I just finished comparing my old 1Ds MkII (mark two!) and the new 5D MkIII

[Kuja]

I just finished comparing my old 1Ds MkII (mark two!) and the new 5D MkIII.

16,7 old megapixels vs 22 new megapixels.

At base sensitivity of 100ASA 1Ds MkII won in detail resolving abilities!
Also, shadow noise in dark areas looks nicer in 1Ds MkII files.

I have compared tiffs made from RAW using Capture One 7.0.2 with default sharpening both on and off,
with all noise reduction/smoothing options turned off.

1Ds MkII files look crisper and with more fine detail.

Compared to them, 5D MkIII files look as if they were made with lower resolution camera than 1Ds MkII,
and then were artificially up sampled with Photoshop to a larger file size.

Here are two 100% crops...
Scroll them to the right to see those windows with white blinds:
real usable detail in 16.7mp 1Ds MkII files,
mushy moire in 22mp 5D MkIII files.

Again, both RAW files were developed with the EXACT same settings in Capture One,
test was done on a tripod, with 50mm lens set at f8:

I will post more crops and pulled shadows samples tomorrow, I must go to work now!

 

[distant.star]

.
Wow!

I thought I had made the right choice with the 5D3.

Looks like now I'll have to kill myself!!

 

[East Wind Photography]

To me it looks like the 5DIII image is slightly overexposed...or perhaps the 1dsII is a bit underexposed. Something is not right there and until you can get the washed out exposure corrected you cant make a good comparison on detail.

 

[Hillsilly]

Hi. Don't know how scientific your tests were. But as a 1Ds Mkii owner myself, I've yet to be convinved that newer cameras offer significantly improved image quality at ISO 100 (where I'm at for most of the time). The 5Diii might have better dynamic range, but I don't know how noticeable this would be in the real world. And of course, if you try to shoot at higher than ISO 3200....

Makes you wonder what Canon have been doing for the last 10 years.

 

[birdman]

I haven't been on here in a while, and honestly I went to the "other" brand. But I'm still a big fan of Canon.

Anyway, I'll say that for one, your test is not very scientific. I would use more lenses and scenarios to compare. If you compare a 5d classic RAW file to a 5d3 RAW at base ISO, it's likely the resolving power will be negligible at best. The major innovations in sensor technology happened years ago -- CMOS sensor.

What we see now are different algorithms in processing, as well as some *slight* changes in sensor design that allow for lower temperatures and other variables to achieve less noise. We're splitting hairs these days. What you do get in newer DSLRs are faster FPS, better ISO performance, better weather sealing, better LCD screens, etc. Don't expect too much -- except large drops in value as the newest model is released. Which is the reason I dumped my 5d2. the Mark 3 had nearly the same IQ, according to reviews. I didn't need FPS, AF points, etc. as a landscape guy.

I'll even go on the record as saying my D800 does not significantly out-resolve my former 5d2. The shadows are much cleaner, however, and RAW files are probably better...uhh....programmed? Outta my league there!! But the 14-24 (I don't own one) and 16-35VR (I do own one) are the reasons I switched from the 5d2 and 17-40L (which I used to own). And all of the newer, highly economical F/1.8 primes they have released.

 

[Kuja]

To me it looks like the 5DIII image is slightly overexposed...or perhaps the 1dsII is a bit underexposed. Something is not right there and until you can get the washed out exposure corrected you cant make a good comparison on detail.

You have used the correct word - "slightly".

There is a small difference in exposure due to slightly changing daylight between the two shots.

I tried to compensate, but the difference was less than 1/3 stop, so this was the best I could do.

Do you think that the difference in exposure of less than 1/3 stop,
can make one camera having lower detail resolving power than the other
and will it introduce moire instead of real detail?

 

[Kuja]

Anyway, I'll say that for one, your test is not very scientific. I would use more lenses and scenarios to compare.

I just wanted to test detail resolving power compared to my older camera,
since I was not 100% satisfied after trying using the 5DIII for the first time on a job.

I immediately noticed some "mushy" quality in 5DIII RAW files compared to my old 1DsII,
so I needed to confirm this.

Sunny afternoon, two cameras set on a sturdy tripod,
50mm prime lens set at f8, manual exposure, manual WB,
RAW files developed with the same settings...

How more scientific can I be in trying to figure out which camera gives better detail resolving power,
in real life use like, let's say, landscape photography?

 

[bdunbar79]

Anyway, I'll say that for one, your test is not very scientific. I would use more lenses and scenarios to compare.

I just wanted to test detail resolving power compared to my older camera,
since I was not 100% satisfied after trying using the 5DIII for the first time on a job.

I immediately noticed some "mushy" quality in 5DIII RAW files compared to my old 1DsII,
so I needed to confirm this.

Sunny afternoon, two cameras set on a sturdy tripod,
50mm prime lens set at f8, manual exposure, manual WB,
RAW files developed with the same settings...

How more scientific can I be in trying to figure out which camera gives better detail resolving power,
in real life use like, let's say, landscape photography?

Uhh, go do landscape photography with each??

