Pages: [1] 2 3 ... 5   Go Down

Author Topic: Michael's Tests of EOS-1Ds  (Read 24255 times)

Larry Smith

  • Guest
Michael's Tests of EOS-1Ds
« on: September 26, 2002, 09:24:54 pm »

[font color=\'#000000\'](Sorry for posting this twice, ...just sent it, and realized that this was the appropriate thread for it :-0

Michael writes:
"I heard today from a large Canon dealer, (in a country and city that will remain unnamed), that the Canon distributor has asked them to get deposits from anyone wishing to place an order for a 1Ds. No reason is forthcoming, but my surmise is ..."

How likely is it that Canon is using this 'deposit' requirement to guage acceptance of their too-high intro price, ...since these depositers will be doing so without knowing whether the price will come down, and to-what. Why not just give Canon a blank check?

Maybe if anyone who puts down a deposit under this current price-rumor had to actually PAY this price, when the camera comes, and KNEW that they would, ...then the deposit-list would give a truer picture of market-acceptance-at-this-price!

Larry
 ???[/font]
Logged

Jens Haas

  • Guest
Michael's Tests of EOS-1Ds
« Reply #1 on: September 27, 2002, 12:31:51 pm »

[font color=\'#000000\']The moire of the 1Ds seems to be an issue not only for fashion but for advertising photography in general. Is this caused by Canon's software settings (to be corrected by a firmware update if necessary) or is it a hardware "issue" that stays with you once you buy the camera?

Thanks for any info on this.

J.[/font]
Logged

Rainer SLP

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 727
    • RS-Fotografia
Michael's Tests of EOS-1Ds
« Reply #2 on: September 29, 2002, 04:58:49 pm »

[font color=\'#000000\']Hi,

I guess you already saw it but if not Michael released part 3 and 4 of his EOS 1Ds review.[/font]
Logged
Thanks and regards Rainer
 I am here for

Rainer SLP

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 727
    • RS-Fotografia
Michael's Tests of EOS-1Ds
« Reply #3 on: October 01, 2002, 09:00:17 pm »

[font color=\'#000000\']Hi Jim,

The 1Ds was far superior than the SD9. I did not even apply Unsharpen Mask.

On this 2 Images I did not see noise.

I took a look at the new images from Phil Askey in the dpreview.com forum and I would say what people saw as noie I can live with it because it looks more like film grain.

I think that since the digital came out the people have forgotten that there was someting like grain and they are now fooled with the digital images because they are so clean that sometimes they think there is more resolution in it which I think is not true.

I still think that film has more resolution than digital compared even to the ~11 MioPixel from Canon and the unseen samples of the Kodak ~14 MioPixel.

OK the software interpolation is in any case cleaner with a digital rather than on an analog filmscan, but with interpolation you do not win sharpness in my opinion. Afterwards applying sharpness just increases the contrast between lowlight borders and highlight borders.

Have taken a deep look what sharpening makes?

I always try to figure out on a landscape ¿ with how many pixels do I resolve some motif scanning it at 4000dpi and this same motif how many pixels are delivered by the DSLR?

4000dpi optical scanning resolution are ~157 pixels/mm and the highest resolution of the D60 is ~135pixels/mm, the 1Ds is ~113pixels/mm and the Kodak will have ~127pixels/mm. So there is still a long way to go to get the 157píxels/mm of the scanning.

The other thing I still do not understand is how the people calculate the resolution in Line pairs per mm for a digital camera?[/font]
Logged
Thanks and regards Rainer
 I am here for

sergio

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 666
    • http://www.sergiobartelsman.com
Michael's Tests of EOS-1Ds
« Reply #4 on: October 02, 2002, 01:31:43 pm »

[font color=\'#000000\']MTF curves could provide more precise info about the system. However I only care for the results I get in my photographs, not in theory or in photographed charts. How good is good enough? Does it do the job? Are the results pleasing? Is it fun to use? Those are things I care more about than scientific sensor tests.[/font]
Logged

sergio

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 666
    • http://www.sergiobartelsman.com
Michael's Tests of EOS-1Ds
« Reply #5 on: October 03, 2002, 11:28:10 am »

[font color=\'#000000\']Are they jaggies in the diagonal lines of the parabolic antenna in the Eos 1Ds review? I dont see them in the film photographs, just in the Eos 1Ds
Does anybody see them too?[/font]
Logged

Marshal

  • Guest
Michael's Tests of EOS-1Ds
« Reply #6 on: October 05, 2002, 03:26:12 am »

[font color=\'#000000\']
Quote
Since the 1Ds fits 11.4 million pixels in a 35.8 x 23.8 mm imager and the 14n will fit 13.89 in the same area, that suggests the 14n's pixel pitch will be 7.2 microns. Working backward (and presuming my math is correct), 16.72 million pixels, each at 6 microns, would meet the ceiling imposed by lens resolving power.

