Review: Canon EOS M5: \

[Woody]

interesting to see that M5 sensor performance is behind 80D.

The sensors may not be the same.

Just as the 77D and 80D sensors are not identical:
http://www.pocket-lint.com/review/140226-canon-eos-77d-preview-lucky-7s-for-the-80d-lite

"On paper the 77D's sensor looks to be the very same 24.2-megapixel one as you'll find in the 80D. That's not precisely the case, however, as it's a new construction, paired with the latest Digic 7 processor."

(The EOS-M5, 77D and 800D all use Digic 7 processor, while the 80D uses Digic 6 processor.)

 

[AvTvM]

interesting to see that M5 sensor performance is behind 80D.

The sensors may not be the same.
...
(The EOS-M5, 77D and 800D all use Digic 7 processor, while the 80D uses Digic 6 processor.)

Yes, sensors are obviously not identical. But is it really too much expecting some *improvement* in IQ after 2 years [80D -> M5] from *oh so innovative, industry-leading* Canon rather than a slight downgrade? Especially when a more powerful CPU is powering image pipeline and should be capable of running stronger/better algorithms on the image data?

 

[AlanF]

interesting to see that M5 sensor performance is behind 80D. as much as i dislike DXO generally, one would hope that the same test method was applied to both sensors and results are therefore comparable.

whether it is a full stop or not, i find it disappointing when a newer camera has less IQ/sensor performance than an older model. would be intetesting to get reasons/explanation for this "phenomenon". one of the questions, dpreview or other website staff could put in front of the Canon exec muppets when they "interview" them next time ... no matter if those guys would find such a straightforward, fact-based question horribly "disrespectful" or not!

Don't believe any measurement that is quoted without a standard error and standard deviation (or estimate of uncertainty) and you don't know how many times that measurement was made and whether it is on more than one copy of the camera. Lensrentals is believable because they do multiple measurments and quote the mean values, and sometimes have shown the deviations from the mean. Unless DxOmark gives an indication of the standard deviation of the measurements, I would not take the difference between two relatively similar numbers as being significant. JUst take DxOmark as giving a rough estimate of the properties of the sensor.

 

[AvTvM]

share most of points and reservations re. DXO.

However, I would expect, that copy-to-copy variation for image sensor performance is not as huge as for lenses [where many elements need to be perfectly adjusted and aligned for optimum performance].

So * under the assumption* that DXO used *exactly the same test method* for 80D and M5 ... I find it disappointing. Even when the difference may not be large or really visible in real luife .. it is on the wrong side. I would have liked to see at least a slight improvement ... as a "rough idea", that Canon is making some progress with sensor performance. After all, they still have some catch-up to do if they want to become "fully competitive" with - or better than! - Sony sensors.

 

[Duckman]

Does anyone know specifics on the M5's AEB abilities? Specifically, maximum number of bracketed shots and the EV range? (i.e. 5div can do 7 sequential bracketed shots +/- 3 EV)
...Anyone know if these spec numbers will carry over to the M6?

Three images with +/-2 stops in 1/3 stop increments, this can be 'added' to +/-3 stops of exposure compensation (again in 1/3 stop increments) too for a 10 stop spread in two bursts of three shots.

The M6 is the same.

Thank you private.
-J

 

[Sharlin]

But is it really too much expecting some *improvement* in IQ after 2 years [80D -> M5] from *oh so innovative, industry-leading* Canon rather than a slight downgrade? Especially when a more powerful CPU is powering image pipeline and should be capable of running stronger/better algorithms on the image data?

Two years? Last I checked there was less than half a year between the 80D and the M5 releases.

 

[AvTvM]

ah thx, you are right, it was not so long as i thought.

 

[rrcphoto]

interesting to see that M5 sensor performance is behind 80D. as much as i dislike DXO generally, one would hope that the same test method was applied to both sensors and results are therefore comparable.

whether it is a full stop or not, i find it disappointing when a newer camera has less IQ/sensor performance than an older model. would be intetesting to get reasons/explanation for this "phenomenon". one of the questions, dpreview or other website staff could put in front of the Canon exec muppets when they "interview" them next time ... no matter if those guys would find such a straightforward, fact-based question horribly "disrespectful" or not!

