Pages: 1 2 [3] 4 5 ... 7   Go Down

Author Topic: 1Ds2 versus 5D  (Read 70925 times)

Jack Flesher

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2592
    • www.getdpi.com
1Ds2 versus 5D
« Reply #40 on: July 06, 2006, 10:41:54 am »

Ray:

Okay, I'll do it.  

Go ahead and send me your 5D body and I will perform a controlled comparison using a very sharp 90 TSE and a more common 50/1.4 at various apertures.  I will test focus accuracy and dynamic range along with color-quality and resolution in a typical landscape shot. (It won't be a particularly interesting image -- it's summer here and landscapes are blah -- but I will endeavor to include significant elements for comparison.)  I'll show crops of the actual files and then try to quantify the print differences based on my observations along with those from a small group of third party viewers in a blind presentation.

I estimate I'll need your camera for about a week.  Email me directly at jbflesher-at-msn-dot-com so we can arrange date and shipping particulars.

Cheers,
Logged
Jack
[url=http://forum.getdpi.com/forum/

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
1Ds2 versus 5D
« Reply #41 on: July 06, 2006, 01:25:53 pm »

Quote
On the general point, I agree with you. I really miss Photodo, but even Photodo didn't take everything into consideration (bokeh, for instance, is completely missing, and vignetting isn't insignificant either).
[a href=\"index.php?act=findpost&pid=69903\"][{POST_SNAPBACK}][/a]

Yes, we could learn a thing or two by referring back to the Resolution, Contrast, and MTF article by Lars Kjellberg on Photodo, from which the following quote is drawn:

"We chose 10, 20 and 40 lp/mm. The most important is 10 lp/mm. This frequency makes up 57% of the total assessment, while 20 lp/mm counts for 29%, and 40 lp/mm 14%."

Thus far in this discussion we have been focusing on examining very fine details in the image, which would correspond to MTF at 40 lp/mm. However, perceived image sharpness is also related to contrast at lower frequencies, which is often more important in determining perceived image sharpness. The relative importances are reflected in the weightings used by Photodo. This aspect of image detail has been largely ignored in the current discussion.

The resolution at an MTF of 50% is probably the best measure of perceived image sharpness, and as I reported previously, it is very similar for the two cameras being discussed. For subjective analysis of sharpness using large prints, it would be necessary to evaluate images with a predominance of high spatial frequencies (such as a landscape or the detail in the engraving of the paper money) as well as images with a predominance of lower frequencies (such as a portrait or a close up of a flower). Michael's test image contains both of these distributions, but thus far everyone has been concentrating on the high frequencies.
Logged

Jack Flesher

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2592
    • www.getdpi.com
1Ds2 versus 5D
« Reply #42 on: July 06, 2006, 03:08:23 pm »

Quote
Yes, we could learn a thing or two by referring back to the Resolution, Contrast, and MTF article by Lars Kjellberg on Photodo, from which the following quote is drawn:

"We chose 10, 20 and 40 lp/mm. The most important is 10 lp/mm. This frequency makes up 57% of the total assessment, while 20 lp/mm counts for 29%, and 40 lp/mm 14%."

Thus far in this discussion we have been focusing on examining very fine details in the image, which would correspond to MTF at 40 lp/mm. However, perceived image sharpness is also related to contrast at lower frequencies, which is often more important in determining perceived image sharpness. The relative importances are reflected in the weightings used by Photodo. This aspect of image detail has been largely ignored in the current discussion.

[a href=\"index.php?act=findpost&pid=69931\"][{POST_SNAPBACK}][/a]

Bob:  I agree that at normal viewing distances 10 LPmm and 20 LPmm are more important than 40 LPmm.  But when viewing a larger print critically and looking at (or for) the high-frequency detail -- and that is what I've been saying all along -- the 40 LPmm performance becomes significantly more important.

It is precisely why Quentin's 8x10 negative printed at 40x50 looks so much better than a 4x5 negative (or anything else) printed at 40x50.  

