Tony I have been following you for almost a year. I have yet to say a word till now. But it’s time now. Tony, plainly you are simply the “Man”. Let me define this. The “Man” tells people just how it is, the pros and cons, doesn’t smooth over things that might god forbid “offend” people and most of all just tells it HOW IT IS WITH RESEARCH and SCIENCE to back it up! I don’t understand how anyone could want anything different or more and still be true to themselves. You are true to yourself and it’s easy to see. You do an incredible job and so I will always be with you learning what the “master” can teach us that are willing to listen.
How cool that they updated the website. I stopped going to the site out of frustration when trying to compare more than one camera or lens. Thank you for this video.
This is a great guide to the testing data from DXOMark. Now begins an arduous re-evaluation of my equipment, followed by buyer's remorse and urge to buy new lenses. If only this were a simple world and the DXOMark score would answer all questions!
Thanks Tony for directing me to this excellent tool and showing me how to use it I am not sure how I missed this most information packed video and noting total score does not tell enough of the story... There is a lot of information in your movie it will take a few more hours/reviews and I feel it is worth the time. I have talked with a lot of friends and family who have spent some serious coin on hardware and still only shoot jpeg how much of your teachings are directed towards or expecting some post processing of the RAW file to get the best results. And yes it is time I get your books so I can get the BIG picture:) Phil... PS :Thanks again for sticking your chin out as professionals you and Chelsea set the bar high and always include the avid armatures in your overviews talking us up not down:)
DxOMark actually is the reason why I mainly have Sigma primes today, and not Sony lenses. Sharpness Field maps rock. I even marked all my lenses with stickers that tell the interval of F-Stops the individual lenses have their sweet spots, so I know what F-stop range i have to work with when throwing them on my cam. Also Great tool for comparing and purchasing. Totally on the same page as you Tony. Cheers.
I found this place when I tried the DXO one camera attachment to the iPhone. Great info on the site. reminds me when you purchased a lens many year ago, you could select a lens based on the lens specs. Thanks Tony, you have made the site more meaningful now that I understand it more then before.
Thank you !! That was very useful. I kinda guessed what the sweet spots on my lenses based on trial and error. This is a great resource helping me get the best images with my lenses of choice. Thanks for explaining.
Roger Cicala over at the extremely useful "Lensrentals" site has this to say about "Perceptual MegaPixels" (P-Mpix as DXOMark describes the idea): "...my major problem is it gives a number that takes into account: 1- lens performance at all apertures, 2- lens performance all across the field, 3- for a zoom, performance at all focal lengths, 4 - a formula of calculation that is not shared (but apparently is normalized down to a lowest common denominator, than multiplied up to a sensor in question, which is pretty insane), 5- a method of measurement that isn't shared and cannot be confirmed or denied by anyone else, and 6 - for lenses that work on multiple cameras and mounts the calculations are done on one camera and mount and therefore aren't valid. So we have this number that equates to something but we aren't sure what, based largely on measurements that we definitely don't know what, and people are making purchasing decisions based on that. More to the point, people are misinterpreting it to say things like 'this lens is rated at 16 megapixels so if I buy a 42 megapixel camera my images won't improve' when, in fact, they will."
Good explanation at ~27 minutes about the dynamic range on ISO differences. I've been shooting with Canon and lately upgraded from 60D to 6D. I managed to check out the competition for the fun, eg. Nikon D600 and noticed on Dpreview.com that it performed better on dynamic range. This wasn't the case at higher ISOs which the sample images proved, and you finally explained why it contradicted with the dynamic range score. Thanks Tony!
Your video certainly cleared up DXO's site. At first you don't really know what you're looking at. Sometimes it's easier to watch someone go through it on a video than plowing into it with no previous experience.
The problem that I see is not that DxO mark is wrong but that there is variation between individual copies of lenses.. so its hard to tell if DxO just used good copies or if you'll just get a bad one. Tony Northrup
In Advanced search the focal range selection works kind of logically. It picks equal or higher value available for minimum focal range. It do not search if the lens is able to work in specified focal range. Overall the change they made is nice.
