Pages: [1] 2 3   Go Down

Author Topic: 5dsr vs hasselblad my first real world experience  (Read 32253 times)

orc73

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 318
5dsr vs hasselblad my first real world experience
« on: July 16, 2015, 04:28:47 pm »

Hi

I shot a few shots with the 5dsr during my latest shooting.
http://www.valentino-photography.com/comparison-medium-format-hasselblad-vs-canon-5dsr/
Maybe light and aperture as well as subject are not an ideal conparison.
It might still give you an idea what to expect.
(I had to use f13 for reasons of dof and background)

The most significant difference is between software converters rather the camera.
Hasselblad still has the edge even at 10mp less.

Best regards
Logged

ddolde

  • Guest
Re: 5dsr vs hasselblad my first real world experience
« Reply #1 on: July 16, 2015, 04:34:12 pm »

Not surprising 40 megapixel medium format beats the Canon
Logged

Gandalf

  • Full Member
  • ***
  • Offline Offline
  • Posts: 112
Re: 5dsr vs hasselblad my first real world experience
« Reply #2 on: July 17, 2015, 12:07:04 am »

Perhaps it's naive of me, but I expected the Canon to be on par with a H4D-40. I have always found C1 to be a little better than Lightroom, and I don't know if C1 got really good or Lightroom got really bad, but these results match my recent testing of the two converters. Of course, I still need Lightroom for DAM.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: 5dsr vs hasselblad my first real world experience
« Reply #3 on: July 17, 2015, 12:24:25 am »

Hi,

Capture One applies pretty extensive sharpening to images by default, while LR applies very little.

Try the enclosed setting, just as an example.

Keep in mind that you probably need to stop down 0.7 stop more on the Hassy for the same DoF. So, you would either use the Canon at f/10 or the Hassy at f/17.

You don't say which Hasselblad, but I assume you have a 1.3X crop factor.

Anyway the sensor on the Hasselblad is significantly larger than on the Canon, so if lenses are of equal quality the Hasselblad will come out on top, and I would assume that the Hasselblad 150 is a very good lens. Now, the Canon 135/2 is also a very good lens, but it probably has maximum performance at f/5.6, see Photozone results below:


This kind of testing is interesting as there are many ways to Nirvana on Earth, but they need to be properly made to be really conclusive:


  • Comparing two formats an equivalent aperture should be used. For 1.3X crop the fstop should be multiplied/divided with 1.3X.
  • Under studio conditions best practices should be used, like focusing with magnified LV and use of MLU and first electronic shutter curtain. What is best practice differs. For instance I always use LV based focusing and prerelease with landscape but it would not be appropriate for action.

It is not a surprise the Hasselblad comes out on top. It takes a lot of optical engineering to compensate for something like 1.7X the image area, stopping down beyond f/8 takes both lenses into the nearly diffraction limited territory. Both lenses will still be able to deliver adequate resolution, though, so much will depend on sharpening skills.

In a few weeks I hope to be able to make a similar comparison between my Hasselblad V-series/P45+ and a Sony Alpha 7rII I have on order. Really looking forward to it and I have little idea of the outcome.

Personally, I like Hasselblad, but I need a 3X ocular for optimum focusing. Also, I need both ultra wides and long telephoto, so I essentially carry double kits. Sometimes I just shoot the Hasselblad sometimes just the Sony. But, I am shooting targets of opportunity, so I cannot really foresee my needs.

Much of my shooting is on travel, requiring flights, most having 10 kg limit on cabin luggage. If I can get a 24x36 equipment fitting into that 10 kg limit matching the image quality of the P45+, that would solve many problems.

Best regards
Erik

Hi

I shot a few shots with the 5dsr during my latest shooting.
http://www.valentino-photography.com/comparison-medium-format-hasselblad-vs-canon-5dsr/
Maybe light and aperture as well as subject are not an ideal conparison.
It might still give you an idea what to expect.
(I had to use f13 for reasons of dof and background)

The most significant difference is between software converters rather the camera.
Hasselblad still has the edge even at 10mp less.