This is not to suggest that even a 16.72 megapixel camera would provide the ultimate resolution for 35mm format. I think Foveon has taught us that. I want a full-frame, 17 megapixel X3 dSLR for XMas.
Dale:

To be more precise about the MP numbers of the Canon and Kodak cameras, the actual effective megapixels are 11.1 and +/- 13.5 respectively, not 11.4 and 14MP. So how many microns they actually are, are still in question. No offense, but these speculated micron numbers are very theoretical.

They may be around 7-8 microns, but micron count doesn't mean nearly as much to me as what I actually see when I'm looking at a 13X19 print and evaluating the image quality. The clarity, sharpness to my eye, not some mathematical equation. That's where it is. How clean or grain free is the image? Does it look like a medium format image?

We can throw around numbers till you know what freezes over, but as the old saying goes, "a picture is worth a thousand words." What matters are actual photos taken by these cameras. A lens resolution chart is a start, but show me some real world samples like the kind we've looked at from the Canon, Kodak and Sigma/Foveon and then it becomes real world stuff.

I'm looking forward to seeing more samples from Phil, the Imaging resource guy(can't remember his name), Kumio Yamada and others. And seeing some 13X19 prints from the 1Ds on the LLVJ coming out in November. I would actually rather see those prints then hear Michael discuss his impressions about them. If Michael wants to discuss his impressions about the AF and AE ability and various other features(voice recording, durability, etc)then cool. But we already have a good idea of Michael's feelings about the image quality. To me, the proof is in the picture(s).[/font]
Logged

Raymond

  • Newbie
  • *
  • Offline Offline
  • Posts: 2
Michael's Tests of EOS-1Ds
« Reply #7 on: October 21, 2002, 09:42:39 am »

Quote
[font color=\'#000000\']Perhaps the following is useful and easy to remember:
1. Nyquist theory says that you need two pixels to resolve a line pair. Yes, I actually verified this with a D30. Not 4 pixels.

3. The "diffraction limit" is not a hard cutoff. It depends on definitions used. If you start with black & white bars and accept very low contrast in the output, lenses can deliver absurdly (and uselessly) high lp/mm. This is the primary source of those 100+ lp/mm claims that you used to see in decades past. Photodo has a lengthy expose on this on their MTF explanations. You need a definition that delivers a usefully high contrast level. Astronomers tend to use FWHM (Full Width at Half Maximum of the Airy disc) since this is what they use to match pixel sizes to telescope configurations, seeing conditions, and the Nyquist criterion (two pixels per line-pair). With this definition the theoretical diffraction limit of a lens in lp/mm = 1000/f-no for a wavelength of 4900 Angstroms (yellow light). Ie a theoretically perfect lens delivers 125 lp/mm at f-8. It would be more lousy in deep red light (< 90 lp/mm) and a lot better in deep blue (> 160 lp/mm).[/font]
[font color=\'#000000\']Samirkharusi,
Thanks for clarifying some of these points. I remember making an assumption a few years ago that one pixel, or scanning dot, was equivalent to one resolution line (a line being just a series of dots joined seamlessly). It seems my original assumption was correct, provided one understands that a black line on a white background has no more significance than a white line on a black background.

I'm not so sure about this Rayleigh's Law figure of 1000. I've seen figures ranging between 1800 and 1000. Again I assumed the lower figure related to red light.

However, I'm surprised at your assertion that there's no advantage in using pixels smaller than 6 microns in a 35mm digital sensor. If this is true, there is clearly a different relationship between lens and sensor, and lens and film.  Can you elaborate on these differences, please?

