I believe that Tony's test was done using a photo of a photo (originally shot with a Canon camera over 10 years ago... and we know how 'good' Canon's dynamic range was! - the worse in the business I dare say back then).
Thanks for this video, it clears up the air a bit... (much needed these days). As you know I normally underexpose 8 or 9 stops... No wonder my images are a bit weird... (Good job!)
I'm actually seeing generally the Z6III has less chroma noise, mainly reds. The Z6III has bolder, more saturated color rendering. The Z6II does have better retention of the mid tones and shadows, but they tend to be splotchy when recovered. So basically it seems the Z6III loses maybe 1 stop at worst, but gains even more accurate color rendering.
You need to tell us a lot more about your methodology before one can say whether these tests represent anything at all. First, you haven't disclosed the exposure settings - or how you controlled the illumination. Only if the exposure for each ISO setting was the same for both, and at the ISO standard for that setting are we comparing like for like. Also, when you use a tool like ACR, there's a lot of difference under the hood between different cameras. To really know of differences in the raw file you'd need to do a direct analysis of the raw data, or use a neutral converter like dcraw (or successors). (Also, re your header, there is no such camera as a 'Z62' or 'Z63'. From what we know of Nikon's naming conventions those would be DX cameras, if they existed.)
@@BobN54 Hi Bob. Your points are good ones. I’ll tell you my controls. Subject, lens & lighting: all the same. Lens: 70-200 mounted on a tripod. (Cameras attached to lens,on and off) Light LED 5600k Naming: I set the file name prefixes in each camera. It’s sometimes hard to see the difference between ii & iii in small print, so I opted for Z62 & Z63 to avoid and confusion. Cameras were each set in aperture mode at f11 in the first test, f9 I believe in the second set of images I posted. ISO was set to manual. SS was automatically set by the cameras. VR was off. WB set to 5600k I did have a slight difference on WB on the first set of images, as the Z62 had a minor WB adjustment, which I corrected in set 2 of the posted raws. At the end of the day, I couldn’t see much of a difference between the two cameras as they relate to ISO. Thanks again for your feedback!
@@kellysparksphotography I would disagree with letting the camera set the shutter speed. You need the same exposure, which if your light is the same (and it should be with a LED light on the same day, they do degrade a bit over time) and the f-number is the same then the shutter speed must be the same. The reason for this is that image noise is mostly related to exposure (times sensor area - but that doesn't matter comparing cameras with the same sensor size). If you centre the meter, then the ISO determines the exposure, at leat in theory. However, the ISO standard allows 1/3 stop latitude, so for instance anything between 712 and 891 can be reported as 800. So a camera manufacturer can set their meter (and processing) to 712 and look better than one that sets it to 891 simply by giving more exposure. In real life, when you're interested in which f-number and shutter speed you can use in the available light, this doesn't help at all. So really, when making this kind of comparison you need to ensure the exposures are the same, which means setting them yourself. You have to choose one meter or another as the master. This also tells you how the relative ISOs are.
@@BobN54 To do it that way, I would need to vary the light output instead of shutter speed. In the end it really wouldn’t change the iso noise.. and it take a lot longer to run the tests. But feel free to run some of your own tests.. I’m satisfied that in the end, there is very little difference in iso performance between the two cameras. My test was performed in about 20 min all in.
@@kellysparksphotography You're misunderstanding what I said. I said 'You need the same exposure, which if your light is the same and the f-number is the same then the shutter speed must be the same'. The same light, f-number and shutter speed results in the same exposure - Photography 101. So, keeping them all the same keeps the same exposure. So you need both cameras to have the same settings - not to change the light. What you'd do at each ISO is meter (using one camera, or a handheld meter) then set both to the same settings. It doesn't matter of you do that by changing the f-number or shutter speed (reciprocity) so no more difficult that what you're doing. If the meter settings from teh cameras are different, then it indicates either that your metering technique is bad or that the cameras are actually using different ISOs, even though they claim to be the same. You say 'it wouldn't really change the ISO noise' that's part of the problem. There is no such thing as 'ISO noise' - it's a commonly used phrase but actually meaningless. What there is is noise caused by decreasing exposure (which is what happens when you raise the ISO and keep the meter centred). As I said, in practice what matters is the quality that you get when you're f-number, shutter speed and light are all constrained, not what you get at some number of the camera dial - so a proper comparison needs to work at the same exposure.