If you don't print larger than 16" and don't look at your large prints up close, then 10 or 12 MP of resolution or medium format film is all you will ever need.  If on the other hand, one wants to see detail when looking at large prints, they have to up the ante.
« Last Edit: July 06, 2006, 03:09:35 pm by Jack Flesher »
Logged
Jack
[url=http://forum.getdpi.com/forum/

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
1Ds2 versus 5D
« Reply #43 on: July 06, 2006, 05:58:17 pm »

Quote
Okay, I'll do it. 

Go ahead and send me your 5D body and I will perform a controlled comparison using a very sharp 90 TSE [a href=\"index.php?act=findpost&pid=69920\"][{POST_SNAPBACK}][/a]

That's very noble of you, Jack, but I'm sure there must be lots of readers of this forum who happen to own a 5D and who live much closer to you than I. You might even consider buying one to avail yourself of that superior performance at MTF 50   .
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
1Ds2 versus 5D
« Reply #44 on: July 06, 2006, 06:16:46 pm »

Quote
"We chose 10, 20 and 40 lp/mm. The most important is 10 lp/mm. This frequency makes up 57% of the total assessment, while 20 lp/mm counts for 29%, and 40 lp/mm 14%."

[a href=\"index.php?act=findpost&pid=69931\"][{POST_SNAPBACK}][/a]


Bill,
As I understand it, these frequencies would have been considered more relevant in the days of 35mm film. 40 lp/mm on the negative, enlarged 8x to make an 8x10 print, results in 5 lp/mm on the print which is close to the limit of anyone's eyesight at close viewing distances, without a loupe. On a 16x24" print that 40 lp/mm on the film becomes just 2.5 lp/mm and on a 24x36" print, less than 2 lp/mm.

If people insist on inspecting large prints from the same distance as small prints (and I think Jack is one of those who does), then a 14% weighting for 40 lp/mm seems too little to me, especially when you consider the effects of USM which can 'bring out' detail but not create it.
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
1Ds2 versus 5D
« Reply #45 on: July 06, 2006, 09:42:38 pm »

Quote
Bill,
As I understand it, these frequencies would have been considered more relevant in the days of 35mm film. 40 lp/mm on the negative, enlarged 8x to make an 8x10 print, results in 5 lp/mm on the print which is close to the limit of anyone's eyesight at close viewing distances, without a loupe. On a 16x24" print that 40 lp/mm on the film becomes just 2.5 lp/mm and on a 24x36" print, less than 2 lp/mm.

If people insist on inspecting large prints from the same distance as small prints (and I think Jack is one of those who does), then a 14% weighting for 40 lp/mm seems too little to me, especially when you consider the effects of USM which can 'bring out' detail but not create it.
[a href=\"index.php?act=findpost&pid=69962\"][{POST_SNAPBACK}][/a]

Ray,

Actually, Photodo used the same criteria for medium format as 35mm, but I think that you are correct in assuming that they were developed for an 8 by 10 inch print viewed at 10 inches. The eye is assumed to resolve 6.7 lp/mm, so 55 lp/mm would be required on 35 mm film (or full frame sensor) to equal the resolution of the eye  with this sized print. That the MTF at 10 lp/mm is given more weight than that at 40 lp/mm tells us that contrast trumps resolution when it comes to perceived sharpness. In other words, rather than examining the magnification where image details are no longer distinguishable as in your large blow ups, you should back off and view the image at lower magnification so as to better evaluate the contrast at lower resolution.

Fujchrome Velvia 100 has a resolution of 55 lp/mm at MTF50, giving resolution/picture height of 1320 lp, compared to about 2100 lp/ph for the digital Canons under discussion, as I reported earlier on this thread. However, the film still resolves 100 lp/mm at MTF of about 25%. In view of these data, you can bump up the resolution weighting values for digital, but the same realationships still apply. I agree that the resolution cutoffs should be altered for digital. However, Canon still uses a maximum of 30 lp/mm in thier MTF charts.