Another of your statements which appears to be in conflict with other information (perhaps misinformation) available on the internet is in regard to 35mm lenses being optimised for best performance at F8. Are you referring to an overall compromise where corner sharpness at F8 might be much improved although centre sharpness is well beyond the diffraction limit? Do you give much credence to the theory that film performs best at large F stops such as F/22 and most 'good' lenses perform best at either one or two F/stops down from maximum aperture? I have an 400mm F/4 Sigma prime lens which is sharpest at F5.6, but at F5.6 there's noticeable light fall off at the corners (Minolta mount!). At F11, there's no noticeable light fall off at the edges or corners. Would you describe this lens as being optimised for best performance at F5.6 or F11?[/font]
Logged

samirkharusi

  • Full Member
  • ***
  • Offline Offline
  • Posts: 196
    • http://www.geocities.com/samirkharusi/
Michael's Tests of EOS-1Ds
« Reply #8 on: October 26, 2002, 12:03:08 pm »

[font color=\'#000000\']
Quote
I'm not so sure about this Rayleigh's Law figure of 1000. I've seen figures ranging between 1800 and 1000.
However, I'm surprised at your assertion that there's no advantage in using pixels smaller than 6 microns in a 35mm digital sensor. If this is true, there is clearly a different relationship between lens and sensor, and lens and film.  Can you elaborate on these differences, please?

Another of your statements which appears to be in conflict with other information (perhaps misinformation) available on the internet is in regard to 35mm lenses being optimised for best performance at F8.
Raymond, pardon the late response. Here goes:
Diffraction Limit has many definitions. Rayleigh's is simply the diameter (radius?) of the Airy Disc to first minimum. Using FWHM is the diameter of the same disc but at half the peak intensity. There's also something called Dawe's Limit. What matters is to note that there's no sharp cutoff as to how many line-pairs you can see. Photodo explains this at length. It's a matter as to how low a contrast you are willing to tolerate. Using the 1000/f-no relates to an MTF contrast of about 33% as plotted on photodo. Astronomers use "critical sampling" because they are all too aware that increasing sensor resolution dramatically increases noise. But when there's plenty of light (shooting planets and the Moon) they do use "over-sampling". This would be equivalent to using pixels much smaller than, say, the D60 7.5 microns with 35mm lenses. It's not that you do not resolve more, it's just that you pay in a lot more noise. The Nyquist criterion is the generally accepted point for compromise.
Lenses optimised for f8? I actually tested 5 prime lenses on a D30. A Sigma 500/7.2 APO was so lousy we need not discuss it further, max resolution 35 lp/mm. The other 4, Tamron 14/2.8, Canon 28/2.8, Canon 50/1.8, Canon 100/2.8 macro all peaked at f8. I expected variations. Coincidence? so I can only conclude that they were actually designed that way ;-)
As for 35mm lenses being the limiting factor with both the D60 and the 1Ds, I would refer any doubters to test their best lenses using Norman Koren's lens test chart:
http://www.normankoren.com/Tutorials/MTF5.html
It's very educational since he brings contrast into his chart, as opposed to the USAF chart.[/font]
Logged
Bored? Peruse my website: [url=http://ww

jeffreybehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 365
  • Happily retired accountant
Michael's Tests of EOS-1Ds
« Reply #9 on: November 01, 2002, 03:38:26 pm »

[font color=\'#000000\']Then don't buy a 1Ds, Doug.  All that cra...er...FILM equipment is NOT the same as a 1Ds and will not produce the same results.  And I know that all that film equipment would produce better results from some perspectives, but at what price in time and processing costs and lack of conveneince?

Pick your poison, boys.  There'll be lots of fotografic soul-searching done the next few months.[/font]
Logged

Marshal

  • Guest
Michael's Tests of EOS-1Ds
« Reply #10 on: September 26, 2002, 02:01:36 am »

[font color=\'#000000\']Amazing how not only the same week and within 2 days mind you, of the official announcement across the pond, Michael has a 1Ds to test already in Canada.

In looking over the sample pics of the Toronto skyline, I have to draw the same conclusions as he did. I look forward to seeing the comparison between the 1Ds and the D60 and in RAW format at that. Even if we have to wait a few more weeks.

And speaking of file formats, apparently the tests Michael made were with Jpegs if I understand correctly. If RAW was not available to test because of the lack of new software, then the sample pics came from Jpeg or Tif files. I apologize in advance if I misunderstood what you said about that Michael. Looking forward to new test results and also any pics shot on a nature shoot. I have a feeling the next edition of LLVJ will be a very popular one.[/font]
Logged

Mike Spinak

  • Newbie
  • *
  • Offline Offline
  • Posts: 22
Michael's Tests of EOS-1Ds
« Reply #11 on: September 26, 2002, 10:00:12 pm »

[font color=\'#000000\']Dear Michael,

Please make sure to include the 1Ds's level of dust proneness/resistance in your review. It has been hypothesized that the dust proneness would be quite low, like the D30/60, because of the CMOS chip. On the other hand, the sample images that Canon released for the 1Ds looked rather dusty. What's the story?

