Leica M8 Noise vs. Canon 5D

[lovcom_photo]

It seems Leica has missed the boat on high ISO noise reduction. Surely you

Leica fans with well heeled wallets would expect better, yea?

 

But you be the judge....see link:

 

http://www.pbase.com/tinamanley/comparison&page=all

[roger_smith4]

Why do the Leica shots look way darker? Were they underexposed?

[blakley]

Yeah, two bad pictures posted on the Internet with some almost-equally-bad pictures from a Canon camera already widely known to be capable of much, much better is pretty conclusive. Must be a complete screw-up. Can't imagine why my results don't look that bad.<div></div>

[blakley]

The discussion <a href="http://www.leica-camera-user.com/digital-forum/10965-m8-canon-5d-comparison-pix-high.html">here</a> also speculates that underexposure is part of the problem with these photos. Without a meter reading, of course, it's difficult to know.

[andy_piper2]

No question the 5D with its big juicy 8-micron pixels (if I did the math right) comes out

way ahead of the M8 (or the Nikon D2X) with 5.8-micron pixels. Real estate still rules

when it comes to light-gathering, and the 5D has about twice the area per pixel.

 

it will be interesting to see which arrives first: A Canon digital RF with a 5D sensor - or a

Leica M9 with a comparable sensor.

 

The M8 has a lot of color noise, so it improves a bit in B&W (whether jpeg or dng - the

former has more noise reduction but lower res, the latter has more noise but stays a bit

sharper).

 

But Tina's samples are a fair representation of reality.

[steve_swinehart]

I don't quite get why this is so earth shaking. You couldn't get over about ISO 1600 with film - and that looked bad.

 

If you shoot the camera differently, it works just fine. I just finished an expiriment with the M8 outdoors using moonlight at ISO 320. The images have little noise and the darkframe subtraction was flawless.

 

What you're seeing in these photos is under exposure. As a test, I went from 30 seconds down to 1/2 second under moonlight at f/2.0. Under 2 seconds the exposure is noisy - 2 seconds and over the exposures had little if any noise.

 

But, I suppose if you just really have to find something wrong with the camera you will.

 

I once saw a Rolls Royce being towed by a wrecker. I immediately made the judgement that all Rolls - past, present, future - were junk and made up my mind I'd never own one.

 

I suppose based on one poorly done "test" - you could make the same judgement about the M8.

[ray .]

I'm waiting for the day when someone who can actually make interesting photographs

does these tests. They don't, because essentially the whole issue is useless

nitpicking.

The testers often aren't even technically proficient. Look at dpreview as an example.

Sample photographs from the 5D or whatever cam with hardly an ounce of soul in the

first place and then on top of that blown highlights running wild. There's nothing

that says that a bit more noise or a little less sharpness is intrinsically

harmful to

the end result- the big picture, if you will. I've seen great photographs that weren't

particularly sharp that blow

away 99% of what gets posted here. I include commercial work on this issue. There

are high ticket wedding photographers using Holgas to make wedding pics. This

endless obsession with miniscule differences in camera performance is a distraction,

or for some, a whole world in and of itself- that really has very very little to do with

photography. It has more to do with commerce related to photographic equipment.

[torben_daltoft]

'But Tina's samples are a fair representation of reality.'

Whauw!!

[lovcom_photo]

The under exposure of the Leica tests may well be true. But I would image that if the Leica exposures were compensated by 1/3, or 1.2 stop I doubt all that noise would be mitigated....it is just too much noise. Also, if the 5D images were under exposed by 1/3, or 1/2 stop I seriously doubt it's pictures would be so noisy. The example B&W picture above in this thread is not a good representation of what the Leica can do at 1250 ISO because it was shot or converted to B&W, that alone will mitigate the noise...try doing that same shot in color...I suspect it will be a lot noisier.

 

Instead of the Leica fanboys coming to the defense of their beloved M8, it would be better for them, of all people to be the most critical of all.

 

But that is not the case...sadly.

 

At ISO 1250 or faster, I strongly expect even the lowly $800 Canon XTi entry-level DSLR to also do better then the M8.

 

How come some of you Leica fans are actually apologists for the M8? If it were me that spent that big $$ on an M8, I would be furious at Leica...or better yet, at myself for having dropped the big $$ on it.

 

I assure you that if my 5D produced that noise I would be furious at Canon. It's best to be and remain objective.

[blakley]

I get noise at 1250 - not like Tina Manley, but noise nevertheless. I get magenta shift, and I get banding. I don't like magenta shift and banding, and I'll get the camera fixed and use filters; I've already spoken to Leica, I've registered for the free filters and am waiting for instructions on how to send the camera in to get it fixed.

<p>

But here's the thing. You think I should complain about noise at ISO 1250, because the Leica has more of it than a Canon 5D. I knew the camera had more noise than a 5D before I bought it. It doesn't bother me, because I don't use 1250. I didn't use 1250 on film - I didn't even use 640 on film, and almost all of my pictures are taken in really low light. So I don't complain about noise at 1250, because I'm getting LESS noise with the M8 at 640 than I'm getting with 400-speed film at 400 in the same light. Go figure. I'm also getting what I consider to be better quality prints than what I've seen from 5D in prints - which is of course not a fair comparison, because I don't have a 5D, don't make prints from one, and don't know how good at printing the folks whose prints I've seen are.

<p>

I'm getting excellent pictures with the M8. If you aren't, you should complain. Maybe you're using it for things it doesn't do well, and if so you should probably send it back and get a camera which does those things better. The things I'm using it for it does very very well indeed - so I'm not complaining.

<p>

<a href=" title="Photo Sharing"><img src="http://static.flickr.com/121/313374369_295251d4ec.jpg" width="403" height="500" alt="ad astra" /></a>

[lovcom_photo]

Bob, could it be you got conditioned over the years not to use ISO 640 and faster film because of the horrible grain?

 

If your M8 provides less noisy images at ISO 800 then film grain at ISO 800, perhaps it is now time for you to take advantage of the more usable high ISO images you can now take with your M8 that you would not dare with film.

 

I have the 5D and I get excellent prints, no less stunning then from your M8 at ISO 800 or slower...I have the best EOS mount lenses, and perhaps they're not as great as your Leica glass but many of them are very darned close.

 

If your M8 produces great ISO 800 images, then perhaps you should consider getting those really dark shots that you would not even think of getting with ISO 800 film....no reason to apply the high ISO film limitations to your M8, yea? Lift the limitations, expand your horizons.

 

But regardless if you never shoot faster then ISO 400...this fact still does not excuse the excessive noise of the M8 at ISO 1250 or faster. I mean, after all, it is a Leica right, and a Lieca is supposed to provide the best image quality regardless of the recording device used; be it film or a sensor.

[forum_shopper]

Troll

[blakley]

<i>Bob, could it be you got conditioned over the years not to use ISO 640 and faster film because of the horrible grain?</i>

<p>

Of course! Film and digital sensors are always best at the lowest possible ISO. What's your point?

<p>

<i>If your M8 provides less noisy images at ISO 800 then film grain at ISO 800, perhaps it is now time for you to take advantage of the more usable high ISO images you can now take with your M8 that you would not dare with film.</i>

<p>

Well, to begin with, I said nothing about ISO 800 film or about film grain. To continue, the M8 doesn't offer ISO 800 - it offers 160, 320, 640, 1250, and 2500. I am in fact using 640, because I get better quality with 640 on the M8 than I get with 400 on film - which I've already said. There are no "high-ISO images I would not dare with film" because there are no "high-ISO images". I often take pictures in dark conditions; I do it at what I consider to be moderate ISO settings, which is one reason I love Leica: I can get lenses which are still spectactularly good when opened up to f/1.4 and f/1.2.

<p>

<i>I have the 5D and I get excellent prints, no less stunning then from your M8 at ISO 800 or slower</i>

<p>

I have no idea how you can make this statement, since you've clearly never used <i>any</i> M8, much less mine. I'm sure your prints are great. The 5D is a fantastic camera.

<p>

<i>...I have the best EOS mount lenses, and perhaps they're not as great as your Leica glass</i>

<p>

They're not.

<p>

<i>but many of them are very darned close.</i>

<p>

Yes, they are. The 85/1.2 L, for example, is a really fine lens.

<p>

<i>If your M8 produces great ISO 800 images,</i>

<p>

As I've already noted, it produces <i>no</i> ISO 800 images.

<p>

<i>then perhaps you should consider getting those really dark shots that you would not even think of getting with ISO 800 film....no reason to apply the high ISO film limitations to your M8, yea? Lift the limitations, expand your horizons.</i>

<p>

As I've already noted, there are no such shots. I'll be very surprised if you shoot in light darker than my average. 1/15 at f/1.2 on ISO 640 is quite a normal experience for me. If you're suffering from the delusion that you need high ISO to shoot in low light, check out <a href="http://www.flickr.com/photos/tommyoshima/">Tommy Oshima</a>. He shoots VERY dark night scenes in Tokyo and elsewhere with his Noct on 100 speed SLIDE film.

<p>

<i>But regardless if you never shoot faster then ISO 400...this fact still does not excuse the excessive noise of the M8 at ISO 1250 or faster.</i>

<p>

OK, so let me get this straight. If Leica offered ISO 12,000 you'd require an "excuse" for "excessive" noise? You seem to be defining "excessive" as "not better than the current best performance by ANY camera" - which is what the 5D offers.

<p>

<i>I mean, after all, it is a Leica right, and a Lieca is supposed to provide the best image quality regardless of the recording device used; be it film or a sensor.</i>

<p>

The Leica is supposed to enable the photographer to produce excellent pictures. The M8 does. It is not required to make you or me unhappy about all other cameras on the earth.

[john kuo]

Noise isn't everything, besides I tend to believe most M8 buyers are not really comparison shopping the 5D, so such a test is really irrelevant. I'm glad you enjoy your 5D.

[Brad_]

<I>I'm waiting for the day when someone who can actually make interesting photographs

does these tests. ...</i><P>

 

What Ray said. I'll take unsharp pix that are <I>interesting</i> any day. This is really

tedious...

[noah_c]

I love comparing a $4800 camera body against a $2500 camera body. Yummy.

[andrew robertson]

"Can't imagine why my results don't look that bad."

 

More likely than not because you are converting to black and white, where noise shows up

much less strongly. I often shoot my Epson R-D1 at ISO 1600 and underexpose by 4 or

more stops (ISO 25600+!) and when converted to black and white the noise is difficult to

detect. Of course, detail suffers at that speed, but you won't see awful chunky color vomit

noise in a monochrome image.

[andrew robertson]

"I'm waiting for the day when someone who can actually make interesting photographs

does these tests. They don't, because essentially the whole issue is useless nitpicking. The

testers often aren't even technically proficient. Look at dpreview as an example. "

 

Dpreview is the only site on the internet that performs useful, standardized tests which

one can easily use to compare different cameras. Artistic sensibility is the least important

skill one could have when designing a useful test.

 

It is difficult or nigh on impossible to make any useful conclusions about a lens, camera,

film, sensor, or anything when looking at semi-random shots of different subjects etc.

You need the images to be as identically framed and lit and shot as can be to be able to

compare anything.

[andy_piper2]

Well, OK. Tina got what she got and I'll withdraw any comments on her specific results one

way or the other. Sean Reid duly noted the noise difference between the 5D and the M8 at

reidreviews.com.

 

Yes, comparison shots are usually boring. It's kinda hard to catch the same interesting

"moment" twice, even with one camera at 2 ISOs, let alone two different cameras.

 

Here's an ISO 1250 color shot from the M8 (.dng) Mixed daylight and indoor tungsten. 50

'cron. Full frame and crop. Definitely boring overall - I'll try to do better in future.<div></div>

[Karim Ghantous]

I don't see a problem at all here. It's like saying "ZOMG teh M8 doesnt h4v3 built-in noizeninj4/ne4timag3 11!111!11oneoneeleveneleven". Tools like Noise Ninja and Neat Image are brilliant (esp. on digital images) - even the free versions. Better let the camera do its work naturally and fix noise in post.

 

Is that such a problem? The 5D has its issues, too. Heaven forbid that people buy it!

[ray .]

The thing is Andy, once you get into actually taking pics, and I know you're capable of

it- you forget all this crap. If testing is a hobby for you, then fine, but let's say what

it is.

Some people just enjoy doing tests and endlessly pursuing theoretical optical

perfection. But

most get the camera they can afford or the one that looks cool, or the one that

handles the way they like. If it's up to basic modern standards, these miniscule

differences that you need a loupe to see don't matter.

 

The time I know I need to keep a camera or lens is once I take a real good photo with

it. Then I know it can do the job I want, isn't a dog and won't get in my way. Or if I

see someone else doing interesting

things with a certain camera, I might decide I want it too. But no amount of cold

testing is going to convince me of it.

 

You're a good shooter. Why don't you forget these samples you think show

characteristic detail or whatever and just shoot with the thing?

[StuartMoxham]

Wow Dan lucky you did not buy an M8. Be happy with your choice and leave others to be happy their's be it Leica, Nikon, Pentax, Sony or anything else that does not have the high ISO performance of the Canon 5D.

[frank]

<sight!>

Noise, noise, noise...

Is THAT the only issue that is relevant for amateurs?

In many years of using M-Leicas for indoor shots ISO400 films

were sufficient for me. I used a test sample of the M8 for three

weeks and I was happy with the results that were taken at

ISO640. So, why worry about ISO 1250 or ISO 2500?

 

BTW, Leica is famous for its lenses - not for its sensors. Guess

this won't change in the near future.

[len_smith]

These results are a matter of concern for people (like me) who have placed an order for the M8. I use a Canon EOS 5D and value the very low noise, even at high ISOs. I don't expect the M8 to be nearly as good, but these results are a disappointment.

 

However, what we have here is a sample of one. And Tina Manley's comments on another forum suggest that she is perplexed why her sample shows such a lot of noise at high ISOs when the M8 she was originally loaned for beta testing did not.

 

Reviews of the camera in "Amateur Photographer" and "British Journal of Photography" make no mention of such severe noise. The BJP review is particularly thorough, and gives the M8 very high praise.

 

Given the conflicting information, we should not place too much reliance on one report. I doubt very much that Tina Manley would wish us to condemn the M8 on the basis of one test, when other reviews are praising its excellence.

 

So I will not be cancelling my order just yet!

[john_r._fulton_jr.]

<I>I'm waiting for the day when someone who can actually make interesting photographs

does these tests.</I><br><P>Generally, they're out taking photographs and not doing

"tests".