Normally, a 16 by 24 print would be viewed at more than 10 inches, so that the needed resolution would not need to double. However, if you or Jack insist on viewing the print at 10 inches, then you would need to double the above resolutions to 20, 40, and 80 lp/mm for 16 by 20 and triple them for 24 by 36 inches to 30, 60, and 120 lp/mm. The point is that the most important resolution figure is well below the limiting resolution of the sensors, which is about 48 lp/mm for the 5D and 58 lp/mm for the EOS 1Ds2.

If you really want to preserve maximum sharpness apparent to the naked eye in an image larger than 16 by 20 inches  by the above criteria, you would have to go to a medium format back or a scanning back such as the Better Light.
« Last Edit: July 07, 2006, 08:47:33 am by bjanes »
Logged

Jack Flesher

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2592
    • www.getdpi.com
1Ds2 versus 5D
« Reply #46 on: July 07, 2006, 01:50:19 am »

Quote
That's very noble of you, Jack, but I'm sure there must be lots of readers of this forum who happen to own a 5D and who live much closer to you than I. You might even consider buying one to avail yourself of that superior performance at MTF 50   .
[a href=\"index.php?act=findpost&pid=69958\"][{POST_SNAPBACK}][/a]

Ray:

First off, I've bought one, tried it, didn't like it and sold it.  So there is no way I am going to buy another just to appease your curiosity -- I already know there is no superior performance there to avail fo me!

Second, of the four other guys I regularly shoot with who have owned a 5D, ALL of them have sold theirs too, so I don't have one close to by to do the test with.

So until somebody sends me one or joins me for a shoot with one, it sounds like this argument is over  The good news is you seem quite happy with yours so it has obviously been a good decision for you.
Logged
Jack
[url=http://forum.getdpi.com/forum/

Jack Flesher

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2592
    • www.getdpi.com
1Ds2 versus 5D
« Reply #47 on: July 07, 2006, 02:04:20 am »

Quote
The point is that the most important resolution figure is well below the limiting resolution of the sensors, which is about 96 lp/mm for the 5D and 117 lp/mm for the EOS 1Ds2.

[a href=\"index.php?act=findpost&pid=69983\"][{POST_SNAPBACK}][/a]

Bob:

I think you will find those figures are the maximum linear resolutions for the sensors in question in "lines per mm" and not "line pairs per mm"...  If you halve your numbers, you essentially get the maximum Nyquist Line Pair resolution.  In my testing the numbers are pretty close too -- I get a maximum resolution on my particular 1Ds2 of about 54 LP/mm.  A bit lower than Nyquist due to the AA filter I suspect.

~~~

Where the discussion gets interesting for larger prints is how well the digital file interpolates and still maintains detail in the 20 to 40 LP/mm range.  With proper workflow, my experience shows the 1Ds2 interpolates very well (and FWIW so does the P45 -- extremely well in its case). And for that matter so does the 1D2.  But for whatever reason, my particular 5D did not interpolate as well as my 1D2 with same pixel pitch -- an anomalie I am at a loss to explain.

Cheers,
« Last Edit: July 07, 2006, 02:11:58 am by Jack Flesher »
Logged
Jack
[url=http://forum.getdpi.com/forum/

dwdallam

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2044
    • http://www.dwdallam.com
1Ds2 versus 5D
« Reply #48 on: July 07, 2006, 02:44:43 am »

Actually, except for a few very technically interested people, which is indeed extremely important information, what matters in the end is how the printed image looks at a respectable viewing distance. You print at 20x30 and you're not going to be looking at it from 10". At 20x30, probably more like 4+ feet. At 12x18 probably 1.5-2 feet at closest.

So tell us, how will they look when viewed with the same image processed and printed at the same rez? If I had a IdsII I could do the test myself. But I only have the 20D and 5D. It would be interesting to me to see a landscape printed image at 12x18 and 20x30 from the 20D and 5D side by side, and the 1Ds.

Like I said, the technical information is extremely important (and interesting) but how the image looks AFTER printing, it seems to me, is even more important. However, I know that some people are not as interested in how the printed image looks as they are in technical superiority. And don't get me wrong. I'm not saying that there isn't a linear relationship between technical superiority and the printed image. But it is a matter of degree, not kind.

Case in point: I asked a guy in my town who bought a 5D and who had a 20D if he was going to take identical landscape shots and blow them up to 20x30 just to see what the printed difference was. His reply was an indignant, "I already know what it will look like, so I'm not going to do that." It is interesting that this particular person, who has a Masters Degree in computer science, would proclaim that he knows what something will look like without looking at it. It just proves the point that for some, the technical information is more important, not a real world test print, for although we may have a good idea what the print will look like, we have no way of knowing exactly what it will look like unless we at first view it. That "exactly" aspect is, I believe, what most people who want to print images are most interested.
« Last Edit: July 07, 2006, 02:50:16 am by dwdallam »
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
1Ds2 versus 5D
« Reply #49 on: July 07, 2006, 08:10:26 am »

Quote
Actually, except for a few very technically interested people, which is indeed extremely important information, what matters in the end is how the printed image looks at a respectable viewing distance. You print at 20x30 and you're not going to be looking at it from 10". At 20x30, probably more like 4+ feet. At 12x18 probably 1.5-2 feet at closest.


dwdallam,
This is very true. If you can position your over-enlarged print so that viewers are unable to peer at it closely, by placing a settee or large TV set in front of it, then you are home and dry. However, many people have an insatiable curiosity and can't resist walking up close to appreciate any fine detail that they think might be there. If the fine detail is not there when it should be, then it can cause disappointment. I do this sort of thing all the time just to check out whether I'm looking at a 35mm or MF enlargement, if for no other reason. If a settee's in the way, I'll climb over it. I'm not respectable, of course   .

Quote
So tell us, how will they look when viewed with the same image processed and printed at the same rez? If I had a IdsII I could do the test myself. But I only have the 20D and 5D. It would be interesting to me to see a landscape printed image at 12x18 and 20x30 from the 20D and 5D side by side, and the 1Ds.


You should be able to work this out. The 1Ds2 has 100% more pixels than the 20D, or approximately 40% more pixels in each dimension. Whatever size print you make from your 20D, the 1Ds2 will produce a print of similar quality, (viewed from the same distance) but 40% greater in length along each side. If you still can't get a clear idea from looking at a print and imagining an extended field of view, then try stitching a few 20D images to a resulting file size of 48MB.

Quote