Thank you.[/font]
Logged

Ray Robertson

  • Guest
Michael's Tests of EOS-1Ds
« Reply #12 on: September 27, 2002, 11:51:11 am »

[font color=\'#000000\']Michael,
Thanks for the detailed review of the 1Ds. In my view, the ISO 50 shot had marginally more detail in the shadows, but overall does the 1Ds display an increase in dynamic range? I don't know, but I think not from the sample pics. Yet you would expect it to because of the larger pixel size. Perhaps the raw files will reveal this.

Also, according to my calculations, the D60 should display marginally better resolution than the 1Ds at equal focal lengths, the trade off being the D60 image will be cropped. Is this in fact the case? Shooting at the long end of the most powerful zoom in the camera bag could be a 'real world' situation, could it not?[/font]
Logged

rwzeitgeist

  • Guest
Michael's Tests of EOS-1Ds
« Reply #13 on: September 30, 2002, 12:33:46 am »

[font color=\'#000000\']One of Michael's comments in Part 4 suggests Canon should change the .TIF file name suffix used by the 1D and 1Ds to something else.  He points out the risk of unintentionally rewriting the raw file when using a tool ignorant of the true nature of the file.

  Canon could substantially lower the risk of that particular error by simply writing the files with the read-only attribute set to "true."  Sounds like a very simple change I'd like to see.

    Bob[/font]
Logged

jwarthman

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 99
Michael's Tests of EOS-1Ds
« Reply #14 on: October 01, 2002, 08:37:25 pm »

[font color=\'#000000\']Rainer,
Thanks very much for the pointers - very interesting! (Now, if only I could read Japaneese!  ;-)

I may print both the SD9 and 1Ds samples, and also the 1Ds sample downsized-then-upsized to simulate this image taken with the D20, just for comparison sake.

Enjoy!

--- Jim

P.S.  I forgot to ask, was it the 1Ds that you thought superior? How does the "noise" appear in your two prints?[/font]
Logged

mmurph

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 506
    • http://
Michael's Tests of EOS-1Ds
« Reply #15 on: October 03, 2002, 11:00:13 am »

[font color=\'#000000\']
Quote
I was surprised to see the venue of Michael's shots change from Toronto to Florida...  Do you think he likes the camera so much that he has taken it on the run??  

I was hoping this was a new reality-based show.  One where Michael takes off with Canon's camera, then posts pics from around the globe while Canon tries to track him down.  Given his recent movements (Toronto ->Florida ->Northern Ontario ->Yellowstone) I may not be wrong.  (Hint: you can hide from Nikon in Cologne ;>)


Michael, I meant to post a "thank you" for your D60 review in the spring.  Thanks also for this review - for listening to and answering the concerns of posters on your site, for keeping a close focus on the pragmatic, and for patiently putting up with of the net BS.

Have a good trip!  (Or better: enjoy your rest when you finally get home.)

Best,
Michael T.[/font]
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Michael's Tests of EOS-1Ds
« Reply #16 on: October 05, 2002, 08:22:14 am »

Quote
[font color=\'#000000\']"...According to Olympus, most current lenses for 35mm SLRs resolve down to about 10 microns, and top-end lenses to around 6 microns."[/font]
[font color=\'#000000\']Dale,
Have you any idea what this statement really means? I think we can safely assume it refers to a 'real world' context of low contrast ratios, say 1.6:1 to 30:1. According to my maths 0.006mm is equivalent to 167 lines/mm or 83 lp/mm. Since this is roughly equal to the resolving power of good film in low contrast situations, the product of the two gives a maximum system resolving power for 35mm of about 40 lp/mm. Would that be about right? Sounds a bit low to me.

But digital sensors are not film and the same formulas do not necessarily apply. What formulas do apply?[/font]
Logged

Howard

  • Guest
Michael's Tests of EOS-1Ds
« Reply #17 on: October 20, 2002, 09:46:12 pm »

[font color=\'#000000\']
Quote
Hi Jim,

The 1Ds was far superior than the SD9. I did not even apply Unsharpen Mask.

On this 2 Images I did not see noise.

I took a look at the new images from Phil Askey in the dpreview.com forum and I would say what people saw as noie I can live with it because it looks more like film grain.