Thank you for this video. So "stuffed" with useful info. Thank you for teaching me to take a deep dive into the measurement data for lenses. Thank you for teaching me that the overall score is a "blend" of parameters where the weighting is subjective - will use the underlying data going forward. And the T-stop point is should be marketed - but maybe the vendors do not want the truth to be told.
Here's a question about "lens" sharpness measurements... As someone below mentioned, DXOMark, when testing a lens, tests it on a body, and that's fine if you're looking at what you can expect given the body you have. What I wonder about is that if we're evaluating a lens by itself, is there any good way to evaluate the extent to which the lens itself vs. the sensor+processor is the limiting factor in being able to resolve a given resolution? I mean, you have the fundamental issue that if you have, say, a 16 Mpix sensor, then in practice, you are probably limited to ~13 P-Mpix even with flawless optics. Yet, when an M43 lens rates 7 or 8 P-Mpix vs. an FF lens on a D810, these two results are put side by side as if the scale is absolute in spite of the fact that you don't really know whether the M43 lens is limited by the M43 sensor or FF lens is "carried" by the high-res FF sensor. I can think of a number of ways in my head to resolve this variation in a more objective way, but of course, the detailed data that is necessary to do that isn't available on the DXOMark website. Alternatively, if DXOMark would actually do tests with adapted lenses, that might give some insight. Also, in regards to your initial point about "I got this one picture super sharp"... it does depend on the subject. Certain subjects -- say, a very fine checkerboard pattern, it isn't even possible in principle to resolve more than 4 Mpix on a 16Mpix sensor because of the RGB mosaic and so on, regardless of how good the lens is. Note that 4Mpix here refers to the checkerboard having 4-million black&white blocks visible in the image. And even that result relies on being able to align the checkerboard cells perfectly with the photosites on the sensor and so on, which is basically impossible in practice. Also, the nature of aliasing is such that even if a 4 Mpix checkerboard resolves perfectly, a 3.99 Mpix checkerboard will not because of those same aforementioned alignment issues. By comparison, a super-sharp photo of someone's cat can look super sharp with some given lens because of the fact that this sort of extreme contrast between neighboring pixels doesn't often exist. It seems that the MTF measurements used to relate to P-Mpix measures at the limits of resolvable contrast. So even if a given MTF shows that you can only get about N P-Mpix when looking at variances over the entire dynamic range of the sensor (hypothetically, say, 12 stops for some sensor X), many real-world images might only have details with 3-4 stop variance between pixels, which means that the same MTF results that got you some N P-Mpix will still be able to resolve that comparatively "low contrast" detail at a higher resolution. So it's often not that straightforward to relate that figure to the real world results exactly.
Hi Tony , I really appreciate the amount of time and detail you go to, no one else does this at this level, I entirely agree with your summing up of how to evaluate CAMRA and lenses using ISO sharpness and dynamic range , thanks once again to you and Chelsea and the team , dave newby
I know you're getting a lot haters, but just know that you're doing something right. Great quality videos (literally) both spoken, explained, and visually. Always room for improvement, but I know you know that and I appreciate you're insight. Better than a college professor and more content, more often!
I agree that DxO marks are very accurate and useful. HOWEVER, it can make you 2nd guess your camera and lens combination, and convince you that you have crap for optics. To get over that I let my lens do the talking and got sharp, crisp results. I don't have any L lenses, but I do have top of the line Sigma and Tamron. So I think DxO marks are based on consistent performance also. Whereas I can get sharp images with my lenses, just not at a consistent rate as L lenses
in regards to focal lenght filter: it filters out all lenses which have higherfocal length than set by filter, ie olympus zuiko 75-300 won't be included in results when you'll set your filters between 100-300, but canon 200-400 will be shown in results. weird, i know :)
Hugely useful and opened a whole new window on photography for me. It also helped me to avoid buying a particular upgrade by revealing that was no significant optical difference between three generations of a particular lens. As a very ex-electronics engineer understanding decibels was the least of my worries!
