Canon M6 Mk II First impressions

Have questions about the equipment used for macro- or micro- photography? Post those questions in this forum.

Moderators: Pau, rjlittlefield, ChrisR, Chris S.

Macrero
Posts: 888
Joined: Thu Sep 01, 2011 8:17 am
Location: Valladolid , Spain

Canon M6 Mk II First impressions

Post by Macrero »

This morning I received the M6 II. Hands-on, it is a decently made, cute, little camera.

First off, the e-shooter is completely electronic/silent. There is no EFSC option, just mechanical or e-shutter.

The touchscreen is nice, manual focus is easy, no flickering with LEDs.

Now... the most important part: IQ.

DR is not great, but that was to be expected from Canon. Not that bad either for an APS-C camera.

My RAW converter does not support yet the M6 II CR3 files, so I have to convert them to DNG, not big of a problem.

Resolution wise, the OLPF is indulged in its ways... Why, Canon? Why on earth did you put a blurring filter on front of a 32.5MP APS-C sensor? :? :evil:

I am thinking of either return the camera or get the darn AA filter removed. Will see if I can get that done locally, sending it to US is much of a hassle and money...

Will post some test stacks latter.

- Macrero
https://500px.com/macrero - Amateurs worry about equipment, Pros worry about money, Masters worry about Light

Macrero
Posts: 888
Joined: Thu Sep 01, 2011 8:17 am
Location: Valladolid , Spain

Post by Macrero »

Asked Kolari about buying the replacement, "clean" filter alone and getting the mod done locally, that's what they said:

"We can arrange something, but this camera is difficult to modify. You will also need to lose the dust reduction filter in the process, but the technician must leave it partially in the camera so that you do not get a camera error. I would not recommend doing this one locally, it is very hard for us to support the technical challenges with this camera remotely. "


Tomorrow I'll do a few more tests, but I think the camera will get back to the seller asap. And I promise myself that this is the very last Canon camera I buy! Silly me... I'm done with you, Canon, this time for good. What's next? 80MP FF camera with an even stronger OLPF?
https://500px.com/macrero - Amateurs worry about equipment, Pros worry about money, Masters worry about Light

ray_parkhurst
Posts: 2819
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

The OLPF, and lack of EOS Utility support, are what has kept me away from the M series cameras.

I had my T2i for a year or two before I sent it in to have the OLPF removed. My impression after comparing with and without is that there was a noticeable, though more subtle than expected, improvement in sharpness, but that false color became a significant problem. The demosaicing algo for the T2i assumes the OLPF is in place, and my hypothesis is that the poor false color performance was due to lack of algo support for this phenomenon. False colors don't improve on the HRT2i until around f9, and are not completely gone until f11.

Based on my experience with the HRT2i, I would not go down the path of OLPF removal again.

Macrero
Posts: 888
Joined: Thu Sep 01, 2011 8:17 am
Location: Valladolid , Spain

Post by Macrero »

Well, maybe I am being too hard on Canon. It is actually my fault, since I did know that there is an AA filter in this camera.

However, I would love to get an explanation from Canon sensor engineers about why they put a blurring filter on this sensor, given the pixel count/pitch/density :?

I'm drinking my morning coffee. In a while I will go to the studio and will make more tests. Maybe the resolution has improved overnight ... :roll: :lol:
https://500px.com/macrero - Amateurs worry about equipment, Pros worry about money, Masters worry about Light

chris_ma
Posts: 283
Joined: Fri Mar 22, 2019 2:23 pm
Location: Germany

Post by chris_ma »

well, I would argue against the reasoning to put a sensor with this pixel count in a camera like this in the first place ;)

I'd guess 95% of the people buying this camera will never print anything bigger then 30x40cm (or not print anything at all), plus there's the question if a picture with tons of high frequency detail really looks more pleasing (personally I find it does only on a few content)

I like high res digital cameras for technical work, but for portraits and moody images I'm still using analog film cameras because I prefer the softness and colors.

really like your signature as well because I catch myself worrying far too much about money and equipment.
chris

Macrero
Posts: 888
Joined: Thu Sep 01, 2011 8:17 am
Location: Valladolid , Spain

Post by Macrero »

Fair enough, but for "normal" photography the lens itself will act as a blurring filter on this sensor in most cases.