I think that since the digital came out the people have forgotten that there was someting like grain and they are now fooled with the digital images because they are so clean that sometimes they think there is more resolution in it which I think is not true.

I still think that film has more resolution than digital compared even to the ~11 MioPixel from Canon and the unseen samples of the Kodak ~14 MioPixel.

OK the software interpolation is in any case cleaner with a digital rather than on an analog filmscan, but with interpolation you do not win sharpness in my opinion. Afterwards applying sharpness just increases the contrast between lowlight borders and highlight borders.

Have taken a deep look what sharpening makes?

I always try to figure out on a landscape ¿ with how many pixels do I resolve some motif scanning it at 4000dpi and this same motif how many pixels are delivered by the DSLR?

4000dpi optical scanning resolution are ~157 pixels/mm and the highest resolution of the D60 is ~135pixels/mm, the 1Ds is ~113pixels/mm and the Kodak will have ~127pixels/mm. So there is still a long way to go to get the 157píxels/mm of the scanning.

The other thing I still do not understand is how the people calculate the resolution in Line pairs per mm for a digital camera?
Hmmm......when I look at both images they are VERY hard to compare. One shows a deep blue sky and the other shows a gray-green sky.  And, the SD9 image appears underexposed.
  Very hard to compare.
  One camera seems to have the color all wrong.
  I'm not so sure which it is. Does the SD9 have a blue cast or is it merely under-exposed? Does the Canon have aslight greenish cast in the sky or is it smog?
  <boggle>
  I honestly don't know.
  I'd sure like to see a Kodachrome 25 of the same shot for reference.
  Regards....Howard[/font]
Logged

samirkharusi

  • Full Member
  • ***
  • Offline Offline
  • Posts: 196
    • http://www.geocities.com/samirkharusi/
Michael's Tests of EOS-1Ds
« Reply #18 on: October 26, 2002, 12:04:19 pm »

[font color=\'#000000\']
Quote
I'm not so sure about this Rayleigh's Law figure of 1000. I've seen figures ranging between 1800 and 1000.
However, I'm surprised at your assertion that there's no advantage in using pixels smaller than 6 microns in a 35mm digital sensor. If this is true, there is clearly a different relationship between lens and sensor, and lens and film.  Can you elaborate on these differences, please?

Another of your statements which appears to be in conflict with other information (perhaps misinformation) available on the internet is in regard to 35mm lenses being optimised for best performance at F8.
Raymond, pardon the late response. Here goes:
Diffraction Limit has many definitions. Rayleigh's is simply the diameter (radius?) of the Airy Disc to first minimum. Using FWHM is the diameter of the same disc but at half the peak intensity. There's also something called Dawe's Limit. What matters is to note that there's no sharp cutoff as to how many line-pairs you can see. Photodo explains this at length. It's a matter as to how low a contrast you are willing to tolerate. Using the 1000/f-no relates to an MTF contrast of about 33% as plotted on photodo. Astronomers use "critical sampling" because they are all too aware that increasing sensor resolution dramatically increases noise. But when there's plenty of light (shooting planets and the Moon) they do use "over-sampling". This would be equivalent to using pixels much smaller than, say, the D60 7.5 microns with 35mm lenses. It's not that you do not resolve more, it's just that you pay in a lot more noise. The Nyquist criterion is the generally accepted point for compromise.
Lenses optimised for f8? I actually tested 5 prime lenses on a D30. A Sigma 500/7.2 APO was so lousy we need not discuss it further, max resolution 35 lp/mm. The other 4, Tamron 14/2.8, Canon 28/2.8, Canon 50/1.8, Canon 100/2.8 macro all peaked at f8. I expected variations. Coincidence? so I can only conclude that they were actually designed that way ;-)
As for 35mm lenses being the limiting factor with both the D60 and the 1Ds, I would refer any doubters to test their best lenses using Norman Koren's lens test chart:
http://www.normankoren.com/Tutorials/MTF5.html
It's very educational since he brings contrast into his chart, as opposed to the USAF chart.[/font]
Logged
Bored? Peruse my website: [url=http://ww

Rainer SLP

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 727
    • RS-Fotografia
Michael's Tests of EOS-1Ds
« Reply #19 on: October 02, 2002, 05:20:18 pm »

[font color=\'#000000\']Hi Sergio,

Yes you are absolutely right. The other point is everybody sees different and has another quality sense.[/font]
Logged
Thanks and regards Rainer
 I am here for
Pages: [1] 2 3 ... 5   Go Up