The Sony a7R III has a well-earned reputation for low-light shooting. It has a superb full-sized sensor and a wide ISO range, going from 50 through 102400. But just how useful are those very high settings?
When you compare an image shot at ISO 102400 with one at ISO 50, you can see obvious differences. There’s more noise or mottling, less dynamic range, and the colors are less rich and true.
But those things aren’t necessarily dealbreakers. If the only way you’re going to get the shot is with a high ISO, just how high can you viably go with the a7R III?
To find out, I’ve shot some photos at the highest ISO settings on the a7R III. I recognize these aren’t necessarily the most compelling visual images, but what I was trying to do was to test and then share some real-world examples in case others were similarly curious.
These were shot in RAW and only minimally processed in Lightroom (with noise reduction zeroed out). You can click on each image to open a full-size version if you’d like a closer look.
Fixing image quality problems in post
The general rule of thumb in photography is that it’s better to get the shot right at the time of capture rather than trying to fix it later.
That’s a great aspiration, but it’s not always possible to do if you’re shooting in especially challenging conditions or bumping up against limitations or flaws in gear, conditions, or technique.
But it’s worth mentioning that there are some excellent tools available to help address common image quality issues with digital images. And they’re getting better and better all the time as the power of AI machine learning ramps up.
They can deal remarkably well with anything from sensor issues like high-ISO image noise to lens issues like distortion, softness, vignetting, and chromatic aberration. (Note: I’m focusing here on corrections related to image quality, not image enhancement tools — that’s a different kettle of fish.)
All-round image processing apps like Lightroom Classic and Capture One have solid tools built in that are very good places to start.
But it’s also possible to take it much further with more specialized tools.
If you shoot in challenging conditions regularly and find room for improvement in the image quality coming out of the camera, these might well be worth a look (and they have free trials). Some are stand-alone apps; some integrate into image editing suites such as Lightroom Classic.
Some of these work only with RAW files, and those are the ones I’ve had the best results with. But there are also good options if you’re working with JPG or TIFF files.
Fixing image noise & high ISO artifacts
The newest generation of tools can do some pretty remarkable improvements to image noise. These are some of the best tools for reducing image noise in photos. Most of them include a free trial to test out the results for yourself.
- Lightroom Classic’s Denoise tool. This was quietly added to recent versions of Lightroom Classic, and you can find it in the Develop module’s Detail panel. It only works on RAW files. It doesn’t have some of the more advanced features and more granular controls of some of the other apps here, but if you’re already using Lightroom, it’s a great option to have available.
- DxO PureRAW 5. Like Lightroom Classic’s Denoise tool, it only works on RAW files. It has become my go-to app for this kind of thing. I’m consistently amazed at how it can rescue photos with otherwise dodgy image quality from noise. It can also help with lens distortion, lens vignetting, and lens softness.
- DxO Nik Dfine. This product originated with the Nik Collection, and it focuses on the more traditional approach to digital image noise reduction. Something that’s especially notable about Dfine is that you can use U-point selections for precise application of different settings to different parts of the image.
- ON1 NoNoise AI. Its “deblurring” features are particularly notable, with an emphasis on restoring sharpness that is often a casualty in denoising operations. It works on JPG and TIFF files, and you can create sophisticated masks to apply corrections to specific parts of an image.
- Topaz Labs Photo AI. Topaz Labs used to offer its NoNoise AI as a standalone product, but it has now been integrated alongside several other image improvement tools in its Photo AI app.
Fixing lens-related optical issues
- DxO PureRAW. Again offers an impressive suite of automatic fixes that are applied before you start editing the images.
- DxO ViewPoint. Correcting for lens distortion and geometry skews. Lightroom Classic and Capture One also have excellent built-in tools for this.
- Topaz Labs Photo AI. Topaz Labs used to offer its sharpening tools as a standalone product, but it has now been integrated alongside several other image improvement tools in its Photo AI app.
Conclusion
The high-ISO performance of the a7R III is pretty remarkable. Yes, there are tradeoffs. You lose some dynamic range, color fidelity, and there is definitely more noise and loss of details, all of which is especially noticeable when you view at 100%. Even with RAW, you can’t recover much from the shadows or highlights. But the costs are remarkably light compared to state-of-the-art sensors of even a few years ago. It wasn’t long ago that you were happy if you got usable images out of ISO 3200. That we can now come to expect usable images at ISOs in the six digits is pretty extraordinary. Even if you wouldn’t necessarily want to be using these high-number ISOs in everyday shooting, it’s sure nice to know that you have the headroom if you need it and that the resulting images will be usable for many situations.
Price & availability
Check the current price and availability of the Sony a7R III at:
The images above were shot with the Sony a7R III. Sony has since released the a7R IIIA. It improves the resolution of the back screen and adds USB-C 3.2 compatibility but is otherwise identical.


















Cloudflare is returning a file access denied error message when you try to open the full size orginal image, stating that you do not allow resource hot linking….
Thanks for the heads up. I think this is now fixed (hard to confirm because I’m having trouble replicating the issue this end).