The Nikon Z9 uses a Stacked CMOS sensor but with a pixel count and base ISO shared with the single layer BSI chip used by the Z7 and Z7 II. That was already one of the best-performing sensors we've ever encountered, so our look at the Z9 is more likely to be a question of 'what's the cost of the stacked chip's added speed,' rather than 'how much better is the IQ.
'
Studio scene
Our test scene is designed to simulate a variety of textures, colors and detail types you'll encounter in the real world. It also has two illumination modes to see the effect of different lighting conditions.
$(document). ready(function() { ImageComparisonWidget({"containerId":"reviewImageComparisonWidget-55953283","widgetId":845,"initialStateId":5578}) })
The details levels on show from the Z9 are a fraction above those of the Z7 II, perhaps because the magnified live view has allowed slightly better focus, but also possibly as a result of the absence of any shutter shock (or mechanical shutter, for that matter). There's enough moir$(document). ready(function() { $("#icl-5579--1483316096"). click(function() { ImageComparisonWidgetLink(5579); }); }) present to suggest there's no AA filter, but this could also suggest the 85mm F1. 8 is readily resolving higher frequencies than the sensor can correctly capture.
There appears to be a (small) price to pay$(document). ready(function() { $("#icl-5580--2084881143"). click(function() { ImageComparisonWidgetLink(5580); }); }) in terms of noise performance at higher ISOs, compared with both the Z7 II and its immediate rivals, but it's only really visible in side-by-side comparisons, even at very high ISO settings$(document). ready(function() { $("#icl-5581--2077258244"). click(function() { ImageComparisonWidgetLink(5581); }); }).
JPEG color$(document). ready(function() { $("#icl-5582-620068565"). click(function() { ImageComparisonWidgetLink(5582); }); }) looks good, with warmish caucasian skin-tones, pinks that skew closer to orange than magenta, and solid (not-too-green) yellows. Sharpening$(document). ready(function() { $("#icl-5583--2122279093"). click(function() { ImageComparisonWidgetLink(5583); }); }) is less subtle than on the Sony, meaning the a1 ends up looking a touch more detailed. And, likewise, noise reduction$(document). ready(function() { $("#icl-5584-1547554799"). click(function() { ImageComparisonWidgetLink(5584); }); }) at high ISOs can't retain as much detail as the Sony. Overall, the Z9 doesn't fall too far behind the Z7 II and isn't too far off the performance of the Sony a1.
Dynamic Range
The slightly higher noise level we saw results in a slight reduction in dynamic range, compared with the Z7 II (and hence, with the a1). If you underexpose, to protect highlights, then pull the dark regions of the image up, you'll encounter a little more noise, if you really push to the camera's dynamic range limits. It's not a major difference in noise, especially if you can make use of the camera's ISO 64 setting, rather than the a1's base ISO of 100.
Raising the ISO gives a decent improvement in noise performance over what we see at base ISO. Particularly once the dual gain sensor moves to its higher conversion gain mode, above ISO 400. From ISO 500 upwards, there's little benefit to raising ISO, which gives the option of using a low-light exposure but pinning your ISO at 500, to preserve highlights that might otherwise be amplified to clipping.
Exposure Latitude | ISO Invariance
Raw compression
We shot the most extreme push in the camera's three Raw compression modes to see if there was any impact on image quality. It appears there is some data loss in the deepest shadows if you engage the HE or HE* compression modes, but it only appears as an extremely slight increase in the noise floor, which is unlikely to have any impact on most photos.
Lossless Compression
Download Raw (45. 7MB)
HE* Compression
Download Raw (30. 0MB)
HE Compression
Download Raw (18. 6MB)
These images were processed with noise reduction minimized. Having separately examined the highlights of a real-world scene, we couldn't find any visible differences there, meaning it's unlikely you'd ever notice a difference except by side-by-side comparison and measurement.
. dpreview.com2022-3-21 17:00
