EOS 90D dynamic range

[MAC]

However, I think it would be safe to assume that a 1.6x crop from the 5D4 has little or no benefit over the 90D at high ISOs.

I hope so

If true, powerful...

Well, it's already true for the 7D2. The 5D4 should, according to noise measurements, have 1/3 stop less read noise per unit of sensor at high ISOs, but with the same exposure, noise looks pretty much the same, because the 7D2 noise is a little bit more random, spatially, as well as being finer. The 7D2 has slightly less noise at ISO 12800 in the DPR studio comparison tool than the 5D4 at ISO 3200 (the same total light or exposure times area) in incandescent mode. In the daylight mode, they look roughly the same. This hides the fact that actually cropping the 5D4 at the same ISO would have even more visible noise, because the spatially-correlated part of 5D4 noise is magnified more, into more visible lower final, cropped image frequencies than when you use the entire 5D4 frame.

looks like you flipped that 7d2 and 5d4 above

for me my 6d at iso12800 looks the same as the 5d4 at iso12800

my 7d2 matches my 6d 12800 at about 6400

and my T7i loses at stop and matches at iso 3200

it looks to me that they are just catching the 90d up to the 7d2 level of noise whereas the 80d was behind the 7d2

 

[John Sheehy]

However, I think it would be safe to assume that a 1.6x crop from the 5D4 has little or no benefit over the 90D at high ISOs.

I hope so

If true, powerful...

Well, it's already true for the 7D2. The 5D4 should, according to noise measurements, have 1/3 stop less read noise per unit of sensor at high ISOs, but with the same exposure, noise looks pretty much the same, because the 7D2 noise is a little bit more random, spatially, as well as being finer. The 7D2 has slightly less noise at ISO 12800 in the DPR studio comparison tool than the 5D4 at ISO 3200

Should be "32000".

(the same total light or exposure times area) in incandescent mode. In the daylight mode, they look roughly the same. This hides the fact that actually cropping the 5D4 at the same ISO would have even more visible noise, because the spatially-correlated part of 5D4 noise is magnified more, into more visible lower final, cropped image frequencies than when you use the entire 5D4 frame.

looks like you flipped that 7d2 and 5d4 above

Sorry, I left a "0" out of what should have been "32000" (thirty-two thousand).

The 5D4 at 32000 should get the same total light as the 7D2 at 12800.

The 5D4 is tested at 32000, which is unusual.

 

[John Sheehy]

This is great news, many thanks for doing this.

I know the normalised values are the correct ones to compare, but as a wildlife photographer I'm often focal length limited, and a major attraction of the new sensor is its significantly higher pixel count - more 'pixels per bird'. However it would be worth much less if it had been done at the expense of dynamic range, so it's great to see these results.

Normalized is the right thing for focal-length-limited concerns, too, unless you would automatically crop a certain fixed number of pixels, rather than a fraction of the frame, which I personally would find a bit strange, unless you are compensating with actual optical focal length. The read noise levels, however, are still way higher than the competition's best. That's not necessarily a big problem in open areas during daylight, even if a bit overcast, but in the deep shade of heavy foliage, where the red channel light capture is guano, Canon is way behind the rest.

Sure, but I'm sticking with Canon so my main concern was that the very attractive big step up in pixel density from the bodies I have wasn't going to be rendered a lot less attractive by IQ issues. I'm greatly reassured by what I've seen so far.

The 20MP m43 sensor in current production has almost 2 stops less input-referred pre-gain (high ISO) read noise than the 90D, with roughly the same size pixels. That means almost 4x the red-channel read noise in the shade for Canon.

Yep, it's very impressive. If only we could pick the bits we like from each brand

Imagine only one lens mount; it's easy if you try ...

 

[Stollen1234]

would be interesting to do some fotos comparsion between 7DII and 90D es pecially like wild life fotos and sports...

 

[Chris Mak]

This is great news, many thanks for doing this.