@@BobN54 I guess I don’t understand then. To test iso ranges, something in the triangle (SS or aperture or the light) must change to get a proper exposure. Or the image will be too dark or too light.
thanks for both sets of images - it seems that in the second set you fixed the WB and the exposure to match on both cameras - the first set of images in fact differ in WB, EV and the noise reduction is set in camera - from Photons to Photos the difference in the photographic Dynamic Range is between 100 and 636 ISO with the Z6III worse of about 0.8 EV - if there is any difference you would see it in that area - maybe would be helpful is you provide a few images with the same technique you shot the second set.
@@anddmt thanks for writing. After a bunch of tests, I just don’t see a significant difference in iso performance between the two cameras. Now I am working on testing the focusing system.. specifically testing bird detection.. more to come.
Some argued that the Z6 III video specs put it on par with the Sony A7SIII. I’m not seeing support for that argument in any videos. The complete video specs are awesome but the ISO and DR of the Z6 III are its weakness in comparison. Thanks for the idea.
Thanks for this. The differences are marginal and won't be apparent in most real world situations at lower ISOs. I plan to get the Z6III so all these vids are a great help.
I don't get it. Is the measurement from photons to photos accurate or not? I'm not talking what do we see in the photos. My S23 is much slower than S24. But that's true in paper. In real life you won't tell the difference. But the measurements are accurate. So are the measurements for Z6 III accurate or not. That's the question. And numbers never lie even though they don't tell the whole truth. They tell their own truth. So what's gonna be?
@@georgemcr1802 as I said in the video, we are nitpicking at 100%. In reality there isn’t much of a difference between the Z6ii and Z6iii on iso performance. IMHO.
I want to mention something as someone who shoots a lot of fashion and gives a perspective most people here won't have. The Magenta in the extreme underexposure comes from IR pollution. You will see this in much less extreme situations with a lot of special, often awful expensive garments and this is a hell to edit out. This is even more pronounced by a couple of flash brands/tubes, something no one is really testing. It is also a huge problem with a lot of Softboxes, they can even empower this behavior. Means with the right garment, you can see an awful magenta shift at iso100 without underexposure, just from the combination of the camera sensor, softbox, flashtube and black garment. If the sensor has now a better IR Cut filter, this can lead to a bit more noise in super weird underexposure tests, but for fashion photographers this can mean hours of less editing in normal shoot conditions.
Is the Z6II "clipping" the blacks because its lower noise let's it reach 0 in the shadows while the higher noise in the Z6III prevents it from ever getting down to 0?
I downloaded raw. Check ISO's 6400 and 3200. You have been used different exposure mode ("spot" on z62 and "matrix" on z63), so you have different SS (on z6iii lower), so this is not relevant comparision.
@@kellysparksphotography You see on your screen not RAW (you can’t see a raw image, btw). You see jpeg generated by RAW converter you are using or embedded in raw. A SS affects to amount of light, that hits the sensor. I think it’s a basics.
@@SmartCaster the common denominator in the images is iso, and aperture, and the amount of light. All constants. The differences are the sensor and ss. I controlled iso, aperture and light, as I think was appropriate for this test. I didn’t care about SS, as that was determined by the sensor. I did have a different metering method, which resulted is slightly different shutter speeds, but in the end I think the tiny adjustment to make them the same was negligible. Oddly, the white balance was set to 5600k on both cameras, and the color was slightly off between the cameras. As I said, this is all nitpicking.. and in the end there isn’t much of a big difference..
@@kellysparksphotography I have no purpose to convince you of anything. Can I ask you to redo the test for 6400 or 3200 (set the exposure camera settings to the same and turn off noise reduction)?