Helpful. Sure wish you would return to this format vs. the one that's more fun & games, i.e. entertainment. Please read this as positive feedback and thanks for all the work you've done.
This is a very informative channel, one of the best in youtube... but there's way too much focus on the equipment. A great photographer is the one who knows how to see the light, to compose a photography and only then who uses his camera the correct way... some types of photo styles might rely more on the hardware (like for sports) but for most types the hand who shoots the photo matters way more
Most of our videos are not about equipment... Just check our channel page. Ironically, people who think we're only about gear are the ones who choose to only watch gear videos.
Do this video explanation again during your weekly LIVE show, when everyone has had a few to drink. It would be a blast to try and understand it plastered.
Tony Northrup Hey Tony, I think a lot of people who are complaining about color depth are actually seeing color data lost due to poor dynamic range performance. Lost data is that, simply lost data. That said, I can tell the difference in Jpeg files from camera or lightroom -vs- RAW files from the camera. Even exported to 16bit Tiff there is noticeably more color range. This is not the cameras fault though, its due to Jpeg being 8bit. Note, Jpeg standard was updated to 12bit recently, but that update has not made it to Lr or our cameras at this time.
Tony, Just started really getting into your videos, which are excellent. They are really opening a whole world to me, especially this one regarding DxOMark and how to use it to compare cameras and lens. Also, how to find a sweet spot for a given lens is very valuable. However, I am confused about the Vignetting rating which shows a negative EV rating like -1.5 EV. This is said to mean the corners are 1.5 EV darker than the center. This seems backwards to me. A smaller EV number is gotten by using a slower shutter speed and/or a wider aperture opening. That should result in a brighter image, not darker. So I would say the corners are +1.5 EV darker. Where am I going wrong? This is not a big deal, but I am trying to get it right. Thanks again! PS: I am an old engineer and IT guy so I do understand math - at least I thought so!!
Color depth is a real issue. I know this is an old video but I've ran into issues concerning color depth doing macro work and some pet photography. Now I think it really comes into play with a camera lens combo from what I found. I won't mention manufacturers as I didn't go into why it was an issue but merely changed to another brand which resolved the issue right away.
Tony, you have made a valued tool "DxOMark" more comprehensible and therefore of greater value to me and I'm sure many others. The mark of a good teacher is to make a complicated idea or concept understandable to others. Here is where you excel in communicating in ways that are tangible and clear in your TH-cam videos. You demonstrate the potential that the internet promised but so rarely delivers. I continue to learn from your generous sharing of our knowledge. Very Kind, thankyou!
Good overview Tony - thank you. Only two things: 1. When comparing T-stops as in your example, 4.2 to 5.1, it is not 1 T-stop. The next T-stop to 4.2 would be 5.9. (see typical stop-row: 2.8 - 4.0 - 5.6 - 8.0). T5.1 compared to T4.2 is only about half a T-stop - even if the numbers difference is almost 1. 2. Distortion: It is true that distortion is auto-corrected in post process, but it should be mentioned that distortion-correction affects the sharpness in a negative way. Similar as vignetting it does not influence the center, but it does at the borders. Do you know if the distortion-correction is already considered by DxO-mark when looking to p-mpix?
Thomas Feldbauer 1. Yes, and I wrote a title over the video to explain that mis-speak. 2. Yes, I mentioned that. I'm sure DxOMark doesn't factor in distortion correction.
Very informative...It is very nice to be able to understand DxOMark better. Thank you Tony. *you would think that if someone has a problem with the information that you provide us, that they wouldn't bust your balls own your channel, but rather set up their own test on their own channel and try to prove you wrong.?
Wow, the way Tony teaches is a perfect match for how I learn. I feel like I want him to keep throwing more and more at me in some sort of teacher -student challenge until I call UNCLE! I had an economics teacher in college that elicited the same response from me. He was some guy named Lester from MIT. What's fun is that I'm a highly analytical person but at the end of the day I just want to take great shot.