Best regards
« Last Edit: July 17, 2015, 01:06:15 am by ErikKaffehr »
Logged
Erik Kaffehr
 

BernardLanguillier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 13983
    • http://www.flickr.com/photos/bernardlanguillier/sets/
Re: 5dsr vs hasselblad my first real world experience
« Reply #4 on: July 17, 2015, 01:16:36 am »

The 5Ds images feels surprisingly soft at f13.

Cheers,
Bernard

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: 5dsr vs hasselblad my first real world experience
« Reply #5 on: July 17, 2015, 03:19:17 am »

The 5Ds images feels surprisingly soft at f13.

Hi Bernhard,

It's not really surprising (to me anyway), because the small (4.14 micron) sensel pitch samples the diffraction pattern induced blur much more accurately in much more detail than the 6.0 micron sensel pitch of the Hasselblad allows. It would take a much larger radius (and deconvolution) Capture sharpening to restore the amplitude of the original detail. By choosing the correct parameters, or better e.g. Focusmagic as deconvolution sharpening tool, the whole situation changes a lot.

The physical resolution limit for a 4.14 micron pitch sensor can be calculated. The aperture at which the diffraction will not only reduce contrast but actually limit resolution can be calculated as:
Aperture = (2 x senselpitch) / wavelength, e.g. (2 x 0.00414) / 0.000555 = f/14.9

So at f/13, for Green light, there is probably already a loss of resolution or at least a very low MTF, even more so if we add some residual lens aberrations and defocus to the system MTF. The DOF zone at the actual perfect focus plane is extremely narrow if we consider the COC to be equal to the sensel pitch.

For the Hasselblad, the physical (diffraction only) resolution limit for Green light would be at:
Aperture = (2 x senselpitch) / wavelength, e.g. (2 x 0.006) / 0.000555 = f/21.6

So at f/13 the Hasselblad would still have something like up to 20% MTF, and the Canon 5DS R only 3%, which would require very good deconvolution sharpening to restore some of the original micro-contrast and actually recorded resolution. Of course the Hasselblad will also be more sensitive to aliasing artifacts, which may make Capture sharpening a more difficult task, whereas on the Canon deconvolution Capture sharpening would be very effective and successful.

Cheers,
Bart
« Last Edit: July 17, 2015, 03:21:58 am by BartvanderWolf »
Logged
== If you do what you did, you'll get what you got. ==

Chris Livsey

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 807
Re: 5dsr vs hasselblad my first real world experience
« Reply #6 on: July 17, 2015, 03:29:10 am »


For the Hasselblad, the physical (diffraction only) resolution limit for Green light would be at:
Aperture = (2 x senselpitch) / wavelength, e.g. (2 x 0.006) / 0.000555 = f/21.6

Bart

I just shot some non commercial close work on the Hasselblad/P45+ using the 80mm HC and 13mm extension, so close but not very, I have shots at f32 that are bordering soft but sharpen up nicely in C1 alone, in expert post work hands I suspect they would be bordering very good. Purely anecdotal but under the circumstances where it is a good idea stopping down is still a valid option in MFD.
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: 5dsr vs hasselblad my first real world experience
« Reply #7 on: July 17, 2015, 04:46:38 am »

I just shot some non commercial close work on the Hasselblad/P45+ using the 80mm HC and 13mm extension, so close but not very, I have shots at f32 that are bordering soft but sharpen up nicely in C1 alone, in expert post work hands I suspect they would be bordering very good. Purely anecdotal but under the circumstances where it is a good idea stopping down is still a valid option in MFD.

That's right, although at a loss of actual resolution.