Honestly that’s a strange result, the z6-3 should definitely have more noise. BTW, the sensor is also a bit sharper which tend to confuse people about the level of detail retained at high iso. The z6-2 should retain a bit more detail even when it’s softer- that’s also a result of the slightly higher dinamic range. Moreover, the Z6-2 has a magenta cast. I don’t have the z6-3 but people seem to find a green cast to it instead.
The Z6III background noise looks better to my eyes. I'm more concerned with the color shift. Background colors are definitely different. Maybe it was settings or color profile, IDK.
@@JET-Photo some of the interpretations are subjective. The color could be managed in post. As I said, at 100% we are nitpicking.. not much that would be seen unless cropped way in.
@@kellysparksphotography Agreed, color can be fixed, just wonder why there's a difference. It's subtle but I could see it. Noise isn't an issue these days anyway, but I think the newer processor is helping the Z6III to be slightly better than the Z6II. Both are great cameras.
It could be that on the Z6III due to the Expeed 7 processor some kind of AI AntI Noising and sharpening is happening, nevertheless a very impresseing result for a stacked sensor, maybe a lot better then with the A9III.
The Sony a9iii looks horrible in the ISO 6400-12800 range. This is an iso comparison site, unfortunately the z6iii is not in their test, but from what can be seen in this video, the z6iii is better than the z6ii at certain iso values from 6400 and up. The z6iii keeps its colors even at high values without turning them into other shades like the green z6ii. It matters if the picture control of the camera or adobe picture control was used in the test and if the sharp Lr is on 0. If you are really not satisfied with the iso, you can reduce from the camera profile-medium sharp (-0.5-1) and you will have a better iso.
Your findings bear out what I discovered when playing with the Z6iii at my local camera store recently. I was shooting at 5000, 10,000 and higher, and the files definitely looked cleaner than the Z6ii at the same ISOs. 👍👍
I left a comment on your dynamic range video explaining I had been running my own tests and they mirrored your DR results. I also said you should do an ISO video as I found the Z6iii was handling the higher iso's much better than the Z6ii. My results were the same as yours, thank you for confirming things. I shoot sports (mainly football and tennis) and I have found the z6iii is spitting out really great higher ISO images. My real world results were tested last week at the Wimbledon tennis championships and I was very happy with the files. Keep up the good work.
I'm almost thinking, given the body of evidence that concluded the Z6 III is slightly worse, that maybe you accidentally switched up the labels, and it's actually the Z6 II on the left. Assuming you shot RAW and treated the files in the same way.
The new Z6III native ISO ranges to 64,000, so that is "the highest ISO with a number." The Z6II range does indeed stop at ISO 51,000, which is a major reason I upgraded to the III. Thank you for this video. It provides more results for people considering the abilities of these two cameras.
@@lawrose4 yes. I had a shot at 64k for the Z6iii, but didn’t have one for the Z6ii, so I couldn’t compare. Things were so bad at 51.2k I didn’t bother..
@@prokremelskidezolati1426, that is .33 of a stop more gain. In critical situations, and what situation where you would use 64,000 ISO isn't "critical," that can make a difference.