I know the normalised values are the correct ones to compare, but as a wildlife photographer I'm often focal length limited, and a major attraction of the new sensor is its significantly higher pixel count - more 'pixels per bird'. However it would be worth much less if it had been done at the expense of dynamic range, so it's great to see these results.

Normalized is the right thing for focal-length-limited concerns, too, unless you would automatically crop a certain fixed number of pixels, rather than a fraction of the frame, which I personally would find a bit strange, unless you are compensating with actual optical focal length. The read noise levels, however, are still way higher than the competition's best. That's not necessarily a big problem in open areas during daylight, even if a bit overcast, but in the deep shade of heavy foliage, where the red channel light capture is guano, Canon is way behind the rest.

Sure, but I'm sticking with Canon so my main concern was that the very attractive big step up in pixel density from the bodies I have wasn't going to be rendered a lot less attractive by IQ issues. I'm greatly reassured by what I've seen so far.

The 20MP m43 sensor in current production has almost 2 stops less input-referred pre-gain (high ISO) read noise than the 90D, with roughly the same size pixels. That means almost 4x the red-channel read noise in the shade for Canon.

Yep, it's very impressive. If only we could pick the bits we like from each brand

Imagine only one lens mount; it's easy if you try ...

It may not be necessary to start putting the Canon big whites on an Olympus body. The Sony A7RIII improved dynamic range with the same sensor as the Sony A7RII. I am sure that Canon can lower the read noise of the new 32,5mp sensor with the right (more expensive) hardware surrounding it. The EM1II is a body aimed at professionals or (very) serious amateurs. The 90D is aimed at pretty much the masses of enthousiast amateurs. There is simply a gap above it when it comes to APS-C bodies. Part of the lower performance will come down to Canon's sensors, but certainly not all of it.

Chris

 

[John Sheehy]

would be interesting to do some fotos comparsion between 7DII and 90D es pecially like wild life fotos and sports...

I would expect that with both focused correctly, the 90D will look a bit better and allow more NR.

32.5MP APS-C, of course, reduces the severity of aliasing vs 20.2MP, especially with slower lenses and softer ones which chip in to reduce color moire.

Oh yeah, I almost forgot. If the 32.5MP sensor is BSI, as some have said, then it should allow more light through at very low f-numbers (lower than f/2.8), giving more of an edge .

 

[Waldemar]

I still don´t understand, in what way DR matters. O.k., I maybe understand it for JPEG shots.

But for raw? As I understand, according to this:

https://www.dpreview.com/articles/7450523388/sony-alpha-7r-ii-real-world-iso-invariance-study

you need an ISO-invariant sensor to gain several stops of highlight headroom by deliberately underexposing the shot.

If the sensor is not ISO invariant, it still can be done. Just to a much smaller amount, before noise and other artifacts become unacceptable. So basically you have to shoot at base ISO as lifting shadows will always introduce noise. Also, there might not be many details in the dark area of the raw file to begin with.

But all this doesn´t have much to do with DR.

The DR as such hasn´t changed that much since the 5D II.

I´d rather speak of "raw DR" now, as underexposing and then brightening the shot in post is not the same as the usual DR measurement / number.

Increasingly, it´s hard to say, what DR is at all or in which way it matters. With the 5D I and 5D II it was hardly possible to pull back shadows, so the DR number was somewhat more clearer. But now? If you use an ISO invariant sensor and the camera does the underexposing / brightening to a JPEG as final result, you could catch a scene with an extreme DR range of, say, 18 EV, without having to shoot raw and postprocess. It´s just not being done to that extreme right now (maybe because the result is hardly predictable and a JPEG can´t be changed much afterwards). But we´ll get there, I guess. And that´s why ISO invariance is probably the most important innovation since the begin of digital imaging.

---------------------------------------------
Wald

 

[Sharlin]

Dynamic range is simply the logarithm of the signal-to-noise ratio. Nothing more, nothing less. In engineering we’d use base-ten log and measured DR in decibels, but as photographers we’re more accustomed to stops and thus use base-two.

