Ad
Ad
Ad
Pages: « 1 [2] 3 »   Bottom of Page
Print
Author Topic: Canon MF Rumor  (Read 12318 times)
ziocan
Sr. Member
****
Offline Offline

Posts: 427


« Reply #20 on: August 10, 2010, 10:57:28 AM »
ReplyReply

There is enough diffraction limits with the current 21/24 MP DSLRs.
IMO In 35mm FF there is not much room for detail improvement with current lenses.


Logged
Phil Indeblanc
Sr. Member
****
Offline Offline

Posts: 1104


« Reply #21 on: August 10, 2010, 07:28:51 PM »
ReplyReply

I agree with this and the previous post. 1Ds Mark IVx lets call it.
I would love to see the flex of Canon speed, ISO in a MF cube. Pop that on a RZ or Sinar...Wow!

I would think they can EASILY make it priced agressively.
As long as the quality is there with large pixels(original 1Ds size or so) for good movements. After this much momentum they have gained, I can't see any reason for them not to do it... I think they would have a very good chance of succeeding. Universal back adapt to MF and LF.They would cut ALL the costs of the camera body, and all into the chip and process.  WOW.
Logged

If you buy a camera, you're a photographer...
dergiman
Full Member
***
Offline Offline

Posts: 171



WWW
« Reply #22 on: August 11, 2010, 01:31:43 AM »
ReplyReply

i think canon can go into 30+ megapixel territory by dropping the AA-filter and gain a reasonable amount of additional sharpness. i hope that they  really work on better dynamic range and better color reproduction (maybe by going 16bits). and they should add a shooting mode that uses an electronic shutter like in video that allows flash sync at all speeds. that would be a huge step foreward and eliminates allmost all mf advantages.

Philipp
Logged

Philipp Derganz Photography
www.philippderganz.com
Architektur Photo
Wien, Austria
erickb
Full Member
***
Offline Offline

Posts: 235


WWW
« Reply #23 on: August 11, 2010, 02:12:23 AM »
ReplyReply

I think not, I shall never buy a new Canon with more pixels, that's already too much
Logged
markhodkinson
Newbie
*
Offline Offline

Posts: 46


« Reply #24 on: August 11, 2010, 05:56:49 AM »
ReplyReply

I have to agree, i just changed from Canon 1dsmk3 to a phase one P30, i think 21 mpx is enough on the 35mm size sensor.
Logged
erickb
Full Member
***
Offline Offline

Posts: 235


WWW
« Reply #25 on: August 11, 2010, 07:02:49 AM »
ReplyReply

or a Hasselblad H3D II 22mp
Logged
Phil Indeblanc
Sr. Member
****
Offline Offline

Posts: 1104


« Reply #26 on: August 11, 2010, 11:10:58 AM »
ReplyReply

A larger sensor is what I was talking about about...not cramming more pixels...

and yes dropping the AA would be SOOO nice!..at least a removable one, or none.  I don't think anyone is really holding out for a 35mm size to be 30 or more mpixels. It is likely to happen, but so what?
Logged

If you buy a camera, you're a photographer...
BJL
Sr. Member
****
Offline Offline

Posts: 5121


« Reply #27 on: August 11, 2010, 02:19:02 PM »
ReplyReply

There is enough diffraction limits with the current 21/24 MP DSLRs.
IMO In 35mm FF there is not much room for detail improvement with current lenses.
You are probably right about the resolution limits of some 35mm format lenses, but not the best 35mm format primes ... which are the main ones relevant to a discussion of higher resolution possibilities.

On diffraction, I have to say this yet again:
diffraction at, say, 30MP or 40MP would not be not any more of a problem in 35mm format than in a larger format. Yes, controlling diffraction to make use of the sensor's full resolution requires using a lower f-stop with the smaller format (about one stop lower for 35mm than 645), but the smaller format gives the same DOF at that lower f-stop, just with more "speed" so that a lower ISO speed and/or shorter exposure time is used: not much of a problem.

Maybe someday diffraction could be more of a problem for 35mm than for MF, but only if and when 35mm format reaches about 100MP (3 micron pixel spacing). Then reducing diffraction effects enough to get full resolution in 35mm format would require opening up to about f/4 whereas MF would use about f/5.6, and aberrations might be more of a problem at f/4 than f/5.6.