DXO doesnt' publish how it runs it's tests and under what ambient condition controls.

however, keep in mind that the sensor may be the same, or as good, but the operating environment be poorer.

if you are running the EVF, it's clocking at 120hz. that means the video feed from the sensor, line skip or not, is running at 120hz.

increase in ambient temperature > less sensor performance.

one of the reasons nikon was always able to goose more out of sony sensors is that sony uses EVF for all it's cameras, thus running at higher ambient.

 

[scyrene]

interesting to see that M5 sensor performance is behind 80D. as much as i dislike DXO generally, one would hope that the same test method was applied to both sensors and results are therefore comparable.

whether it is a full stop or not, i find it disappointing when a newer camera has less IQ/sensor performance than an older model. would be intetesting to get reasons/explanation for this "phenomenon". one of the questions, dpreview or other website staff could put in front of the Canon exec muppets when they "interview" them next time ... no matter if those guys would find such a straightforward, fact-based question horribly "disrespectful" or not!

Don't believe any measurement that is quoted without a standard error and standard deviation (or estimate of uncertainty) and you don't know how many times that measurement was made and whether it is on more than one copy of the camera. Lensrentals is believable because they do multiple measurments and quote the mean values, and sometimes have shown the deviations from the mean. Unless DxOmark gives an indication of the standard deviation of the measurements, I would not take the difference between two relatively similar numbers as being significant. JUst take DxOmark as giving a rough estimate of the properties of the sensor.

Good point!

 

[lescrane]

for me, too little, too late

Sorry, Canon.
This looks like a great camera,with most of the features I need in a small, light package. I'm not sure how many lenses there are for this, when I sold my original "M" there were only 3 system lenses.(I don't count adapted lenses)

After 30 years of using Canon SLRS/DSLRS, I sold most of my lenses, and kept the two bodies(60,70d)I had since they fetched little on the market. I am now fully invested in the Panasonic Lumix m43 system.

I had no complaints about Canon ergonomics or IQ, none at all. However, weight and size became more of a factor every year. If Canon had developed an M system faster, including long zoom lenses, I probably would have stayed. Frankly, Panasonic Mirrorless are harder to use if you like making a lot of adjustments. However, I have reduced the weight factor 50% from the DSLRs, which is a load off when I go on a hike or nature walk.

As far as IQ, well, I really see no difference in noise up to 1200 ISO, or resolution. I'm not saying there IS no difference, just that I'm not seeing it, even on A3 prints. Obviously still check Canon Rumors once in a while, and if a friend is intent on a DSLR, I recommend Canon. However, the axiom that the "Best camera is the camera that you have with you" means I have no regrets about lightening up with the Lumix system.

 

[Orangutan]

Re: for me, too little, too late

Sorry, Canon.... I am now fully invested in the Panasonic Lumix m43 system.

Enjoy!

 

[KeithBreazeal]

Some "mad science" taking place here. I need to go find some birds that will sit still and pose for me.

Canon EOS M5 Canon 300L f2.8 Canon 2X Converter © Keith Breazeal by Keith Breazeal, on Flickr

 

[bholliman]

Some "mad science" taking place here. I need to go find some birds that will sit still and pose for me.

I actually used this combo a few times last winter for perched birds and it performed well. The small camera and grip make it a bit unwieldy to handle.

I tried the 300 + 1.4x once as an experiment for shooting birds in flight with some geese, just to see how it would do... and let's just say I've used my 5DsR for all my BIF ever sense. The M5 performs nicely for perched birds however. Attached is a shot with the M5 with 300 f/2.8 II + 1.4xiii. I have some with my 2x as well, but don't have access to those currently.

 

[AlanF]

That's a very sharp photo, as it should be - the M5 has an excellent sensor and focus.

 

[dcm]

Truck rolled/slid down the hillside across the lake earlier this year - just over 1000 meters away. Thought I'd get a closer look with the M5, 100-400L II, and 2xIII on a tripod so see how the combination performed.

Not bad near noon in 90 degree weather. You can make out the Sheriffs Office markings on the cab if you open the image and enlarge it. Only adjustment was white balance. Need to try it out on some closer subjects next time.