I will try a different RAW converting method, and see if I can offset the cataracts effect :evil: without ending up with a full of noise and artifacts image...
https://500px.com/macrero - Amateurs worry about equipment, Pros worry about money, Masters worry about Light

ray_parkhurst
Posts: 2819
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

Macrero wrote:Fair enough, but for "normal" photography the lens itself will act as a blurring filter on this sensor in most cases.
I think this is what is happening with my HRT2i. The filter had a similar effect to stopping the lens down to f9-f11.

Macrero
Posts: 888
Joined: Thu Sep 01, 2011 8:17 am
Location: Valladolid , Spain

Post by Macrero »

ray_parkhurst wrote:The OLPF, and lack of EOS Utility support, are what has kept me away from the M series cameras.

I had my T2i for a year or two before I sent it in to have the OLPF removed. My impression after comparing with and without is that there was a noticeable, though more subtle than expected, improvement in sharpness, but that false color became a significant problem. The demosaicing algo for the T2i assumes the OLPF is in place, and my hypothesis is that the poor false color performance was due to lack of algo support for this phenomenon. False colors don't improve on the HRT2i until around f9, and are not completely gone until f11.

Based on my experience with the HRT2i, I would not go down the path of OLPF removal again.
Ray, dou you have some "with and without" comparison pics? I would mod the M6 II, if there were an european company that offers an AA filter removal, I am not aware of any. I am reluctant to send it to US just for the mod.

Btw, LifePixel sell already no-AA converted M6 II for $1,124.00 + shipping. But that would be more than $1,500 with customs fees. A lot of money for this camera, with or without AA filter...
https://500px.com/macrero - Amateurs worry about equipment, Pros worry about money, Masters worry about Light

ray_parkhurst
Posts: 2819
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

Macrero wrote:
ray_parkhurst wrote:The OLPF, and lack of EOS Utility support, are what has kept me away from the M series cameras.

I had my T2i for a year or two before I sent it in to have the OLPF removed. My impression after comparing with and without is that there was a noticeable, though more subtle than expected, improvement in sharpness, but that false color became a significant problem. The demosaicing algo for the T2i assumes the OLPF is in place, and my hypothesis is that the poor false color performance was due to lack of algo support for this phenomenon. False colors don't improve on the HRT2i until around f9, and are not completely gone until f11.

Based on my experience with the HRT2i, I would not go down the path of OLPF removal again.
Ray, dou you have some "with and without" comparison pics? I would mod the M6 II, if there were an european company that offers an AA filter removal, I am not aware of any. I am reluctant to send it to US just for the mod.

Btw, LifePixel sell already no-AA converted M6 II for $1,124.00 + shipping. But that would be more than $1,500 with customs fees. A lot of money for this camera, with or without AA filter...
Sorry, but I don't have any good comparison pics to share.

Macrero
Posts: 888
Joined: Thu Sep 01, 2011 8:17 am
Location: Valladolid , Spain

Post by Macrero »

ray_parkhurst wrote:
Sorry, but I don't have any good comparison pics to share.
Ok, no worries.

After the disappointing first couple of test stacks, I changed my RAW converting / editing workflow according to the new sensor and RAW files and results are improving.

Btw, the guys at Kolari say that "on a high MP camera the AA removal only has marginal sharpness benefits"

So, I don't think it is worth it the time, money and potential side issues.

Best,

- Macrero
https://500px.com/macrero - Amateurs worry about equipment, Pros worry about money, Masters worry about Light

ray_parkhurst
Posts: 2819
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

Macrero wrote:
ray_parkhurst wrote:
Sorry, but I don't have any good comparison pics to share.
Ok, no worries.

After the disappointing first couple of test stacks, I changed my RAW converting / editing workflow according to the new sensor and RAW files and results are improving.

Btw, the guys at Kolari say that "on a high MP camera the AA removal only has marginal sharpness benefits"

So, I don't think it is worth it the time, money and potential side issues.