DXO website has gone from bad (as in pretty confusing) to worse with their recent make-over. Small wonder Tone pulls up two browser windows to compare 2 lenses. That used to be easily done within a single window on that site. When you filter down say cameras and want to look at the best re ISO, you get the first three places in a little box right in the graph. But ... all the rest are given as per their launch year (how silly is that) and *not* with their ISO indicated but their general sensor mark, which Tony rightly says is rather superfluous. I really wish they introduced a more sensible design at DXO, it could be such an invaluable site.
On DXOMARK the D5's low light score is even lower than the D810. However, even when viewed at the SAME SIZE, the D5's pictures are significantly cleaner. In fact, for high ISO, the D5 is a very, very good camera for that. Probably even better than the A7SII. Practically noise-free at 12800, excellent at 25600, very usable at 51200, and usable at up to 102, 400.
F-stop is based on the size of the aperture, not the front element, by definition. (If that were true, you couldn't adjust the f-stop of a lens without swapping out glass.)
doubledeej right, it's technically the iris size relative to the focal length. I was just trying to convey that f/stop describes the light coming into the lens while t-stops describe the light that exits the lens.
Tony please. This frightens the hell out me! I watched this video and then looked at DxOMark tests to see what lenses would be best for my D500. That is, to find lenses where the PMP equalled the actual MP of my camera. A sharp intake of breath and a Gerry mouse style gulp later, I find that unless I have a few grand to spare on each lens I need, I may as well use an empty Coke bottle with a bayonet mount!.. Ok, just a fun response from an English guy with a weird sense of humour..
Great stuff. When you said you were bringing it out I was thinking more post-summer! I don't think you've gone into enough detail to convince those who dislike it for whatever reason despite saying to totally disregard the ballsed up subjective formulas but appreciate the effort gone to here to help people past the fluff and onto useful info. Those who are scared by the numbers latch on to the idea that it's garbage and those who could understand them if they looked analyse the rubbish seen in clickbait articles then on page one of the site and (quite fairly) never give it a chance beyond that. A couple of tripped over words that will no doubt be latched on to and perhaps a couple of opinions that will be mixed in with the facts you're presenting here (I do agree with ISO holding the most value in sensor rating but disagree a little in what you're saying about print ISO being whole story when you yourself promote cropping (though I obviously see where you're coming from)). Not sure the disclaimers for photog>>>gear and you shouldn't be pixel peeping are going to make it through to enough people either but we can be optimistic. With the video coming up to an hour I understand why you may have edited out some more detail which people will stumble upon by clicking through the measurements tab. Would be really curious to see stats for how many got through to near the end. Also value you not breaking it down into 3 separate videos etc. I'd argue that even the ISO score loses a lot of its value when they combine it to that 'sports' rating. Would be nice to see that broken down into colour, luminance and test shots but that's asking a lot of them when you do pixel crops vs print crops too. Besides, DP review tends to have a relatively accurate way to compare these subjective things if you're really looking at a new body. Also along with that focus breathing vs effective field of view it's also worth mentioning that even on some high priced primes that distortion often adjusts as you move through the focal range. Would be nice to see an extra few measurements on that field map tab to click through. (Though I admit it's not really fair to be asking anything of them considering what people pay to use it - nada) I can't see them ever moving away from things like their overall score because that's what the media use in their sensationalist clickbaity articles as you said, which drives more people to DXO mark (even if only long enough to look at the front page and rightfully decide those first few scores are pretty garbage) Anyways, good job and much respect for spending the time on a video that won't generate nearly as many clicks as some of the others.
James R Thanks, James. For the record, my goal isn't to convince people to trust DxOMark if they don't now--there's definitely no way I'm going to change those peoples' minds. I just want to help everyone else understand it, because their website is hella confusing.
