D700 vs. M9, high iso.

[ulrikft]

M9 superior to the D700 at 3200 ISO??! What are you smoking?

(Sorry, couldn't resist - no offense)

Jarle

It is superior, when it comes to detail, and it has more noise. As I have said quite a few times now. The M9 retains more detail at high iso, than the d700, at the cost of more noise. I have printed A3+ versions of the same shot with D700 at iso3200 and M9 pushed to iso3200, you are of course welcome to Oslo to check them out 🙂

 

[Pickett Wilson]

There was a long thread once here on the M8 pushed to astronomical ISO's (I think we saw images supposedly at ISO 12,500) "proving" that he M8 was better than any other camera at high ISO. There were also endless posts all over the web "proving" that the M8 had higher resolution and the sensor delivered sharper images than any DSLR and some medium format digital backs.

Trying to get objective conclusions about Leica digital M's is simply an exercise in frustration. There is little objectivity, it seems, when talking about Leicas. These threads simply go in circles.

 

[ulrikft]

There was a long thread once here on the M8 pushed to astronomical ISO's (I think we saw images supposedly at ISO 12,500) "proving" that he M8 was better than any other camera at high ISO. There were also endless posts all over the web "proving" that the M8 had higher resolution and the sensor delivered sharper images than any DSLR and some medium format digital backs.

Trying to get objective conclusions about Leica digital M's is simply an exercise in frustration. There is little objectivity, it seems, when talking about Leicas. These threads simply go in circles.

If this is directed at me, I'm a very leica m8-sceptic person, and I was very, very sceptical towards the M9 until I got the chance to borrow one and test it. It has it's share of flaws indeed, but if you compare the m9 with the d700 with equal framing at iso 80-3200, the M9 will deliver both more noise AND more detail. That is, as close I can get to an objective fact, without having multiple M9s, multiple lenses, multiple d700s and multiple lenses and a far more rigid test setup. The M9 has more noise, especially color noise, which is not too hard to clean up, the D700 has a more smooth look, but then again less detail too. The M8 never was a very good low light camera, I tested that too, and realized that even at 1250 you get close to the limit for iso noise, but the M9 is, as far as I can tell from testing, close to/just behind the class leading d700 in this area.

I would love to borrow a 5d MKII and test that against the M9 too, to see how two high resolving bodies compete.

 

[Pickett Wilson]

ulrikft, my point is that it doesn't matter. Any of these cameras, the M9, D700, D3x, 5D, 5D MkII, all of them will deliver image quality beyond the needs of 99 percent of the people using them and beyond the ability of 99 percent of the people looking at the images to perceive the difference at any size they are likely to be printed.

I don't think, because of the emotions involved, that any objective tests are really going to change minds. If you must own a Leica, you are going to own a Leica. (I own a number of Leicas, and think they are great cameras) When the M8 was the only game in town, it was portrayed by many owners as the "best camera in the world."

 

[Jarle Aasland]

I don't think, because of the emotions involved, that any objective tests are really going to change minds.

Well said. My feelings exactly. Not that it matters much - let people decide for themselves. If someone is happy with their 5000 ISO M9 photos - good for them.

Jarle

 

[skimmel]

It is superior, when it comes to detail, and it has more noise. As I have said quite a few times now. The M9 retains more detail at high iso, than the d700, at the cost of more noise. I have printed A3+ versions of the same shot with D700 at iso3200 and M9 pushed to iso3200, you are of course welcome to Oslo to check them out 🙂

Sorry if you've discussed this already, but, in print, what do you see comparing these two?

 

[ulrikft]

Sorry if you've discussed this already, but, in print, what do you see comparing these two?

Hey!

I see more detail and more color noise in the M9 shot, less detail, but also less color noise in the D700. Held at an arms lenght, the most striking difference is the more saturated colors of the M9, not the noise. Both files print very well up to A3+ at iso 3200, absolutely usable. I would have no worries selling portraits or concert shots pushed to 3200 with the M9

 

[KM-25]

I will say what I said again. I spent over two hours shooting the M9 using my 28 cron, 35 and 50 lux Asph lenses and then my D700 with the Carl Zeiss 35/2.

The level of noise was far better on the D700 at 1,600 and above and the detail, color saturation and contrast on the D700 with that lens about equal to the 35 lux until above ISO 1,600 when it simply ran away with it while the M9 looked blotchy.

These were not tests, this was professional photography. The only person you are testing for and trying to convince is you, the rest of us have different methods, needs and levels of expertise.

I want an M9, but I have other priorities, like 4 Leica M bodies loaded with Kodachrome. If the M9 had been a little better in terms of noise, I would have sold my D3 and happily used the M9 along with a D700 for paid digital work.

Tests can only do so much and in your case, what you are seeing is not what a lot of us are seeing, so just keep that in mind as you seem to make tests top priority instead of meaningful photography....

 

[ulrikft]

I will say what I said again. I spent over two hours shooting the M9 using my 28 cron, 35 and 50 lux Asph lenses and then my D700 with the Carl Zeiss 35/2.

The level of noise was far better on the D700 at 1,600 and above and the detail, color saturation and contrast on the D700 with that lens about equal to the 35 lux until above ISO 1,600 when it simply ran away with it while the M9 looked blotchy.

These were not tests, this was professional photography. The only person you are testing for and trying to convince is you, the rest of us have different methods, needs and levels of expertise.

I want an M9, but I have other priorities, like 4 Leica M bodies loaded with Kodachrome. If the M9 had been a little better in terms of noise, I would have sold my D3 and happily used the M9 along with a D700 for paid digital work.

Tests can only do so much and in your case, what you are seeing is not what a lot of us are seeing, so just keep that in mind as you seem to make tests top priority instead of meaningful photography....

Condecending if nothing else.. that is great. Of course i don't do meaningful photography, that is your domain o-great one!

If you want to test quantifiable things, tests are what you have to use. Level of detail and level of noise, those are things you have to use tests to determine, quantifiable, where are. The Leica have far more detail than the d700 at base iso, both beacause it is a higher resolving sensor, and beacause it do not have a AA-filter. This is quite easily shown. At higher iso, the d700 has less noise, but also less detail, to about iso5000-6400. What you are seeing, i can not control, but I can control test methodology, and the "science is stupid"-attitude some people seem to have, is great, just don't apply it in the territory of science. What am i trying to say? If you think an image is great, noise or no noise, fantastic! But then don't participate in a technical debate about noise levels, beacause you are in the wrong thread.

 

[Richard Marks]

Condecending if nothing else.. that is great. Of course i don't do meaningful photography, that is your domain o-great one!

If you want to test quantifiable things, tests are what you have to use. Level of detail and level of noise, those are things you have to use tests to determine, quantifiable, where are. The Leica have far more detail than the d700 at base iso, both beacause it is a higher resolving sensor, and beacause it do not have a AA-filter. This is quite easily shown. At higher iso, the d700 has less noise, but also less detail, to about iso5000-6400. What you are seeing, i can not control, but I can control test methodology, and the "science is stupid"-attitude some people seem to have, is great, just don't apply it in the territory of science. What am i trying to say? If you think an image is great, noise or no noise, fantastic! But then don't participate in a technical debate about noise levels, beacause you are in the wrong thread.

If you are talking science (which you are attempting to do) you actually need a randomised blinded trial in which the observers do not know which images are from which camera. The images would then be scored by an independent panel of observers for noise (defined in terms of colour / shadow areas highlights / what ever is pre agreed at the outset) and detail also pre defined and perhaps a final category for colour accuracy. This is how i would aim to achieve an objective result experimantally. Ideally a good range of images with different exposures and different colours and scenes and also a range of ISO's.
You could add an additional number of very noisy images and some noisless film prints to callibrate the sensitivity of each of the assessors and their biass could then be adjusted for in the final markings.

Its tricky to do but it would be a worthwhile excercise and at least it might produce something of publisheable value rather than all this flannel.

I am not sure how many of the big camera manufacturers would openly volunteer for a randomised blinded evaluation but this is probably what is needed. It may well reveal some surpising results. Anything else is so frought with brand loyalty biass as to be ultimately meaningless.

Richard

 

[JNewell]

These were not tests, this was professional photography. The only person you are testing for and trying to convince is you, the rest of us have different methods, needs and levels of expertise.

If you want to test quantifiable things, tests are what you have to use. Level of detail and level of noise, those are things you have to use tests to determine, quantifiable, where are. The Leica have far more detail than the d700 at base iso, both beacause it is a higher resolving sensor, and beacause it do not have a AA-filter. This is quite easily shown. At higher iso, the d700 has less noise, but also less detail, to about iso5000-6400. What you are seeing, i can not control, but I can control test methodology, and the "science is stupid"-attitude some people seem to have, is great, just don't apply it in the territory of science. What am i trying to say? If you think an image is great, noise or no noise, fantastic! But then don't participate in a technical debate about noise levels, beacause you are in the wrong thread.

Maybe, but at the end of the day, popular trends on the internet notwithstanding, it's about real life images, not test results, right? In my own experience, I've shot test targets of various kinds and come to one set of conclusions. Then I've taken the lenses out into real situations and taken real, practical, everyday pictures and gotten different results. I am *definitely* not saying that "science is stupid," but at the end of the day, real world pictures are what matters to me.

 

[ulrikft]

If you are talking science (which you are attempting to do) you actually need a randomised blinded trial in which the observers do not know which images are from which camera. The images would then be scored by an independent panel of observers for noise (defined in terms of colour / shadow areas highlights / what ever is pre agreed at the outset) and detail also pre defined and perhaps a final category for colour accuracy. This is how i would aim to achieve an objective result experimantally. Ideally a good range of images with different exposures and different colours and scenes and also a range of ISO's.
You could add an additional number of very noisy images and some noisless film prints to callibrate the sensitivity of each of the assessors and their biass could then be adjusted for in the final markings.

Its tricky to do but it would be a worthwhile excercise and at least it might produce something of publisheable value rather than all this flannel.

I am not sure how many of the big camera manufacturers would openly volunteer for a randomised blinded evaluation but this is probably what is needed. It may well reveal some surpising results. Anything else is so frought with brand loyalty biass as to be ultimately meaningless.

Richard

As said, I could have done with 10-15 M9s, 10-15 5dMkIIs and 10-15 d700s, and of course, if I could randomize it all, it would do things evenb etter. But this is as close, in methodology one can come alone. But also as said, I have learned that people here don't like tests... good to know. I won't be posting any further results with loaner here, done deal.

 

[Richard Marks]

As said, I could have done with 10-15 M9s, 10-15 5dMkIIs and 10-15 d700s, and of course, if I could randomize it all, it would do things evenb etter. But this is as close, in methodology one can come alone. But also as said, I have learned that people here don't like tests... good to know. I won't be posting any further results with loaner here, done deal.

Dear Ulrik

Its not that people do not like tests. They simply do not agree with your conclusions. You can not walk off just because you do not like what people are saying. You have attempted an objective test but clearly if you have just shelled out £5,000 you must be at least ever so slightly biassed? (love is blind 🙂) Equally just because some one else does not agree with you does not mean you have chosen a poor camera. If you are happy thats fine. You do not need the support of those to whom this has been posted. But be prepared for people here to question your methodology if you make such a conslusive statement when quite honestly the images you have shown are at best equivocal in my opnion. I am not being condescending just objective about the evidence you have presented.

best wishes

Richard

 

[ulrikft]

Dear Ulrik

Its not that people do not like tests. They simply do not agree with your conclusions. You can not walk off just because you do not like what people are saying. You have attempted an objective test but clearly if you have just shelled out £5,000 you must be at least ever so slightly biassed? (love is blind 🙂) Equally just because some one else does not agree with you does not mean you have chosen a poor camera. If you are happy thats fine. You do not need the support of those to whom this has been posted. But be prepared for people here to question your methodology if you make such a conslusive statement when quite honestly the images you have shown are at best equivocal in my opnion. I am not being condescending just objective about the evidence you have presented.

best wishes

Richard

I have, as said for the third time, not bought the M9, neither will i before it costs far less, I have _loaned_ a camera I was very sceptical about, to test it out.

When it coms to the evidence, discuss this, instead of implying that I love the M9 and therefore, I am biased. Weirdly enough, the forum where I ususally hang out, where 99% of the user group are SLR users, and the crowd is highly negative towards leica beacause of the M8, people are far more positive about the M9 high iso results, but here? People seem to _want_ the M9 to be a bad iso achiever.. and they seem to ignore evidence, not interpret it differently, just ignore it..

 

[Richard Marks]

I have, as said for the third time, not bought the M9, neither will i before it costs far less, I have _loaned_ a camera I was very sceptical about, to test it out.

When it coms to the evidence, discuss this, instead of implying that I love the M9 and therefore, I am biased. Weirdly enough, the forum where I ususally hang out, where 99% of the user group are SLR users, and the crowd is highly negative towards leica beacause of the M8, people are far more positive about the M9 high iso results, but here? People seem to _want_ the M9 to be a bad iso achiever.. and they seem to ignore evidence, not interpret it differently, just ignore it..

Dear Ulrik

You have attempted to perfom an objective test. People have given you their opinion and you do not agree with it. You then acuse posters of being negative about the M9.

If you think that I imply that you "love the M9 and are therefore biassed" please remember you were the one who said earlier in this thread that you were "drooling" with reference to the advatages of the M9. Your words not mine. Do not therefore be suprised if people do not take your test too seriously.

Best wishes

Richard

 

[ulrikft]

Dear Ulrik

You have attempted to perfom an objective test. People have given you their opinion and you do not agree with it. You then acuse posters of being negative about the M9.

If you think that I imply that you "love the M9 and are therefore biassed" please remember you were the one who said earlier in this thread that you were "drooling" with reference to the advatages of the M9. Your words not mine. Do not therefore be suprised if people do not take your test too seriously.

Best wishes

Richard

If people were actually not agreeing with conclusions, and based this on reality, that would be all fine and good, but it seems to be more important to make snide comments, imply different things and do anything but*argue in polite, mature and rational fashion. If nothing else, you guys sure know how to chase someone away from your forum 🙂

 

[Richard Marks]

If people were actually not agreeing with conclusions, and based this on reality, that would be all fine and good, but it seems to be more important to make snide comments, imply different things and do anything but*argue in polite, mature and rational fashion. If nothing else, you guys sure know how to chase someone away from your forum 🙂

Dear Ulrik

You have your reality. Do not expect others to necessarily share the same.
Posters can agree or disagree that is their privelidge.
It does not mean they are personally after you!
If I may be constructive, perhaps a post with a slightly less attention seeking title than "D700 vs M9 high Iso" may be a little less conspicuous for your return visit with the S2 vs M9 evaluation!

Good luck if you choose an M9, I hope you enjoy it.

Best wishes

Richard

 

[Thardy]

I understand what he is saying. At 18mp the M9 will have more resolution and more noise. That is the take home point.

How can 12mp have more resolution than 18?

 

[Richard Marks]

Resolving power is lost where there is noise.
Noise is pixels registering a signal that has not come from the intended image. Some noise may not affect the important aspects of an image such as in shadow areas. However resolution will decrease as a function of increasing noise if all other factors are equal.

Increasing the number of pixels should increase resolution but this assumes other factors such as lens resolving power have not been reached. 12 mp full frame is likely to resolve the order of 200 line pairs / cm and this is above most 35 mm format lenses at the edges. Hence the benefit of increasing pixels may not add resolving power but can add more noise.

Richard

 

[gnarayan]

However resolution will decrease as a function of increasing noise if all other factors are equal.

And this is my issue with the OPs comparisons. All his high ISO comparisons for noise and resolution are with the D700 uprezzed to 18MP. I'm not sure what resolution comparisons you can make doing that to begin with. After that they are using different lenses. I've no idea if he did this on a tripod or handheld. The shots of the boxes are all very well but I'm not sure how you can say anything about resolution with that. Not all other factors are equal.

I'd be curious to see how the M9 does against a 5D2. Dpreview will have more through tests in a few weeks.

Meh. They will be bandided about as explanations of why the m9 is the best or worst camera ever, regardless of what the tests actually show.

Cheers,
-Gautham