Even then: at the same f-stop, increasing pixel count in the same format has no effect on diffraction spot size and improves sensor resolution, so overall resolution gets better, not worse!
Logged
erickb
Full Member
***
Offline Offline

Posts: 235


WWW
« Reply #28 on: August 11, 2010, 02:26:38 PM »
ReplyReply

you can turn it as you like but with f/11 f/16 on 35 and  f/22 on MF you get diffraction if the pixels are too smalls, using a lower f-stop  is not always what you want or need
Logged
revaaron
Sr. Member
****
Offline Offline

Posts: 333


« Reply #29 on: August 11, 2010, 02:33:32 PM »
ReplyReply

man... do I hate the new llok of this forum.


anyhow, there will be no DSLR that removes or has a removable AA. The last one that did that (which I'm currently waiting for the battery to charge) was the SLR/n. Camera makers aren't going to remove this because the standard user would complain about moire patterns in every single picture they shoot.
Logged

BJL
Sr. Member
****
Offline Offline

Posts: 5121


« Reply #30 on: August 11, 2010, 03:08:25 PM »
ReplyReply

you can turn it as you like but with f/11 f/16 on 35 and  f/22 on MF you get diffraction if the pixels are too smalls, using a lower f-stop  is not always what you want or need
Agreed: my points are only that

A. the problem is equally bad in any format, so long as lenses are good enough to give the needed resolution, so it is wrong to offer this as a reason for going to a larger format than 35mm.

B. more pixels still give a clear increase in resolution at lower f-stops, and no decrease in resolution at any f-stop, so it is silly to say that there is no point increasing 35mm format sensor resolution simply because at some choices of f-stop (f/11 and above?), there is not much resolution gain.


As to why it is equally bad in any format (with good enough lenses):
1. The dominant reason for stopping down to an aperture small enough to cause diffraction problems, like your example of f/11 in 35mm, is to get enough DOF.
2. Changing to MF, that same need for DOF requires a higher f-stop: about f/16 in your example.
3. At the different f-stops needed to get the same DOF in the different formats, diffraction hurts resolution just as much.

The hard fact is that pushing resolution to the levels offered by sensors of 20MP and up imposes noticeable limits on the DOF attainable in conjunction with that high resolution, through the f-stops needed to adequately control diffraction, and this happens to a degree which does not depend on the format. I suspect that this will someday set the practical resolution limits for still photography, short of special techniques like DOF stacking. (Stitching does not really help: it just effectively turns a lens for one format into a wider angle lens in a larger format.)


P. S. What about all those giga-pixel images? The ones I have seen are all wide-angle views of distant panoramas, avoiding nearby subjects, which is the easiest case for DOF.
« Last Edit: August 11, 2010, 03:10:18 PM by BJL » Logged
ziocan
Sr. Member
****
Offline Offline

Posts: 427


« Reply #31 on: August 11, 2010, 10:20:28 PM »
ReplyReply

You are probably right about the resolution limits of some 35mm format lenses, but not the best 35mm format primes ... which are the main ones relevant to a discussion of higher resolution possibilities.

On diffraction, I have to say this yet again:
diffraction at, say, 30MP or 40MP would not be not any more of a problem in 35mm format than in a larger format. Yes, controlling diffraction to make use of the sensor's full resolution requires using a lower f-stop with the smaller format (about one stop lower for 35mm than 645), but the smaller format gives the same DOF at that lower f-stop, just with more "speed" so that a lower ISO speed and/or shorter exposure time is used: not much of a problem.

Maybe someday diffraction could be more of a problem for 35mm than for MF, but only if and when 35mm format reaches about 100MP (3 micron pixel spacing). Then reducing diffraction effects enough to get full resolution in 35mm format would require opening up to about f/4 whereas MF would use about f/5.6, and aberrations might be more of a problem at f/4 than f/5.6.

Even then: at the same f-stop, increasing pixel count in the same format has no effect on diffraction spot size and improves sensor resolution, so overall resolution gets better, not worse!


I use Zeiss primes on the a900 (85mm and 135mm), maybe because they are fast lenses and optimized to give the best quality at around f5.6, but they begin to show diffraction at f11 and at f16 it is really there.

On the Phase with the P30+ and the older 80mm and 150mm non D, I need to stop at f16 or 22 before seing the effect of diffraction.