Case in point: I asked a guy in my town who bought a 5D and who had a 20D if he was going to take identical landscape shots and blow them up to 20x30 just to see what the printed difference was. His reply was an indignant, "I already know what it will look like, so I'm not going to do that."


Fair enough! Like many of us, that guy in your home town has probably already been through his period of extensive testing of various issues. I recall a few years ago using a lot of ink and paper to test the visual differences between prints made from images at 150 ppi resolution, 180 ppi, 240 ppi, 360 ppi etc etc and lots of other ink and paper-consuming tests. If you've ever done any stitching of images, you will get a pretty good idea of the quality that larger file sizes produce.

It's not going to mean very much to you if a complete stranger on this site tells you that he regularly makes 20x30 prints from 20D images and that they are perfectly satisfactory, is it? Peoples' standards vary. Jack Flesher probably wouln't have a bar of such a print   .
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
1Ds2 versus 5D
« Reply #50 on: July 07, 2006, 09:10:49 am »

Quote
That the MTF at 10 lp/mm is given more weight than that at 40 lp/mm tells us that contrast trumps resolution when it comes to perceived sharpness. In other words, rather than examining the magnification where image details are no longer distinguishable as in your large blow ups, you should back off and view the image at lower magnification so as to better evaluate the contrast at lower resolution.


Bill,
Again, I wonder how much of this practice belongs to film legacy. My instinct on this tells me I want the highest MTF possible at around 40 to 60 lp/mm. I don't care if the performance dives above that, and I'm not too concerned whether the MTF at 10 lp/mm is 98 or 92% because I figure I can increase contrast with USM to mimic the effect of a contrasty lens. But I can't create detail that was never captured in the first instance as a result of the contrast being too weak. I figure there's not much that won't be captured at 10 lp/mm because of a 10% loss in contrast rather than a 5% loss in contrast.

Some lenses are high resolving but low contrast (presumably as a consequence of a relatively poor MTF at 10 lp/mm), and other lenses are contrasty, producing better looking prints with minimal adjustment, but are not particularly high resolving.

I believe Leaica lenses tend to be of the contrasty variety rather than the high resolving variety. I think it's difficult here to get the best of both worlds.

I don't have any Leica lenses, so can't test my gut feeling on this, but I know I can increase contrast dramatically in PS with various amounts of pixel radius.
Logged

Eric Myrvaagnes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 22814
  • http://myrvaagnes.com
    • http://myrvaagnes.com
1Ds2 versus 5D
« Reply #51 on: July 07, 2006, 10:33:14 am »

Ray,

Obviously you and Jack need an objective observer to mediate this dispute. I am willing to offer my time and expertise at a very low cost to do this for you. So Ray: please send me your 5D, and Jack: send me your 1DsII, and I'll be happy to do some comparisons when I get around to it.

I'm also willing to send the cameras back as soon as I get tired of either or both of them. How can you resist an offer like that?    

Eric
Logged
-Eric Myrvaagnes (visit my website: http://myrvaagnes.com)

Jack Flesher

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2592
    • www.getdpi.com
1Ds2 versus 5D
« Reply #52 on: July 07, 2006, 12:04:37 pm »

Quote
I believe Leaica lenses tend to be of the contrasty variety rather than the high resolving variety. I think it's difficult here to get the best of both worlds.

[a href=\"index.php?act=findpost&pid=70014\"][{POST_SNAPBACK}][/a]

Actually, Leica lenses are of the higher-resolution, lower-contrast variety and by comparison I have heard Nikkor's are of the higher-contrast, lower-resolution variety.  And I have also heard you cannot design in both attributes at the same time -- not sure why though I think it has to do with the fact that fancy coatings increase contrast and at the same time reduce resolution. (It's supposedly the  reason Heliopan does not offer multi-coated filters.)

I agree with you on contrast:  No way max performance at 10 LP/mm is going to create detail at 40 LP/mm when none existed there in the first place.  10 LP/mm performance is great for billboards at 300 feet and 4x6 prints of aunt Meg's 80th birthday party, but not for 24" landscape prints at 3 feet.
« Last Edit: July 07, 2006, 12:05:56 pm by Jack Flesher »
Logged
Jack
[url=http://forum.getdpi.com/forum/

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
1Ds2 versus 5D
« Reply #53 on: July 07, 2006, 12:29:43 pm »

Quote
Bob:

I think you will find those figures are the maximum linear resolutions for the sensors in question in "lines per mm" and not "line pairs per mm"...  If you halve your numbers, you essentially get the maximum Nyquist Line Pair resolution.  In my testing the numbers are pretty close too -- I get a maximum resolution on my particular 1Ds2 of about 54 LP/mm.  A bit lower than Nyquist due to the AA filter I suspect.