The maximum signal level is limited by the full well capacity of the sensor, and by the bit depth of the analog/digital conversion. The minimum signal level is bounded by photon counting statistics.

 

[John Sheehy]

Increasingly, it´s hard to say, what DR is at all or in which way it matters. With the 5D I and 5D II it was hardly possible to pull back shadows, so the DR number was somewhat more clearer. But now? If you use an ISO invariant sensor and the camera does the underexposing / brightening to a JPEG as final result, you could catch a scene with an extreme DR range of, say, 18 EV, without having to shoot raw and postprocess. It´s just not being done to that extreme right now (maybe because the result is hardly predictable and a JPEG can´t be changed much afterwards). But we´ll get there, I guess. And that´s why ISO invariance is probably the most important innovation since the begin of digital imaging.

"ISO invariance" is not any kind of absolute noise metric; it simply describes a state where the ISO setting can vary a bit for a manual exposure without affecting noise, on that camera. It has nothing to do, whatsoever, with how noisy that camera is compared to others.

Cameras may seem more ISO-invariant in charts than they do visually, because charts can't see when the post-gain read noise has heavy spatial correlation. IME, banding noise can be an order of magnitude more visible than spatially-random noise with the same standard deviation. So, while a camera may give 1.6 DN of read noise at ISO 100, and 3.0 DN at ISO 200, what makes the ISO 100 1.6 instead of 1.5 may be highly patterned post-gain read noise, which is reduced to 0.1 difference in quadrature, but is actually much large, if measured by itself, in addition to also being a more visible noise.

Years ago the poster here, Chato, read that his Nikon D2X was ISO-invariant. So, he put it to the test, and his ISO 100 had chunkier and greater visible chromatic noise than his ISO 400 with the same exposure.

 

[John Sheehy]

Dynamic range is simply the logarithm of the signal-to-noise ratio. Nothing more, nothing less. In engineering we’d use base-ten log and measured DR in decibels, but as photographers we’re more accustomed to stops and thus use base-two.

The maximum signal level is limited by the full well capacity of the sensor, and by the bit depth of the analog/digital conversion.

That's the maximum ADU or DN; not the maximum signal. Gain plays a role there matching signal to DN.

The minimum signal level is bounded by photon counting statistics.

Where can I buy that camera?

Every existing camera is infested and brought to its knees by read noise (especially spatially-correlated read noise) in very deep shadows and/or high ISOs. It is only in the relatively non-problematic higher tones of lower ISOs where photon noise can dominate.

As far as photon noise is concerned, we are also in an age of extremely inefficient color filters. The vast majority of light, especially red light, never makes it through the color filters and onto a sensor.

John

http://www.pbase.com/image/55384958.jpg

 

[Waldemar]

"ISO invariance" is not any kind of absolute noise metric; it simply describes a state where the ISO setting can vary a bit for a manual exposure without affecting noise, on that camera. It has nothing to do, whatsoever, with how noisy that camera is compared to others.

I understand this.

I´ve experimented a lot with underexposure on my A7 I and D750, actually before stumbling upon the mentioned DPR article on ISO invariance and it´s practical consequences.

It´s probably time to leave the usual DR definition behind and find something better, as it seems to have no practical meaning anymore. As long as you can brighten the dark parts of a shot without an excessive noise and artifacts penalty, you can have any DR you want (basically by mapping the brightness values). It should be possible to find a definition and a calculated number for that, instead of using examples of visual noise.

--------------------------------------------

Wald

 

[Sharlin]

The minimum signal level is bounded by photon counting statistics.

Where can I buy that camera?

Every existing camera is infested and brought to its knees by read noise (especially spatially-correlated read noise) in very deep shadows and/or high ISOs. It is only in the relatively non-problematic higher tones of lower ISOs where photon noise can dominate.

Yeah, sorry. Should have written a bit longer post going into read noise and stuff. What I meant was a theoretical lower bound.