E-M1 Mark II Sensor Results

[uMad]

At first glance, it seems like something significant may have finally changed and maybe this isn't the same sensor as in other cameras such as the GX8 at all. The performance has improved by a significant margin.

How many times Olympus needs officially say that Mk2 doesn't have any previous sensors but totally new one designed by Olympus?

Simply not believable. NR is

This. Especially since a Oly technical rep said, that noise reduction on sensor data before writing RAW files was one of the biggest improvements in the E-M1 mk2.

 

[Tommi K1]

It is unbelievable to you because you don't even know that this ain't Olympus first own designed sensor....

Olympus is not unheard or small player in sensor research and development... It just doesn't have a own fabrication lines to mass-produce sensors.

Olympus has own global shutter sensors, sensors with dramatic DR improvements, ultra high speed sensors and all kind things.

 

[Tommi K1]

And the same linked tech representative said he doesn't know about the sensor tech in the same sentence.

 

[bclaff]

Would love to get files from someone to test this camera myself.
Email (or Private Message) is you're willing to help.

I have been doing subjective tests comparing with a D700, D500 and D7200. I am happy to send you raw files. Some are well organise, some are not.

I appreciate the offer. But the files I use are taken in a particular way.

 

[Hiphopapotamus]

If you want evidence, you need to do prints or resize the pictures after good post processing to web sizes.

I have evidences too that show regardless of the sensor, all from 4/3" to best FF sensor produce identical quality at identical exposure values at larger than usual prints from photos at bad light or good light.

Even all these improvements in Mk2 doesn't matter as the difference is already non-existing or negligent. So only pushing the differences in a manner that doesn't matter.

Of course if we process to print then it doesn't really matter my E-M5 will do 30x20 prints which are hard to tell what they came from when they're processed properly. The thing is, six of one half a dozen of the other. This is a gear forum, and while I say that, like most people I want more from the gear that I already have.

I might as well post some files while I'm here that demonstrate I'm not wasting your time, but the problem is I'm running into issues where dynamic range, colour and ISO would only become slightly less of an issue if I were shooting with a full frame camera.

 

[bclaff]

Available at DxOMark, my DxOMark Derived figures are at my site,eg .:



Would love to get files from someone to test this camera myself.
Email (or Private Message) is you're willing to help.

Just a reminder that my charts use the ISO setting, not "measured ISO", along the x-axis.

DxOMark measured a whopping 0.75 stop advantage to the E-M1 Mark II over the D500 because Olympus messes so much with their ISO.



As for Noise Reduction (NR), I should be able to detect that once I get the appropriate files (I have one possibility so far).

--
Bill ( Your trusted source for independent sensor data at http://www.photonstophotos.net )

 

[uMad]

As for Noise Reduction (NR), I should be able to detect that once I get the appropriate files (I have one possibility so far).

How else could such an improvement be explained otherwise? (especially with small photosites at an FSI process).

If I am remembering right, you are calculating QE values for different cameras. Thinking about the D500 (probably a peak QE of 50 or even 60% if I had to guess), how would the E-M1 II with smaller photosites get very similar SNR values? (the QE would have to be improved by the difference in area (APS-C is over 70% bigger). So we would be talking about a peak QE of over 90%)

 

[bclaff]

The sensor characteristics give the QE as 81% which I really don't see. Is it really that far ahead of the A6500 and RX100M5? I think it's very likely to be wrong. So does that mean anything else might be less reliable or is it a stand-alone figure (which I suspect it is)?

DxOMark QE numbers are notoriously inconsistent.

I will have better data when I have the appropriate raw files.

Regards,

 

[photofisher]

Available at DxOMark, my DxOMark Derived figures are at my site,eg .:



Would love to get files from someone to test this camera myself.
Email (or Private Message) is you're willing to help.

Just a reminder that my charts use the ISO setting, not "measured ISO", along the x-axis.

DxOMark measured a whopping 0.75 stop advantage to the E-M1 Mark II over the D500 because Olympus messes so much with their ISO.



As for Noise Reduction (NR), I should be able to detect that once I get the appropriate files (I have one possibility so far).

--
Bill ( Your trusted source for independent sensor data at http://www.photonstophotos.net )

Thanks for clearing this up. Makes more sense now.

--
Amateur Photographer

 

[bclaff]

Thanks for the report.

But I have a question - when comparing the E-M1mk2 to Sony A6500 on your site it looks like the Olympus is above the Sony starting from ISO 200, but on DxO the Sony is above the Olympus on all measurements. What could explain this?

Not sure what you mean but curves crossing at low ISO settings is not unusual because of different sensor designs.; the A6500 looks more ISO Invariant.
As for how the curves compare at my site just remember I use the ISO setting at not measured ISO. According to DxOMark measured ISO the E-M1 Mark II has a 0.74 stop ISO setting advantage.

 

[bclaff]

As for Noise Reduction (NR), I should be able to detect that once I get the appropriate files (I have one possibility so far).

How else could such an improvement be explained otherwise? (especially with small photosites at an FSI process).

I try to make measurements, not assumptions.

If I am remembering right, you are calculating QE values for different cameras.

Yes, and I'm trying to do a "sanity check" on my numbers before publishing them.
I can say that mine appear to be more "sensible" than many of the DxOMark derived ones.