We know at which spatial frequency the diffraction pattern of a perfect circular aperture will reduce image contrast (MTF) to zero amplitude, e.g.  for 555nm wavelength and f/32:
cy/mm = 1 / (wavelength x aperture) , e.g. 1 / (0.000555 x 32) = 56.3 cy/mm

The P45+ back has a sensel pitch of 6.8 micron. We also know that the sensor array has a limiting resolution of maximum:
Nyquist frequency in cycles/mm = 0.5 / senselpitch, e.g. 0.5  / 0.0068 = 73.5 cy/mm

Therefore, at f/32, the optical signal has a diffraction limited maximum of 56.3 / 73.5 = 76.6% of the maximum resolution that the sensor could resolve. So the sensor is oversampling the blurred image signal, and would then be a good candidate for deconvolution Capture sharpening, without risk of aliasing artifacts caused by discrete sampling (which would only be possible if the signal resolution exceeds the sensor resolution). However, the lost resolution will not be recoverable, only the signal up to the 76.6% resolution limit will stand a chance of being restorable.

The Aperture at which resolution will be totally eliminated by diffraction, and will prevent all aliasing, by reducing contrast to 0% at the Nyquist frequency:
Aperture = (2 x senselpitch) / wavelength, e.g. (2 x 0.0068) / 0.000555 = f/24.5

So stopping down to narrower apertures than that will lose resolution on a P45+ back, but also eliminate the risk of sampling aliasing, and in absence of that make it a perfect candidate for deconvolution sharpening.

It's not rocket science, but it helps to do some calculations to see where the limits are. Not to avoid them, but to understand the consequences before shooting. Of course, if the DOF dictates such narrow apertures (and we cannot use focusstacking), then that's it, but we know what needs to be done to optimize the result and minimize the loss of resolution (deconvolution Capture sharpening).

Cheers,
Bart
« Last Edit: July 18, 2015, 11:26:42 am by BartvanderWolf »
Logged
== If you do what you did, you'll get what you got. ==

Chris Livsey

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 807
Re: 5dsr vs hasselblad my first real world experience
« Reply #8 on: July 17, 2015, 07:36:27 am »

Bart, Thanks for laying that out so clearly, I will print and save. I would if doing it for commercial have stacked focus probably as the easiest answer, but sometimes good enough is good enough.
Which package allows the best compromise between excellent result and ease of use in devolution sharpening? (I just used C1)
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: 5dsr vs hasselblad my first real world experience
« Reply #9 on: July 17, 2015, 09:20:44 am »

Bart, Thanks for laying that out so clearly, I will print and save. I would if doing it for commercial have stacked focus probably as the easiest answer, but sometimes good enough is good enough.
Which package allows the best compromise between excellent result and ease of use in devolution sharpening? (I just used C1)

Chris, I also use C1, but I only have the sharpening set for judging the image quality. Upon final saving, in the export recipe I have the checkbox for disable sharpening ticked. I then use Photoshop for Capture sharpening and those tonal adjustments that Capture One doesn't have, followed by resizing for the specific destination, adding a copyright notice, etc.

This allows to make use of the best tools for the different tasks, C1 for Raw conversion, Photoshop as a host for actions that combine repetitive steps and for plug-ins like FocusMagic for Capture sharpening, Topaz Clarity for local contrast, Topaz Detail to control Creative 'sharpening', and other tools for masking, replacing backgrounds, etc.

But you're right, sometimes good enough is good enough, and then one can do a lot with only the C1 controls, maybe on a clone of the image if additional control settings need to be saved but the original high quality rendering also needs to be preserved.

The main hurdle with Capture sharpening is that at different apertures, a different radius is required, and it may differ between lenses. Values can range from 0.7 at optimal aperture (often around f/4 or f/5.6) to e.g. 1.2 (at narrow apertures), but it also depends on the sensel pitch, On a large sensel a radius of 1.0 is covering a larger part of the diffraction pattern than a small sensel pitch on the same diffraction pattern. The latter might require a larger radius, but there is no simple rule of thumb.

FocusMagic makes it easy, because in the small preview it shows that the resolution improves as one increases the radius, and then it suddenly produces artifacts when taken too far. So the optimum can be found interactively, and that also takes care of defocus if needed (perhaps in a masked layer, only applied locally).