I'm not saying that a 32/36 megapixel 35mm sensor will not shows its advantages relative to the current 24 and 21 megapixels sensors, but I think the increase of perceived and effective sharpness will be negligible.

In any case, if Sony, Nikon or Canon will do a body with a 30+ mp sensor I will be happy to buy it.
« Last Edit: August 11, 2010, 10:23:20 PM by ziocan » Logged
ziocan
Sr. Member
****
Offline Offline

Posts: 427


« Reply #32 on: August 11, 2010, 10:29:11 PM »
ReplyReply

man... do I hate the new llok of this forum.


anyhow, there will be no DSLR that removes or has a removable AA. The last one that did that (which I'm currently waiting for the battery to charge) was the SLR/n. Camera makers aren't going to remove this because the standard user would complain about moire patterns in every single picture they shoot.
I had one of those Kodaks for a while and the moiré was not as bad as you describe. It was slightly worse than the 1ds, yet not on all the circumstances.

I think with the newer and better algorithms to process raw files and more computing power we have today, it is possible to have a 35mm FF with a very low pass or no filter at all.
I remember that with the Kodak, as software was being updated (not the Kodak one) and the algorithm became more efficient on demosaicing, though it required more processing time, the results were getting significantly better.
Logged
Phil Indeblanc
Sr. Member
****
Offline Offline

Posts: 1104


« Reply #33 on: August 12, 2010, 12:49:17 AM »
ReplyReply

I had one of those Kodaks for a while and the moiré was not as bad as you describe. It was slightly worse than the 1ds, yet not on all the circumstances.

I think with the newer and better algorithms to process raw files and more computing power we have today, it is possible to have a 35mm FF with a very low pass or no filter at all.
I remember that with the Kodak, as software was being updated (not the Kodak one) and the algorithm became more efficient on demosaicing, though it required more processing time, the results were getting significantly better.
I have one also, the SLR/C. It really pulled the gap between my 35mm images vs my Phase One images.

I agree 100%. the moire was mostly an issue for clothing/fashion, which is significant, but other than that, I had no issues with it for any of the studio work i did. I did have a STORNG magneta cast if I had any light aimed towards it.
I wish the ISO wasn't locked in with the time, so it would be fully manual. In fact, if that magenta cast is something that can be fixed, I would still be using it.
Even the processing software was very advanced in many ways.

As you say, with newer sensors, I think it would be less problems from the SLRx days.   I have not tried Foveon, but I think the Sigma cams don't have a AA either(?)



Yes and 16 is where I top out for MF def issues. Maybe its the blades or my memory is blurry, but I remember using my Leica 100macro lens on the 1Ds and was able to stop down much more than the 100 macro or 180 .
Logged

If you buy a camera, you're a photographer...
BJL
Sr. Member
****
Offline Offline

Posts: 5121


« Reply #34 on: August 12, 2010, 01:10:13 PM »
ReplyReply

I use Zeiss primes on the a900 (85mm and 135mm) ... they begin to show diffraction at f11 and at f16 it is really there.

On the Phase with the P30+ and the older 80mm and 150mm non D, I need to stop at f16 or 22 before seing the effect of diffraction.
It seems that all three of us agree then: with a larger format the diffraction effects kicks in at a somewhat higher aperture ratio, (about one stop higher for MF than 35mm) ... which balances the different DOF at equal f-stop and equal FOV in different formats to give equal DOF at the "diffraction limited aperture".

But perhaps I am misunderstanding your comment, so let me ask what display and viewing conditions you are using when making these judgements of visible diffraction?
- Prints or onscreen viewing at equal PPI? (e.g. 100% pixels on-screen)
- Prints or onscreen viewing of the image at the same displayed size (maybe on-screen viewing of a crop to the same portion of the image)
- Other?
Logged
erickb
Full Member
***
Offline Offline

Posts: 235


WWW
« Reply #35 on: August 12, 2010, 03:28:35 PM »
ReplyReply

printing big size or 100% on the screen
Logged
BernardLanguillier
Sr. Member
****
Online Online

Posts: 7762



WWW
« Reply #36 on: August 12, 2010, 03:32:25 PM »
ReplyReply

Canon are significantly ahead of Nikon noise-wise.
http://www.clarkvision.com/articles/evalua...1div/index.html

The article you point to is not a comparison. It only states that the 1DIV has low read noise.