Best,

- Macrero
The number of MP in the camera should make little difference to the effectiveness of the AA filter removal. Maybe they are thinking that most of the time people shoot at small apertures, or with lenses that are not diffraction-limited until well above the DLA of the sensor, but for the work many of us do in pushing toward the maximum possible sharpness, the AA filter is indeed a stumbling block. Now, it is true that some cameras have AA filters with less blurring. Each camera model is different in this way, since the AA filter is designed along with the other filters and sensor to give a certain target performance. Thus some camera models have "strong" or "weak" AA filters per various reports. I don't know if it is true or not, but if there is a direct correlation between number of MP and the strength of the AA filter, then indeed they may be correct that removing the AA filter has less effect on higher MP cameras. However, I don't believe there is such a correlation, and it would be a difficult thing to figure out.

chris_ma
Posts: 283
Joined: Fri Mar 22, 2019 2:23 pm
Location: Germany

Post by chris_ma »

ray_parkhurst wrote:I don't know if it is true or not, but if there is a direct correlation between number of MP and the strength of the AA filter, then indeed they may be correct that removing the AA filter has less effect on higher MP cameras.
I'd rather say that the AA filter is related to the pixel size since it's purpose is to spread the light ray partially over several pixels to reduce moiree from debayering.
so in that regard it totally makes sense that a sensor with 5um pixels need a stronger AA filter then one with 3um etc.

like, in a typical bayer pattern a circle the confusion on a light point source hitting a red pixel straight on needs to have a radius of 12.5um to hit the 4 pixels next to it, and 18um to also get the next 4 in the diagonal.
with a 3um sensor, that CoC radius would be reduced to 7.5um and 11um respectively.
with a 1um sensor we'd only need a CoC radius of 2.5um and 3.5um to reduce moiree, which would be a very weak AA indeed.

chris

[edit: corrected an error Ray pointed out and could lead to confusion)
Last edited by chris_ma on Mon Oct 28, 2019 3:13 pm, edited 1 time in total.

ray_parkhurst
Posts: 2819
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

chris_ma wrote:
ray_parkhurst wrote:I don't know if it is true or not, but if there is a direct correlation between number of MP and the strength of the AA filter, then indeed they may be correct that removing the AA filter has less effect on higher MP cameras.
I'd rather say that the AA filter is related to the pixel size since it's purpose is to spread the light ray partially over several pixels to reduce moiree from debayering.
so in that regard it totally makes sense that a sensor with 5um pixels need a stronger AA filter then one with 3um etc.

like, in a typical bayer pattern a circle the confusion on a light point source hitting a red pixel straight on needs to have a radius of 12.5um to hit the 4 pixels next to it, and 18um to also get the next 4 in the diagonal.
with a 3um sensor, that CoC radius would be reduced to 7.5um and 11um respectively.
with a 1um sensor we'd only need a CoC radius of 2.5um and 7um to reduce moiree, which would be a very weak AA indeed.

chris
Isn't your example of a 7um filter with a 1um sensor considerably "stronger" than 11um filter with 3um sensor? I would not think that the strength of the AA filter is directly related with the pixel physical dimension. Wouldn't spreading light over a certain number of adjacent pixels be independent of the physical pixel size?

I can see in the limit of extremely small pixels that indeed an AA filter would not be needed at all, since eventually lens diffraction spreads the light over multiple pixels and smooths any moire. Perhaps this trend is what the Kolari guys were considering, but I don't think we're there yet in terms of pixel size vs available optics, certainly not on CMOS.

chris_ma
Posts: 283
Joined: Fri Mar 22, 2019 2:23 pm
Location: Germany

Post by chris_ma »

ray_parkhurst wrote:Isn't your example of a 7um filter with a 1um sensor considerably "stronger" than 11um filter with 3um sensor?
you're right, there's a major calculation error here - the 2.5um for next to the side is correct (2.5 * 1um) but it should be 3.5um for the diagonal (2.5 * 1.4um)
I would not think that the strength of the AA filter is directly related with the pixel physical dimension. Wouldn't spreading light over a certain number of adjacent pixels be independent of the physical pixel size?
I'm confused - seems clear to me that if the pixels are larger, the adjacent ones are farther away (and bigger too), so the CoC needs to be larger too.
what makes things worse, we'll have to spread it out so it also hits those 2 rows over since the adjacent ones are color blind to the same color (specially for red and blue)

Macrero
Posts: 888
Joined: Thu Sep 01, 2011 8:17 am
Location: Valladolid , Spain

Post by Macrero »

Just asked them about the reason behind that "theory". Will report back when/if they respond.

Truth be told, the sensor is doing pretty well when paired with righ-res optics.

Here is a quick test stack of a dusty fly I found in the studio. With Oly 10/0.30 + 105mm tube lens @ 5.8X on sensor. RAW pre-sharpened, no additional sharpening.

https://easyzoom.com/image/158677/album/0/4?mode=manage

- Macrero
https://500px.com/macrero - Amateurs worry about equipment, Pros worry about money, Masters worry about Light

Post Reply Previous topicNext topic