 

[kbmelb]

The thing I noticed between the my 1dsII and the 5DII & III is, the 1DsII has less DR and in a sense makes the photo seem crisper since the image is natively more contrasty.

I find myself adding apx 5-10% contrast bump to the 5DIII's images to match my 1DsII's images.

I'll even go out and say the 5DIII is 2-5% less contrasty as the the 5DII.

 

[Aglet]

.. I'll even go on the record as saying my D800 does not significantly out-resolve my former 5d2. The shadows are much cleaner, however, and RAW files are probably better...uhh....programmed? Outta my league there!! But the 14-24 (I don't own one) and 16-35VR (I do own one) are the reasons I switched from the 5d2 and 17-40L (which I used to own). And all of the newer, highly economical F/1.8 primes they have released.

Off-topic here, how do you find the 16-35mm in the corners?.. I've been considering it for my D800 as the 17-35/2.8 is not up to my desires for landscapin.'
The 17-40L was not very good until stopped WAY down and even then it depended on focus distance.
As for 5d2 vs d800, I sold the former, bought 2 of the latter.

ON topic, I'm not surprised the old 1D series produces a slightly more pleasing image than the 5D3 but there should be some reasons for that. 5D3 should be producing slightly better raw files..

Then again, I sold all my newer Canon bodies and kept the old ones because they deliver nicer looking images. And then added bodies from the competition because they provide even better raw files.

 

[Zlatko]

Here are two 100% crops...
Scroll them to the right to see those windows with white blinds:
real usable detail in 16.7mp 1Ds MkII files,
mushy moire in 22mp 5D MkIII files.

If you down-res the the 5DIII photo to match the 16.7mp of the 1DsII, it will look sharper. Or enlarge the 1DsII image to match the 22mp of the 5DIII and it will look less sharp. The 5DIII is presenting a bigger image of the same scene, so it looks less sharp. If you had a 36mp photo of the same scene and viewed it at 100%, it would look even less sharp. Also the unfortunate moire effect in the 5DIII photo is usually a result of greater resolution, not less.

Both look extremely good — very, very similar. If I had to choose, the 5DIII photo looks slightly more natural to me. Extra sharpness in 100% crops can look overly digital and is not necessarily welcome.

 

[Zlatko]

The thing I noticed between the my 1dsII and the 5DII & III is, the 1DsII has less DR and in a sense makes the photo seem crisper since the image is natively more contrasty.

Oh yes, that's a good point. The camera with greater dynamic range will present a less contrasty image. The camera with less dynamic range will present an image that looks crisper overall. But greater dynamic range is usually preferred as you can always add contrast.

 

[ishdakuteb]

very very "smart and scientific" comparison... ;D

 

[Kuja]

How more scientific can I be in trying to figure out which camera gives better detail resolving power,
in real life use like, let's say, landscape photography?

Uhh, go do landscape photography with each??

I just did!

The examples above are crops from an urban landscape.

Landscape photography can be more than pics of mountains and forests.

 

[Kuja]

Here are two 100% crops...
Scroll them to the right to see those windows with white blinds:
real usable detail in 16.7mp 1Ds MkII files,
mushy moire in 22mp 5D MkIII files.

If you down-res the the 5DIII photo to match the 16.7mp of the 1DsII, it will look sharper. Or enlarge the 1DsII image to match the 22mp of the 5DIII and it will look less sharp. The 5DIII is presenting a bigger image of the same scene, so it looks less sharp. If you had a 36mp photo of the same scene and viewed it at 100%, it would look even less sharp. Also the unfortunate moire effect in the 5DIII photo is usually a result of greater resolution, not less.

This really makes no sense.

If I have tiny real details and texture in the white blinds on the windows in 1DsII sample
and I only have moire in the 5DIII sample,

why do you think that just by downsampling the 5DIII file to 1DsII file size, the detail that is not present in the 5DIII file will magically appear,
...or that just by upsampling the 1DsII file to 5DIII file size, the visible detail in the 1DsII file will magically disappear?

The resolved details are present in the files or they are not.

You can just try what you have suggested - use the crops that i have posted above,
resize them both way to match each other
and see if the lost details will appear in 5DIII file when you downsize it,
or if the fine resolved details will dissapear from 1Ds file by upsizing it.


Illusion of perceived sharpness achieved by stronger microcontrast is not the same as detail resolving ability.

1DsII files have both - more visible resolved detail and higher microcontrast.

Remember that RAW files from both cameras were developed with the EXACT same settings,
so you can not say that file from one camera was sharpened more.




...


I'm a professional photographer for 20 years.
I have been shooting with 1Ds MkII since the day it was introduced
and I know every bit of its possibilities very well.

After all this years I wanted to upgrade to a more modern body.
I was hoping for a high MP body but we got 1Dx and 5DIII.

I (naively?) decided to try the 5DIII since it has more resolution
and the 1Dx is not good investment for my needs (studio) if the big MP camera is around the corner.

The moment I started using the 5DIII in the studio (100 ASA, etc),
I realised that I don't really like its RAW files compared to my old 1DsII files.

