I just have converted a a dpreview sample picture to test the DR. The results are slightly better compared to the 80D:
90D ISO 100:
DR at 32MP: 12.468 DR at 8MP: 13.480 (+1.012)
read noise: 2.80125
80D ISO 100:
DR at 24MP: 12.435 EV
DR at 8MP: 13.2343 (+0.7993)
read noise: 2.86601
The improvement is tiny, close to the measure error. But Canon has managed to keep the same performance despite the smaller pixels (not too difficult, that was the trick the 5DS used to stand out in the old times). After taking into account the normalization, the new sensor is around 0.25 EV better.
EDIT: see this post bellow for updated data at the whole ISO range
I'd be curious also about performance in the 1600-6400 ISO range. But I suppose I can remain patient, given that the camera isn't even on the shelves yet.
Tom_W wrote:
I'd be curious also about performance in the 1600-6400 ISO range. But I suppose I can remain patient, given that the camera isn't even on the shelves yet.
+1
Not a lot of complaint's regarding DR @ base ISO of the 80D. .25 EV improvement @ ISO 100 is inconsequential (welcome, but still). How quickly and how much it falls apart as you move north is likely the bigger question on folks mind. Whether you are shooting in the low light, or need the higher SS for moving subjects (often times both simultaneously) or longer focal lengths, the north end is where the most improvement is desired, imo.
Good to see the ISO 100 data held its own (marginal improvement) along with the 33% gain in px's ... now we wait to see how the North 40 plays with the extra px's.
The A9 and 80D only vary by less than a stop below ISO 800, that is usually where the bulk of the DR will be had across any camera. However the A9 kicks in another amplifier of sorts to take ISO 800 - 1/3 stop back up again by about 1/2 stop and then it goes down from there, like others, but at more DR.
I expect the 90D to barely improve on that, but we shall see.
TeamSpeed wrote:
The A9 and 80D only vary by less than a stop below ISO 800, that is usually where the bulk of the DR will be had across any camera. However the A9 kicks in another amplifier of sorts to take ISO 800 - 1/3 stop back up again by about 1/2 stop and then it goes down from there, like others, but at more DR.
I expect the 90D to barely improve on that, but we shall see.
Understood ... not expecting miracles, but a gain (vs. a loss) in the North end (accompanying the more px's) would be welcome.
Tom_W wrote:
I'd be curious also about performance in the 1600-6400 ISO range. But I suppose I can remain patient, given that the camera isn't even on the shelves yet.
Sy Sez wrote:
There are some downloadable 90D raw files on the Photography Blog site, in ISO's up to 6400.
Thanks for these! The RAW ISO 100s look great. Only 1 ISO 1600 between ISO 100 and ISO 3200. What about some ISO 800s? Overall I'd say looks about the same for noise as my 80D. Having said that, ISO 6400 is way too noisy for my use, ISO 3200 is marginal and ISO 1600 is about my max tolerance. I have max auto ISO set to 1600 on my 80D and 3200 on my 5D4 for reference.
I'll add that the lighting for the bumper cars is way bad, I'd probably not shoot in those conditions.
Canon's DPP software works remarkably well for noise reduction so I expect very satisfactory images at 6400, maybe even 12800. In a pinch even images at 25600 should be salvageable.
The trick to lure me in would be to make it so good that I don't reach for the 5D4 for bird photography. I'd enjoy the pixel density immensely with this or the M6 II, but high ISO favors the full framer. Granted, I could get by with faster glass along with the high pixel density, but that would mean more $$.
Time will tell - this could be really interesting.
Tom_W wrote:
The trick to lure me in would be to make it so good that I don't reach for the 5D4 for bird photography. I'd enjoy the pixel density immensely with this or the M6 II, but high ISO favors the full framer. Granted, I could get by with faster glass along with the high pixel density, but that would mean more $$.
Time will tell - this could be really interesting.
It's a great time to be a photography gearhead!
Plus, this sensor will very likely be showing up in every new APS-C camera Canon sells. This means the Rebels, SL4 (I really hope there will be an SL4), all new M cameras and maybe even an RF mount APS-C camera. It will end up in very affordable cameras.
cgarcia wrote:
I just have converted a a dpreview sample picture to test the DR. The results are slightly better compared to the 80D:
90D ISO 100:
DR at 32MP: 12.468 DR at 8MP: 13.480 (+1.012)
read noise: 2.80125
80D ISO 100:
DR at 24MP: 12.435 EV
DR at 8MP: 13.2343 (+0.7993)
read noise: 2.86601
The improvement is tiny, close to the measure error. But Canon has managed to keep the same performance despite the smaller pixels (not too difficult, that was the trick the 5DS used to stand out in the old times). After taking into account the normalization, the new sensor is around 0.25 EV better.
Thanks for extracting the data and doing the calculations!
If the same pixel design and readout / conversion accuracy is used in a full frame camera, it will translate to 14.16 Ev dynamic range normalized to 8 MP. That is about 1.7 Ev better than the 5DS.
If feasible I would like to see the results of the same analysis with the Jabber images. Although shot in the same place, they look much better than the similar images posted by DPR.
PicGuy wrote:
Plus, this sensor will very likely be showing up in every new APS-C camera Canon sells. This means the Rebels, SL4 (I really hope there will be an SL4), all new M cameras and maybe even an RF mount APS-C camera. It will end up in very affordable cameras.
And, I have a couple of M lenses and the EF adapter, so I would be leaning towards something in the M series as it is.
alundeb wrote:
Thanks for extracting the data and doing the calculations!
If the same pixel design and readout / conversion accuracy is used in a full frame camera, it will translate to 14.16 Ev dynamic range normalized to 8 MP. That is about 1.7 Ev better than the 5DS.
vs 14.8 d850 best per dxomark and 13.6 5div dxomark
dclark wrote:
The link to the sample image seems to be broken.
It seems that FM doesn't always properly redirect when a link is embedded...
Well, there are GREAT news... I tested the ISO 6400 by curiosity and, after detecting a big improvement, I have done a full comparison with the 80D, except for ISO 3200 which is missing in dpreview's samples. Since the sensor is far from being ISO-less, and one needs to crank up the ISO to improve the shadows, the new DR at high ISO is very valuable.
The 90 has much better performance at high ISO. To the point that from ISO 1600 it seems nearly an stop ahead of the 80D. That is, the ISO 6400 should look as clean as ISO 3200 on the 80D (ok, in the shadows... in the midtones, physics laws can't make miracles). These are the read noise values:
90D vs 80D:
ISO 100: 2.77497 vs 2.84197
ISO 200: 4.43868 vs 3.99936
ISO 400: 5.55092 vs 6.35349
ISO 800: 8.65067 vs 11.304
ISO 1600: 14.2535 vs 20.1949
ISO 3200: ? vs 36.077
ISO 6400: 42.5752 vs 61.7794
The ISO 100 values differ a bit from yesterday because I have used both green channels instead of a single one. Well, we are not sure if the ISO sensitivity is equal in both cameras, but this is looking very good.
I have retested ISO 200 and 400 since there were two available images on each and the results (read noise 4.40524 and 5.4736) confirm that the DR "dip" at ISO 200 is real. I don't say it has two base ISO settings (another one optimized for high ISO) but this is interesting. I think that other Canon cameras already had this behaviour. The 90D also applies scaling to the red and blue channels (likely white balance correction applied to the raw) as the 5D4 did.
These are the 90D images used (ISO 100 - ISO 6400, except for ISO 3200):