~~~

Where the discussion gets interesting for larger prints is how well the digital file interpolates and still maintains detail in the 20 to 40 LP/mm range.  With proper workflow, my experience shows the 1Ds2 interpolates very well (and FWIW so does the P45 -- extremely well in its case). And for that matter so does the 1D2.  But for whatever reason, my particular 5D did not interpolate as well as my 1D2 with same pixel pitch -- an anomalie I am at a loss to explain.

Cheers,
[{POST_SNAPBACK}][/a]

Jack,

Thanks for the correction. Yes, the resolutions for the cameras should be in lp/mm, not lw/mm. I corrected my posts.

The Nyquist frequency for the 1DsM2 is 69 lp/mm. For vertical MTF50, Imitest reports 34 lp/mm without sharpening and 45 lp/mm with standardized sharpening. In his review on DPReview, Phil reported an absolute vertical resolution of 2400 lines/picture height for the 1DsM2, or 50 lp/mm. Actual resolution is less than Nyquist because of because of the AA filter and imperfections in the de-mosaicing algorithm and defects in the lens.

The Imitest results from downloads of the images on DPReview are of interest. These are from in camera JPEGS, and your RawShooter Pro would probably do a better job in de-mosaicing than the camera's ASIC.

The 1DsM2 image is not sharpened, and Imitest shows the results for the non-sharpened image and the image with standardized sharpening. As the plots show, sharpening is critical for good MTF with digital images and that is why I think Bill Atkinson used sharpening in his tests and Ray should consider doing so in his comparisons.

[attachment=801:attachment]

The 5D image has been sharpened in camera as indicated by the overshoots in the top diagram, and Imitest attempts to correct for this sharpening. However, the increased response at and above Nyquist indicates problems related to a weak AA filter or oversharpening, and unfortunately Imitest can not distinguish the two. Therefore, I do not think that the Imitest results are valid for the 5D.

[a href=\"http://forums.dpreview.com/forums/read.asp?forum=1021&message=16693479]http://forums.dpreview.com/forums/read.asp...essage=16693479[/url]

[attachment=802:attachment]

Perhaps the 1DsM2 does have a better AA filter as suggested presiously in this thread, but tests with unsharpened images would be necessary to check for this.
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
1Ds2 versus 5D
« Reply #54 on: July 07, 2006, 02:09:09 pm »

Quote
Bill,
Again, I wonder how much of this practice belongs to film legacy. My instinct on this tells me I want the highest MTF possible at around 40 to 60 lp/mm.
[{POST_SNAPBACK}][/a]

Ray,

Actually these considerations have more to do with the characteristics of human vision than the medium used for photography. According to sceintific studies, human vision has a peak MTF at 6 cycles per degree, which translates to 1 cycle per milimeter when an image is veiwed at 34 cm (13.5 inches). This means that the MTF at 1 cycle/mm (1 lp/mm) is most important in determining the apparent sharpness of the image for this viewing distance. If you are viewing an 8 by 10 inch print at that distance, the corresponding resolution on a 35 mm negative or full frame 35mm type sensor would be 8 lp/mm. This corresponds closely to the 10 lp/mm criterion for MTF. Of course it would be preferable that the lens also have high MTF above 10 lp/mm but in lens design there is a trade off between resolution and contrast.

For 16 by 32 inch image, the corresponding resolution on the sensor would be 16 lp/mm and it would be 24 lp/mm for a 24 by 36 inch print. So, where you want the highest possible MTF varies with the size of the image. According to this analysis, the MTF at 24 lp/mm would be most critical for the 24 by 36 inch print viewed at 13.5 inches and one would want to maximize MTF for that resolution. Of course it would be better if the MTF would remain high right up to the Nyquist limit of 69 lp/mm.

[a href=\"http://bobatkins.com/photography/technical/mtf/mtf4.html]http://bobatkins.com/photography/technical/mtf/mtf4.html[/url]
« Last Edit: July 07, 2006, 02:21:44 pm by bjanes »
Logged

Jack Flesher

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2592
    • www.getdpi.com
1Ds2 versus 5D
« Reply #55 on: July 07, 2006, 04:35:07 pm »

Quote
Phil reported an absolute vertical resolution of 2400 lines/picture height for the 1DsM2, or 50 lp/mm. Actual resolution is less than Nyquist because of because of the AA filter and imperfections in the de-mosaicing algorithm and defects in the lens.
[a href=\"index.php?act=findpost&pid=70026\"][{POST_SNAPBACK}][/a]

FWIW, I never was able to hit 50 LP/mm converting with ACR2 (the original version in CS2, whatever version that was), but got up to 54 with RSE/P and C1. So yes, the demosaicing algorithm plays a significant role.
Logged
Jack
[url=http://forum.getdpi.com/forum/

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
1Ds2 versus 5D
« Reply #56 on: July 08, 2006, 11:56:50 am »

Quote
Actually, Leica lenses are of the higher-resolution, lower-contrast variety and by comparison I have heard Nikkor's are of the higher-contrast, lower-resolution variety.
[a href=\"index.php?act=findpost&pid=70023\"][{POST_SNAPBACK}][/a]



Whatever! I had in mind that either Zeiss or Leaica had a reputation for high contrast. Perhaps I got that the wrong way round. I own neither a Zeiss nor a Leica. In any case, I think that is now changing. Aren't Zeiss now boasting about absurdly high resolution for their latest lenses?
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
1Ds2 versus 5D
« Reply #57 on: July 08, 2006, 12:27:53 pm »

Quote
As the plots show, sharpening is critical for good MTF with digital images and that is why I think Bill Atkinson used sharpening in his tests and Ray should consider doing so in his comparisons.