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

Bo Dez

  • Full Member
  • ***
  • Offline Offline
  • Posts: 112
Re: 5dsr vs hasselblad my first real world experience
« Reply #10 on: July 17, 2015, 10:15:45 am »

There really is not much to compare between 645 medium format digital and high res 35mm.

I think the Canon files are good in a small window of operation. The extra file size means you can do with more it but put the two side by side and medium format still kills it.

It's the fine detail, the much deeper colour and fidelity. The Medium Format looks so much more smooth and rich. Skin looks like it has 3 dimensional depth to it. On 35mm it has a biting sharpness without all the goodness that colour and tonality provide. The result is far more harsh and "anaemic"/flat about look at.
« Last Edit: July 17, 2015, 10:36:53 am by Bo Dez »
Logged

CATProductions

  • Newbie
  • *
  • Offline Offline
  • Posts: 1
Re: 5dsr vs hasselblad my first real world experience
« Reply #11 on: July 17, 2015, 11:43:40 am »

Thanks for the effort but I really don't think one can draw definitive conclusions based on that test. 
Logged

Hans Kruse

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2106
    • Hans Kruse Photography
Re: 5dsr vs hasselblad my first real world experience
« Reply #12 on: July 17, 2015, 01:45:31 pm »

Hi

I shot a few shots with the 5dsr during my latest shooting.
http://www.valentino-photography.com/comparison-medium-format-hasselblad-vs-canon-5dsr/
Maybe light and aperture as well as subject are not an ideal conparison.
It might still give you an idea what to expect.
(I had to use f13 for reasons of dof and background)

The most significant difference is between software converters rather the camera.
Hasselblad still has the edge even at 10mp less.

Best regards

The two shots does not have the same DOF, so if you shot the equivalent f-stop on the Canon I'm pretty sure that it would be easily as sharp as the Hasselblad. In that case the Canon would be more like f/8 than f/13.

Seeing studio shots from the dpreview scenes and comparing e.g. the 5DsR with the 645Z then the details from the 5DsR do not look inferior to the 645Z in my eyes.

Attached are 1:1 views done via Lightroom and the sharpening parameters a=60, r=0.7, d=70, m=20 in both cases. The first one is from the 5DsR and the second 645Z.

Joe Towner

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1365
Re: 5dsr vs hasselblad my first real world experience
« Reply #13 on: July 17, 2015, 04:45:39 pm »

I've got test shots of a 5DsR v H4D-50 in studio lighting of a rock with some amazing details (also shot with a 1Dx for reference).  The 5DsR is a nice camera, and a great system camera if you have the latest Canon glass, but it isn't a competitor for the medium format ability. 

DM me and I'll send you a link to the 2 raw files.

-Joe
Logged
t: @PNWMF

Ken R

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 849
Re: 5dsr vs hasselblad my first real world experience
« Reply #14 on: July 17, 2015, 05:32:39 pm »

The two shots does not have the same DOF, so if you shot the equivalent f-stop on the Canon I'm pretty sure that it would be easily as sharp as the Hasselblad. In that case the Canon would be more like f/8 than f/13.

Seeing studio shots from the dpreview scenes and comparing e.g. the 5DsR with the 645Z then the details from the 5DsR do not look inferior to the 645Z in my eyes.

Attached are 1:1 views done via Lightroom and the sharpening parameters a=60, r=0.7, d=70, m=20 in both cases. The first one is from the 5DsR and the second 645Z.

I agree, they are quite close. If anything the 645Z is showing more moire.
Logged

yashima

  • Full Member
  • ***
  • Offline Offline
  • Posts: 154
Re: 5dsr vs hasselblad my first real world experience
« Reply #15 on: July 17, 2015, 09:33:54 pm »

I dont see them as being quite close (Hans photos). The Pentax are noticeably crispier with more texture & definition (Canon file are a little bit darker though). For example, if you look at the King's hair, its very clear.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: 5dsr vs hasselblad my first real world experience
« Reply #16 on: July 17, 2015, 11:03:09 pm »

Hi,

I have downloaded two raws from DPReview. On left their Pentax 645Z image on the right their Canon 5DsR image with the Otus 85/1.4 at f/5.6.