Not sure if there is a problem with RU-vid? But it is hard to tell any difference in 720P? Based on this video looks like the Z6III is actually better than the Z6II. I was waiting for the Z6III but think I will just buy the Z8 instead. Thanks for sharing 👍🏼🙏🏼
Interesting results, on your test it certainly seems that the Z6III manages higher ISO's a tad better than the mk II. I'm inclined to think that the dynamic range tests carried out by other RU-vidrs perhaps reveals more about the Z6 III sensor's " invariant" capability compared to the Z6 II. That is to say that compared directly and assuming exposure is correct to start with that the Z6III improves upon the Z6II in purely higher ISO capability. That would make sense given it us utilizing a more powerful expeed 7 processor. Where the Z6II beats it is in situations where you have to recover a poor exposure in post be it due to under or overexposure. It seems that the Z6II just has a tad more headroom in post processing. I own both the Z6III and Z6II. Even if the Z6III performed less favourably than the Z6II in terms of ISO and dynamic range I'd still take it over the Z6II purely for the improved AF, better EVF and faster fps not to mention the faster sensor readout. These aspects are more use to me in real world image taking than being able to obtain more dynamic range. I very rarely shoot above ISO 6400 anyway. Its also extremely rare I would ever have to recover any image 5 stops or more, heck if you cant get exposure right with WYSIWYG in the EVF then perhaps photography isnt for you 😋
Exposure for newbies..put it in AUTO😂 and if YT newbie surfers complain about camera A vs camera B Dynamic range..then they should go get a Dslr to learn more and hone their skills…more cheaper to learn the basics of photography
@@TheJugaadProduction thanks. Current the only way to save footage to an external SSD, is via HDMI. I have a NinjaV. It would be awesome if Nikon would enable USBC to SSD, but that isn’t a feature currently.
Tony just wanted to get some clicks/integrations to feed the algorithm. He chose a composition with an overexposed subject and a barely exposed subject and pushed that because he knew that the resulting exposure required to not blow out the overexposed subject (1/80th at ISO 100) would tank SNR on the shadows given the limited number of photons that could be captured. There are so many other examples of real-world photos, including unevenly exposed subjects such as in wedding photos he described in this video, that clean up very nicely on the Z6iii. His example was the most outlier of outliers. I started watching his videos back in 2011, specifically his star trails videos. I admire his passion, but he, like many other big name photogs, sold out to the algorithm and sponsorships. He might say he makes all his money off book sales and that he just has our wallets in mind when he makes such videos, but, sad to say, I've not bought it for a long time. His methods are based on science, but he's performing the most obsurred test that can be devised. It may have given the nod to the Z5, but consider that the QE (Quantum Efficiency) of the Z5 is much lower than the Z6iii given the latter is a BSI sensor, and you realize he's really pushed the limits of any sensor. (Sorry if I repeated a bit but it's late) In short, while other sensors may "win" under those conditions, how many photographers will find themselves taking photos in those conditions, with those exposure settings, and then pushing their files that much. What I know is the Z6iii and other Nikon cameras produce amazing images but compared to the Z6/ii Z7/ii Z5, the Z6iii's AF is in a league of its own and that is a difference I know I will see in all my images.
Tony has left the chat, lol. Great test! I'm a working photographer and pay my bills with my camera. I felt like his test was nonsense and you showed exactly why. Kudos!
This is a total fail. This is not how you test DR or noise. You could have saved all that time and looked at Bill Claff results who actually knows what he's talking about.
The Photons to Photos site that Toneh points to in his "test" actually shows that the Z6III matches the dynamic range of the z8 & z9 up to ISO 400, between 400 and 650ish the Z8 and Z9 have slightly more dynamic range, but by ISO 800 and onward they all match up again. But I don't hear anyone complaining about the dynamic range in the z8 or z9. Sure, the z6II, z7 and Zf seem to have consistently slightly higher dynamic range but, again, by ISO 800 the z6II, z6III, z7, z8, z9, zf ALL have roughly the same DR. The Sony a7III lags behind all of them between ISO 630 and 1200ish before catching up. And as you and many other folks on RU-vid are now demonstrating the real world differences in DR if you are not shooting magazine pictures are negligible.
Bill Claff knows how to test, unlike this utuber. And no, z6iii and z8/9 don't have the same DR, z6iii (10.44) is almost a stop behind z8 (11.31). DR at high ISOs rarely matters, it's rarely about highlights clipping (and if you care you can always use lower ISO and correct in post). At high ISO's SNR is what matters, and here again Z6iii is about a stop behind Z6ii. The same is true about z8/9 at high ISO, but at least they have ISO 64 so they manage to get as good DR as Z6ii.
Tony Northrup has already blocked my comments after I wrote some negative things about his reviews. This just shows that this man has an agenda and cannot be taken seriously.