who owns both or has access to both make the comparison. This kind of speculation is just hot air, until you see a couple of images side by side and decide you like one better than the other.

[stevebrot]

I have a friend who owns the 5D Mk II and when I can afford a K3, we can do a comparo.

Stay tuned...

[stevebrot]

I have a friend who owns the 5D Mk II and when I can afford a K3, we can do a comparo.

Stay tuned...

NOT!!!

[JinDesu]

NOT!!!

You've dashed our hopes so expertly.

[normhead]

NOT!!!

I been sucked in, and spit out…
Honestly though, I wasn't holding my breath in anticipation.

[Heie]

Hi!

I guess you want to find out if you lose the longer reach a APSC camera gives you when you switch to a FF camera or if you would have the option of getting that reach by cropping your FF image by 1.5 in post?

To simplify, i only looked at sensor width.
Here's what i got:

Sensor width:
Canon: 5616 px on 35.8 mm
Pentax: 6016 px on 23.5 mm

Crop: When i use only 23.5mm of the Canon sensor, i have 3686 px left.

If i then print those pixels on 16 inches, i can print at:
Canon: 230 pixels per inch
Pentax: 376 pixels per inch

However, we're ignoring whether the lens actually delivers that resolution in both cases, and that the Canon has an AA-Filter
Greetings,
Micha

This is the best explanation I've seen of comparing the effect of crop sensor and cropping FF to mirror the field of view. Thank you for not going into DOF/equivalence, as regarding the reach argument, it's obvious that a capable high resolution crop sensor (aka K-3) offers more reach than even the highest resolution FF (D800) because of the higher pixel density.

P.S. meine Frau ist aus Stuttgart und wir wohnen in Bamberg

[normhead]

However, we're ignoring whether the lens actually delivers that resolution in both cases, and that the Canon has an AA-Filter

The lenses deliver the resolution. In fact the resolution of both APS-c and FF cameras increase their lw/ph at about 70 line widths visible per 100 pixels. Add about a 100 lw advantage to the FF score for the picture and you're good.

So based on your example if the canon delivers 230/pixels per inch, you can expect 160 distinct lines visible per inch. If the Pentax is 376 pixels per inch you can expect..250. You would expect the same level of sharpness, if you printed the K-3 image 50% larger than the 5D Mk II image.

Another way to look at this is, a D800 crop image is 15 MP, that's not as big as a K-5 file. And a D800 is 50% larger than a 5D MkII. A K-3 is 50% bigger.

Anyway you look at it, if you crop your FF image to APS-c size, you can't even exceed a K-5 using an FF. A k-3 gives you 50% more than a K-5

[Khukri]

Thank you all for your comments/inputs.

In response to TIMD: Your question is answered aptly by ALCAZAR (thank you Micha!) so no point repeating it. RONDEC also touches on the main aspect of it:

I’ll explain briefly: I use a 400mm lens on the Canon 5DII for birds & butterflies. Often, cropping is necessary. Suppose I used a Pentax K3 or say a Sony NEX7 (both 24mp) with a Canon adapter (effectively framing 600mm equivalent on a FF Canon), I may not have to crop all that much and overall may end up with a higher resolution and - hopefully - a better quality picture, not so sure of the latter though.

As pointed out by Rondec, buying a Canon 600mm/f4 weighing in at maybe 6kg and using it hand held like I do the 2kg 400mm is a no-no. Even if one can afford the 600 mill, I wonder how many can afford the LBA that goes with it. (Low Back Ache, another ‘affliction’ of many dedicated nature photographers!)

Now to RAWR: DXO does not do the kind of comparisons I am looking for - as far as I can tell. If you disagree, a link to such a test/comparison by DXO is welcome!

NORMHEAD: I agree. My last sentence in my opening post says just that. Like you, I am waiting….! I have asked over at Sony too. The NEX7 and K3 both, use Sony 24mp sensors and so the end result should be comparable (?).

Till then, thanks again and the best to all of you!

PS: Although a long standing Canon guy in 35mm format, there was a time I also used to wield ( yep, I use that word correctly!) a Pentax 67 + 300mm lens hand held at 4000 meters in the high Himalayas. Those images…. ! I never thought I’d think so poorly of Canon glass! (Maybe it was just the film and format that made all the difference? Same as FF and APS-C now-a-days?)