Little doubt in my view is sharper, that said the both have a lot of moiré.

The Pentax 645Z was used with the Pentax 90mm f/2.8 D FA 645 Macro ED, at $4,496.95 (at B&H) while the Canon used the Otus 55 at $3,990.00 (at B&H)

Best regards
Erik
« Last Edit: July 19, 2015, 12:55:21 am by ErikKaffehr »
Logged
Erik Kaffehr
 

Hans Kruse

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2106
    • Hans Kruse Photography
Re: 5dsr vs hasselblad my first real world experience
« Reply #17 on: July 18, 2015, 04:49:12 am »

Hi,

I have downloaded two raws from DPReview. On left their Pentax 645Z image on the right their Canon 5DsR image with the Otus 55/1.4 at f/5.6.

Little doubt in my view is sharper, that said the both have a lot of moiré.

The Pentax 645Z was used with the Pentax 90mm f/2.8 D FA 645 Macro ED, at $4,496.95 (at B&H) while the Canon used the Otus 55 at $3,990.00 (at B&H)

Best regards
Erik


What can be a little deceiving when you look really close is that the 645Z is enlarged more than the 5DsR because of the way to scene is framed. The Canon has more pixels horizontally and the Pentax more pixels vertically.

I usually do not zoom in closer than 1:1 and on my MacBook retina screen they both look equally sharp and detailed. If I zoom in to 3:1 I do see more details in the 645Z however what kind of print would I see this (small) difference in?

What sharpening parameters did you use Erik?

synn

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1235
    • My fine art portfolio
Re: 5dsr vs hasselblad my first real world experience
« Reply #18 on: July 18, 2015, 04:51:34 am »

Thanks for the comparison, OP. This was pretty much my conclusion when I compared the 40MP Leaf back against the 36MP Nikon. I see that the same holds true even with a 50MP DSLR.
Logged
my portfolio: www.sandeepmurali.com

Hans Kruse

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2106
    • Hans Kruse Photography
Re: 5dsr vs hasselblad my first real world experience
« Reply #19 on: July 18, 2015, 05:34:07 am »

In the comparisons with MF I think it is useful to distinguish cropped MF versus FF MF. The 40MP Hasselblad in the comparison is a cropped sensor  with the dimensions 33x44mm. The Phase One CCD cameras IQx60 and IQx80 have sensor dimensions 40.4x53.7mm.

What this means is that for sensor size there is a relatively much larger size difference vertically than horizontally as shown in the attached calculation where the dimensions are shown relative to 35mm FF. So for the FF MF sensors there is a 1.7x linear size difference but horizontally for cropped MF is down to 1.2. So therefore in comparisons it is important how the shots are framed. Either aligned horizontally or vertically. Attached is also the pixel comparison between the 5DsR, 645Z and the Phase One IQ3 80MP sensors. Here you see that the 645Z has less resolution horizontally but more resolution vertically than the 5DsR (7%).

I think the numbers speak for themselves and differences comes as much down to lens differences and differences in focus and careful use of f-stops and post processing parameters.

Try and download the test shots of the same scene using the IQ160 and you will see a huge difference compared to the 645z and which is really visiible in 1:1 view. Again a print need to be quite large in order to show this is normal viewing distance. I use an Epson 3800 which can print up to Super A2 and comparing shots from Canon 5D III, Nikon D800E and Phase One IQ160 the difference in resolution was not visible to at normal viewing distance and very marginal at closeup inspection. So tiny differences viewing 1:1 in Lightroom are really that, tiny and marginal.

What cannot be taken away is the different feeling of shooting a MF camera and human beings are not rational in many cases and especially when a large sum of money has changed hands :) So we will continue to see comparison shots that shows what superiority of MF in cases where there really isn't. Technique, choice of lenses, post processing and photographic ingenuity of the photographers can mask (in my opinion) quite large differences.





Pages: [1] 2 3   Go Up