When I opened first 5DIII RAW file in Capture One (confirmed later in DPP and Adobe CR),
I thought for a couple of seconds that I must have missed the focus a little bit.

But I did not.

I was just used to the crispness and details in 1DsII files that are just not there in 5DIII files.

I was disappointed since I wanted badly to like the 5DIII and to keep it for use in the studio.

 

[Kuja]

Shadow detail and noise as promised...

100% crops again, shadows are pulled in Capture One using HDR control, shadows option set to 100%.

Similar results this time, maybe I like noise in 1DsII files better,
or at least I'm not disappointed in it's performance since it is old technology.

And yes, check the fine detail in those branches compared to 5DMkIII files:

 

[Cannon Man]

This is great!!!! Everyone go buy a 1Ds II now =)

 

[Studio1930]

Forum rule #1. If you are going to do a comparison (no matter what it is) then you are going to get arguments from others unless you are a world famous person and then you will just get arguments behind your back.

Personally, when I buy gear that I don't like then I just sell it or return it. We all know that not everything is as good in real life as it is on paper.

For the record, I sold my 1Ds2 and 1Ds3 for newer bodies like the 1D4 and 1DX. To this day, my 1Ds2 produced more interesting images with my 85L than any other camera. Not scientific, but certainly noticeable in my eyes (and some of my clients').

 

[commercialshooter]

I shot with the 1DS2 for years and upgraded to the 5D3 so I have the same exact experience that you have with the bodies. A couple of things...

IF you only have to shoot at ISO 100 then your 1DS2 takes fantastic images. No doubt about it. IF you have to bump your ISO above 400 the images SUCK in comparison. At 1600 they are a complete mess. The 5D3 at 6400 is way better than the 1DS2 at 1600. There are other things that are hugely different as well.

Better screen on the back of the camera by far which in the real work world I need to see if I got the shot I need at a shoot for my clients.

Much better interface, The 1DS2 had the worst interface ever with the strange 2 button controls for everything.

Live view

Video

more frames per second

Better dynamic range

The list goes on and on.

The other thing is maybe you need to try a different program to edit your RAW files in. The 1DS2 had a VERY soft out of the camera file because of the high pass filter they put on it. The files sharpened up extremely well in PS but they NEEDED it badly. I would try to do more tests with a different software and also change up your lens. Maybe your lens is calibrated perfectly to your 1DS2 but off a little when mounted on your 5D3. The sensor is not "Sharper" on the older camera. Trust me.

The other thing that is quite noticeable in your test is that the 5D3 shots are exposed brighter in EVERY instance. The test would certainly be more accurate if you match exposure. The settings are irrelevant. Its the final image exposure that matters. Need to compare apples to apples.

At the end of the day if you are are only shooting studio work or landscapes at ISO 100 you would never see the detail difference anyway unless you are pixel peeping. PRINT the images out at 24x36 after making them both look the best they can. Don't limit the sharpening, contrast, etc of one to what you do to the other. That is not what you would do in the real world. You would optimize each image individually for it's intended output. Do that and then lets look and see if the 1DS2 is a better camera. I doubt it. I loved that camera for years and I do love the way the images look from it. That said the sensors are much improved since then and there is more to a camera than a sensor.

Happy shooting!

 

[vmk]

+1

I shot with the 1DS2 for years and upgraded to the 5D3 so I have the same exact experience that you have with the bodies. A couple of things...

IF you only have to shoot at ISO 100 then your 1DS2 takes fantastic images. No doubt about it. IF you have to bump your ISO above 400 the images SUCK in comparison. At 1600 they are a complete mess. The 5D3 at 6400 is way better than the 1DS2 at 1600. There are other things that are hugely different as well.

Better screen on the back of the camera by far which in the real work world I need to see if I got the shot I need at a shoot for my clients.

Much better interface, The 1DS2 had the worst interface ever with the strange 2 button controls for everything.

Live view

Video

more frames per second

Better dynamic range

The list goes on and on.

The other thing is maybe you need to try a different program to edit your RAW files in. The 1DS2 had a VERY soft out of the camera file because of the high pass filter they put on it. The files sharpened up extremely well in PS but they NEEDED it badly. I would try to do more tests with a different software and also change up your lens. Maybe your lens is calibrated perfectly to your 1DS2 but off a little when mounted on your 5D3. The sensor is not "Sharper" on the older camera. Trust me.

The other thing that is quite noticeable in your test is that the 5D3 shots are exposed brighter in EVERY instance. The test would certainly be more accurate if you match exposure. The settings are irrelevant. Its the final image exposure that matters. Need to compare apples to apples.

At the end of the day if you are are only shooting studio work or landscapes at ISO 100 you would never see the detail difference anyway unless you are pixel peeping. PRINT the images out at 24x36 after making them both look the best they can. Don't limit the sharpening, contrast, etc of one to what you do to the other. That is not what you would do in the real world. You would optimize each image individually for it's intended output. Do that and then lets look and see if the 1DS2 is a better camera. I doubt it. I loved that camera for years and I do love the way the images look from it. That said the sensors are much improved since then and there is more to a camera than a sensor.

Happy shooting!