Back to the point, it would be great if Canon delivered a CMOS based 50+ megapixel rugged MF camera. But actually the 645D already is all of that, it just lacks live view.

For those photographers thinking that live view is not relevant, there is no reason to wait for Canon/Nikon to enter MF, the Pentax 645D already is all we hoped these mainstream brands would bring to MF in terms of reliability, price and usability in the field.

I actually don't quite understand why 35mm users would have to care much about buying MF from the same brand. True some sensors are better than others (if anything Nikon leads in DR and system noise right now), but overall the absolute level is excellent accross the board. We can expect the next generation of Canon sensor to have catched up 2 years later, which would mean MF sensors levels of DR.

Cheers,
Bernard
Logged

A few images online here!
ondebanks
Sr. Member
****
Offline Offline

Posts: 820



« Reply #37 on: August 13, 2010, 08:56:36 AM »
ReplyReply

The article you point to is not a comparison. It only states that the 1DIV has low read noise.

Cheers,
Bernard


True, Bernard; my case was incomplete. I should also have linked to this:
http://www.clarkvision.com/articles/digital.sensor.performance.summary/index.html
...scroll down to Figure 3 and see where the Nikon's best (D3 & D300) are with respect to a whole slew of the Canons (50D, 7D, 1DIV, 5DII). Heck, even the old 350D is better than them!
Figure 11 is perhaps fairer as it normalizes for pixel size differences (using readnoise/mm^2), but the Nikons still lag a bit.

Now the D3x is something else (it's not measured there) but I gather its real strength is noise at _low_ ISO, thanks to a clever system to reduce the ADC noise contribution which dominates readnoise at low ISOs in all cameras. This plot is about sensor readnoise, which dominates at higher ISOs, and is the more important fundamental characteristic (plug Nikon's best ADC into one of Canon's sensors and you might have the best of both worlds). Does anyone know the sensor readnoise values of the D3x and D3s? I'd be astonished if they're around or under 1.7 electrons, but I have humble pie standing by in the fridge if necessary. However, based on results like http://forums.dpreview.com/forums/read.asp?forum=1021&message=33852632, I don't think I'll be needing it.
 
Logged
williamrohr
Full Member
***
Offline Offline

Posts: 100


WWW
« Reply #38 on: August 13, 2010, 10:12:36 AM »
ReplyReply

Once you handle and shoot a Leica S2, you can sense where Canon might go.  It has the fire power of a medium format in a camera not much different in size and weight than the 1Ds models .... and with the "hand holdability" of an SLR.  Of course, the difference between "might" and "will" is a wide chasm.  Smiley
Logged
ziocan
Sr. Member
****
Offline Offline

Posts: 427


« Reply #39 on: August 14, 2010, 12:28:59 AM »
ReplyReply

It seems that all three of us agree then: with a larger format the diffraction effects kicks in at a somewhat higher aperture ratio, (about one stop higher for MF than 35mm) ... which balances the different DOF at equal f-stop and equal FOV in different formats to give equal DOF at the "diffraction limited aperture".

But perhaps I am misunderstanding your comment, so let me ask what display and viewing conditions you are using when making these judgements of visible diffraction?
- Prints or onscreen viewing at equal PPI? (e.g. 100% pixels on-screen)
- Prints or onscreen viewing of the image at the same displayed size (maybe on-screen viewing of a crop to the same portion of the image)
- Other?
Yes, it seems that we agree.

As for answering to your questions, I do make those assessments on screen at 100% or 66% and on print at about A3 size, even if the final media will be smaller than that.

I have to say that when I need to shoot  at f16 or more, with the a900, I can get rid of the diffraction softness, just increasing sharpening to the point that on screen at 100% looks quite evident. But since my final output is normally full bleed magazine size, wich is about 9" by 12", the image looks fine on print. All the artifacts and little stair steps that may appear on diagonals, just do not show on print and the image simply look sharper without showing any deterioration.
On the printed magazines, I can see the difference of "crispiness" from the times that I did not use to sharpen the images as much.
I have found that the sweetest spot for the a900 is between f5.6 and 8 (nothing new I know...), but sometime is not enough and I need to go f 11 and more.

If I work with the P30, I normally do not use any sharpening at all. Though the P30 images do not look much better than the a900, a little but not much.






Logged
Pages: « 1 [2] 3 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad