High ISO Performance

[Ed V]

Just curious what you think of the M5's performance at ISOs of 1600 and above?

I shot some street yesterday at ISO 1600 and the images seemed pretty noisy... useable but pretty noisy (compared to what I get out of my 5D3 at the same ISO). I've heard some trumpet the performance up to ISO 3200 and even 6400 but I'm not sure, in my admittedly limited experience with this camera, that I would be happy with the noise levels at those ISOs.

Your thoughts and experiences...

Ed

 

[JMZawodny]

Play around a bit here and see for yourself. To my eye it is comparable to the 7D2, a stop worse than the 5D2, and 2 stops behind the 1Dx2.

https://www.dpreview.com/reviews/image-comparison?attr18=daylight&attr13_0=canon_eosm5&attr13_1=canon_eos7dii&attr13_2=canon_eos5dmkii&attr13_3=canon_eos1dxii&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=1600&attr16_1=1600&attr16_2=3200&attr16_3=6400&normalization=full&widget=1&x=0&y=0

 

[dak723]

Just curious what you think of the M5's performance at ISOs of 1600 and above?

I shot some street yesterday at ISO 1600 and the images seemed pretty noisy... useable but pretty noisy (compared to what I get out of my 5D3 at the same ISO).

Your 5D3 is Full Frame. No way the M5 can compete with that.

 

[Ed V]

Just curious what you think of the M5's performance at ISOs of 1600 and above?

I shot some street yesterday at ISO 1600 and the images seemed pretty noisy... useable but pretty noisy (compared to what I get out of my 5D3 at the same ISO).

Your 5D3 is Full Frame. No way the M5 can compete with that.

I understand that. I'm not expecting the M5 to compare to the 5D3, the same way I am not expecting the 15-45 kit lens to compare to my L-glass. But at the same time the 5D3 is my current point of reference.

Interestingly enough, if I had to compare the noise at 1600 to anything I've used, I would have to go back before the 5D3 to the Leica M9. Great camera but the noise level at 1600 was not to my liking. One of the reasons I de-Leica'd and returned to Canon a year ago. What I see noisewise in the M5 is similar to what I would see at 1600 with that Leica.

I guess I was just responding to reviews I've seen that say M5 noise is good at 1600 and 3200 and at 6400 is still usable. Guess it's a matter of what you are comparing it to... and what one's result expectations are.

 

[serendipidy]

Hi Ed V
I recently got a 7D2 to replace my 7D. I took a bunch of BIF shots in low light and I thought the IQ above ISO 1600 was not to my liking. I basically never apply noise reduction (mostly because I don't do a good job of it). JMZawodny stated he thought the M5 was comparable to the 7D2 as to high ISO IQ performance.

In another thread, my opinion of the 7D2 performance above ISO 1600 got this reply

"I find comments like this interesting because I was watching Andy Rouse's review of the 1Dx2 where he took a shot of an otter at 50,000 ISO and said it was sellable. That is a professional talking. Someone who relies on image quality to make a living. And the 1Dx2 has maybe 2 stops benefit over a camera like the 7D2.
Many comments about 'ISO above 1600 being too noisy' are often from amateurs - I am not disparaging the standards that such people may set (it is, after all, a personal opinion), but I sometimes wonder if the ability to zoom in at 100% and criticise any noise that they see means that people are losing sight of the important thing. That is, the image. You only need to look at the comments on what are often quite apalling shots (oof, poorly composed etc) to realise that the vast majority of people out there do not give a monkey's fart about image noise."

 

[Ed V]

In another thread, my opinion of the 7D2 performance above ISO 1600 got this reply

"I find comments like this interesting because I was watching Andy Rouse's review of the 1Dx2 where he took a shot of an otter at 50,000 ISO and said it was sellable. That is a professional talking. Someone who relies on image quality to make a living. And the 1Dx2 has maybe 2 stops benefit over a camera like the 7D2.
Many comments about 'ISO above 1600 being too noisy' are often from amateurs - I am not disparaging the standards that such people may set (it is, after all, a personal opinion), but I sometimes wonder if the ability to zoom in at 100% and criticise any noise that they see means that people are losing sight of the important thing. That is, the image. You only need to look at the comments on what are often quite apalling shots (oof, poorly composed etc) to realise that the vast majority of people out there do not give a monkey's fart about image noise."

Let me begin by saying I am not here to disparage anyone or anything. Rather I am just raising a question. Anyway to say a photo is sellable is kinda like saying a photo is acceptable. That said, I think my images shot at ISO 1600 are acceptable and sellable. All I need is a buyer. Heck I have some indoor shots at ISO 6400 that are acceptable. I just like a bit less noise in my digital photographs. (Note I don't have as much of an issue with grain in my film photographs.)

I was thinking last night that I will print a couple of the photos shot at ISO 1600 in sizes up to 12x18 or even 16x24 and see how they look. Heck I had a grainy 35mm film photo win a competition and be printed on a 14' x 44' (yes, feet) billboard and posted for a year. It looked great from the road. I just wouldn't want to look at it from a foot away!

Anyway, I found your post interesting and though-provoking. And that's all I can ask. Thanks.

Ed

 

[ajfotofilmagem]

The problem is that high megapixel cameras have LESS sharpening per pixel, and when viewed at 100% give a misleading perception that the ENTIRE PICTURE would be less sharp.

I know the difference between 22 to 24 megapixel is small, but in addition to the fact that the APS-C sensor has its PHISICALLY larger image on the computer screen (compared to the physical size on the sensor), it will put the M5 at a disadvantage .

If you do a controlled test, with the same lens, lighting, object, etc. being equal, and cut the two images to the same output size as 15x10 centimeter for example, the advantage of 5D Mark iii over M5 will decrease. In this case you will realize that ISO1600 in M5 looks very much like ISO3200 in 5D Mark iii. This is really expected, with a sensor with more than double the area of light capture.

 

[JMZawodny]

Let me begin by saying I am not here to disparage anyone or anything. Rather I am just raising a question. Anyway to say a photo is sellable is kinda like saying a photo is acceptable. That said, I think my images shot at ISO 1600 are acceptable and sellable. All I need is a buyer. Heck I have some indoor shots at ISO 6400 that are acceptable. I just like a bit less noise in my digital photographs.

Do you use noise reduction software in your post processing? I'm a big fan of DXO Prime NR and it allowed me to shoot routinely at ISO 6400 with my 5D2 or ISO 3200 on the 7D2 and get acceptably clean imagery. Some of the NR SW available these days is nothing short of amazing at reducing noise while preserving detail.

 

[hclarkx]

Just curious what you think of the M5's performance at ISOs of 1600 and above?

I shot some street yesterday at ISO 1600 and the images seemed pretty noisy... useable but pretty noisy (compared to what I get out of my 5D3 at the same ISO). I've heard some trumpet the performance up to ISO 3200 and even 6400 but I'm not sure, in my admittedly limited experience with this camera, that I would be happy with the noise levels at those ISOs.

Your thoughts and experiences...

Ed

Do you "shoot to the right" when you can? The M5 makes this easy with the histogram in the viewfinder shortly after releasing the shutter. Run the compensation up and re-take until the whites or brightest elements in the frame are near the right edge and you've increased exposure of the shadows to some extent thereby reducing noise. Shifting the image to the left in post to compensate for shooting to the right will not increase noise as does bringing up the shadows in a "properly" exposed shot. Some noise and dynamic range critiques assume this is done to get the most out of the sensor.

 

[Mt Spokane Photography]

Everyone has a different tolerance for noise. I use my 5D MK III at 25800 and print or show small photos on the internet that may look awful full size. I'd say the M5 is comparable to the 80D, individual cameras may vary slightly.

Here is one from the other night at 25,800 with my 5D MK III. Its just a trial photo with the teacher and the students getting ready for the dress rehearsal. I used it to setup my camera.