[a href=\"index.php?act=findpost&pid=70026\"][{POST_SNAPBACK}][/a]

When doing comparisons, I like to keep the variables to a minimum. However, sharpening will help make the image look better (cosmetics) and I always do some sharpening for a finished print.

I took some test shots today of my neighbour's fence with my 100-400 plus 1.4x extender, with 20D and 5D. There's no sharpening and no interpolation, yet it's quite clear the 20D/560mm at f22 is the sharpest and the 5D at f11 the least sharp. There might be some doubt comparing the 20D at f11 and the 5D at f22. They look so close they are as good as equal as far as I'm concerned, but I give the edge to the 5D at f22.

[attachment=804:attachment]
Logged

Jack Flesher

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2592
    • www.getdpi.com
1Ds2 versus 5D
« Reply #58 on: July 08, 2006, 03:01:32 pm »

Shifting gears a bit, I think I should offer the following summary of these cameras so perhaps everybody better understands my reasoning for my current choices.

1) If I am primarily hand-holding the camera and using zooms (even L zooms), as say one might in a travel photography situation, then I suspect I would see little to no detail differences in the 5D and 1Ds2 images.

2) I might see some slight differences in DR or focus accuracy, and even detail in a few perfectly captured files, but even I doubt those would be significant enough to justify the cost or added weight of the 1Ds2 in that travel photography situation.

3) In fact, I'm not convinced that in hand-held, zoom-lens imaging I see a significant enough difference between the 20D (30D) and 1Ds2 in detail to justify the added weight/cost for travel.  (Obviously I do see differences, though they are relatively minor.)  

4) I can see color and DR advantages in the 1Ds2 over the 20D, and am still debating with myself as to whether or not these are significant enough to justify the weight/cost when traveling with zooms.

4a) Even I think a 5D and 20/30D combo with a few L zooms would be an ideal travel pair of cameras.  

5) In the studio on a tripod or in the field on a tripod, I can see enough of a detail difference for me to justify the cost/weight of the 1Ds2 to use it over other DSLR options. The added benefit of slightly broader DR and better color fidelity add to this justification.

6) If money were no object, I would own a 30D, 5D, 1Ds2 and would add the 1D2 for fast-action.  

7) Unfortunately money is an object for me, hence I needed to make a decision regarding which cameras I was going to keep.  

8) If I owned a Hassy H2 and P45 or other 30-plus MP digital back, I would NOT need the 1Ds2 detail/DR/color, would sell it and get a 5D as my casual or back-up full-frame DSLR replacement.  

9) Believe it or not, I have in fact considered #8 since I do own the Betterlight scanning back.  It serves my studio needs superbly, but it suffers from the weaknesses in the field that only direct single-capture can overcome -- hence my reservations.  

10) I suspect that #8 and #2 are precisely why Michael Reichmann has gone to the route he has with the 5D and H2...  

Cheers,
« Last Edit: July 08, 2006, 03:54:46 pm by Jack Flesher »
Logged
Jack
[url=http://forum.getdpi.com/forum/

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
1Ds2 versus 5D
« Reply #59 on: July 08, 2006, 10:10:53 pm »

That's a very reasonable and well-balanced assessment of the options, Jack.  

We should not lose sight of the fact that a lot of photography is about catching the moment. My interest in such hair-splitting differences is mainly so I can make an informed choice whether or not to ignore them. All my digital camera purchases have been made on the basis of substantial improvements over what I previously used.
Logged
Pages: 1 2 [3] 4 5 ... 7   Go Up