Thinking about the D500 (probably a peak QE of 50 or even 60%

I measure about 59%

if I had to guess), how would the E-M1 II with smaller photosites get very similar SNR values? (the QE would have to be improved by the difference in area (APS-C is over 70% bigger). So we would be talking about a peak QE of over 90%)

With good microlens QE (as we measure it) has little to do with pixel pitch.

 

[Skeeterbytes]

Looks very promising, thanks for your efforts, Bill!

Cheers,

Rick

 

[Jon555]

The sensor characteristics give the QE as 81% which I really don't see. Is it really that far ahead of the A6500 and RX100M5? I think it's very likely to be wrong. So does that mean anything else might be less reliable or is it a stand-alone figure (which I suspect it is)?

DxOMark QE numbers are notoriously inconsistent.

I will have better data when I have the appropriate raw files.

Regards,

 

[Jan Chelminski]

Very cool, thanks I like graphs.

I'm seeing it. I had the mk1 and shot it quite a bit, and the current mk2 is a step up, it is noticeable, to me. I can't quantify it but I would describe it as getting the extra resolution, without paying a noise penalty...plus a little more. I am satisfied.

 

[assaft]

Thanks for the report.

But I have a question - when comparing the E-M1mk2 to Sony A6500 on your site it looks like the Olympus is above the Sony starting from ISO 200, but on DxO the Sony is above the Olympus on all measurements. What could explain this?

Not sure what you mean but curves crossing at low ISO settings is not unusual because of different sensor designs.; the A6500 looks more ISO Invariant.

I mean that this is how the photographic dynamic range looks on your site, where the E-M1mk2 tops the A6500 starting from ISO 200:





And this is how the DR plot looks on DxO site, where the A6500 tops the E-M1mk2:



Are you saying that the difference is only because your scale is the ISO setting and DxO's is the measured ISO?

As for how the curves compare at my site just remember I use the ISO setting at not measured ISO. According to DxOMark measured ISO the E-M1 Mark II has a 0.74 stop ISO setting advantage.

--
Bill ( Your trusted source for independent sensor data at http://www.photonstophotos.net )

 

[bclaff]

Thanks for the report.

But I have a question - when comparing the E-M1mk2 to Sony A6500 on your site it looks like the Olympus is above the Sony starting from ISO 200, but on DxO the Sony is above the Olympus on all measurements. What could explain this?

Not sure what you mean but curves crossing at low ISO settings is not unusual because of different sensor designs.; the A6500 looks more ISO Invariant.

I mean that this is how the photographic dynamic range looks on your site, where the E-M1mk2 tops the A6500 starting from ISO 200:



And this is how the DR plot looks on DxO site, where the A6500 tops the E-M1mk2:



Are you saying that the difference is only because your scale is the ISO setting and DxO's is the measured ISO?

Yes, bring my E-M1 curve left 0.71 stops and it will be under the A6500 curve just like at DxOMark.
Also note, they don't show the lowest extended ISO and I do, that's the "flat" section on the left.

--
Bill ( Your trusted source for independent sensor data at http://www.photonstophotos.net )

 

[Max Iso]

At first glance, it seems like something significant may have finally changed and maybe this isn't the same sensor as in other cameras such as the GX8 at all. The performance has improved by a significant margin.

Seeing a 5 point integer jump has me thinking where this is coming from, or maybe its an account of something that has yet to be made obvious about the low light ISO performance.

Speculatively, about baked in noise reduction, but I don't know.

First off this is what we have been hoping for, a significant jump in some kind of sensor IQ. If it's NR at the sensor level we should be able to see it in comparison photos of detail, but i haven't heard anybody complain yet about softness or whatnot. It will be interesting to see if this is a different sensor or if Oly has just done something different in processing. Exciting either way and now im already thinking over my tax return budget.... Just wish it wasn't so spendy.

 

[David]

Available at DxOMark, my DxOMark Derived figures are at my site,eg .:



Would love to get files from someone to test this camera myself.
Email (or Private Message) is you're willing to help.

--
Bill ( Your trusted source for independent sensor data at http://www.photonstophotos.net )

Bill,

This is very interesting and I am trying to make sense of all of this on your website. I'm trying to compare the E-M1 Mk II against my current E-P5; is there image quality numbers I can see and compare with?

Thank you..

 

[Skeeterbytes]

Not to wander completely off-topic but has there been any question whether the E-M1-II sensor is new? So far as I know it's a bespoke item developed by Oly for this camera and built by Sony. It's not shared with any other camera model. The new TruPic VIII deserves co-billing too, but is little-discussed.

Cheers,

Rick

 

[bclaff]

Available at DxOMark, my DxOMark Derived figures are at my site,eg .:



Would love to get files from someone to test this camera myself.
Email (or Private Message) is you're willing to help.

Bill,

This is very interesting and I am trying to make sense of all of this on your website. I'm trying to compare the E-M1 Mk II against my current E-P5; is there image quality numbers I can see and compare with?

Thank you..

Since I haven't tested those camera you must relly on the DxOMark site or my DxOMark derived numbers.
For PDR we see :



According to DxOMark Measured ISO :



There's a 0.23 ISO advantage to the E-M1 Mark II so some people would say you need to move PDR curves 0.23 stops closer horizontally to compare.

Personally, I think of ISO setting from the photographers standpoint and don't make that adjustment.

--
Bill ( Your trusted source for independent sensor data at http://www.photonstophotos.net )