This is the single most useful camera *tech* video on the web. Spot on. Explains perfectly what is unique and useful (a lot) and not useful (also a lot) on DxO. FWIW, you can see that Leica doesn't lie about their ISO performance like the others. Leica T at 3200 indicated is 4100 ISO while the D7200 is at 2300. I've never seen a camera reviewer bring up the "this one goes to eleven" problems. You can't find that info anywhere else. (Suggest use 'Select' and "Compare camera"/"Compare lens" to overlay 3 of each on same chart. New menu requires another step, 'View comparison selection')
Most helpful video I've seen in years! Before seeing it I was always frustrated when trying to navigate through DXOMARK. I don't understand why they don't they improve their website.
I don't have a problem with what DXOMark measures. My issue is that there are things that they don't measure and may not be measurable, like micro-contrast and color rendition. Looking at DXOMark you would think that a Sigma Art lens was just as good as (or better than) a Zeiss lens, but the Sigma lenses lack that Zeiss pop and color rendition.
Tony, just a few comments while I watch your video ... The DxO color depth IS useful. It is DxO's way to measure color noise. All other metrics are black & white only. W/o this metric, vendors could have very wide color Bayer filters which produce a lot of color noise and possible bad CRI and rock DxOmark. DxO prevents this by including the color depth metrics.
Omesh Singh Camera A+light_source+Camera B Maybe not that crazy ,after all ...if you want to measure ISO with exactly the same conditions and time stamp!
I really wish DxOMark had more lenses to compare as well as more lens/body tests. They have a really incomplete list. I know it isn't easy to test as many lenses as they have but there was an website that I can't seem to access outside of waybackmachine called photonotes that had way more lenses and comparisons you could actually look at yourself. The comparison page doesn't work on waybackmachine but it was a really awesome website that seems to have died.
theduckfu actually the site photonotes is just down because of internet issues. It should be back later check it out they have a ton of EOS samples to compare lenses. It is incredibly exhaustive. Also I prefer looking at actual samples at different apertures then a 9 or 10 sharpness rating. I don't know how the 9 or 10 is quantified and how the graphs actually look on performance. I do know a lens can very greatly being wide open vs being stopped down a few pegs.
Camera von Theuns Yeah, that anti-aliasing filter makes a big difference! I did a bunch of research on this for this: th-cam.com/video/Ww6QGpryrLM/w-d-xo.html
I've watched many of the Northrup videos and I believe this one to be the most educational, overall, than any other video. It may be focused on dxomark but in explaining the site Tony teaches about many important aspects of photography. Right on!
I was debating about upgrading my Tamron 150-600 G1 to a G2. The data from Dx0Mark and how to interrupt from this video has helped me decide it is not worth it. And actually, if sharpness is the only consideration, G1 is better.
I think this is a great presentation, love the clarified detail and recommendations, a great service to photographers and DXO. ps What do you think of their DXO software?
Thanks Tony - great video. I always wonder how far we should take these scores. Taking say the 6D as an example, the highest sharpness lens with that body is rated at 17 which I guess is broadly expected for a 20 MP camera. Does that mean that there is no point mounting say a 50mm 1.4 over a 50mm 1.8 from a pure sharpness perspective? I guess the argument is academic for the 6d because the speed difference is enough to justify changes but if you take the same logic and apply it to the A7. It being a 24 MP camera would indicate its unlikely to score much above 24 so if there any point in buying the expensive 55mm 1.8 which is a much sharper lens over the 50mm 1.8. According to DXOMark the 50mm is already going to give a resolution higher than what the camera body itself is likely to support (26 when mounted on the A7R II). Are A7 users better off buying the 50mm and saving money? Obviously if you have an A7R its a different question.
T&C... Just let the haters hate. Don’t give them energy by reacting to them. I think it’s really cool that you do stuff like this. :) BTW. Still waiting for that EM-1 video. 😂✌🏼😇
DXOMark also doesn't include the capacity of the sensor to produce details for a given lens, if you take for example the D5300 with nikon 35mm 1.8G DX the Perceptual MPix are 13 MPX but only 12 MP mounted on the D7100 (The D5300 Sensor can resolve more details this should be included on the final score), the gap should be more obvious on camera like Samsung NX1 with their 28Mpix BSI Sensor a lens like the Samsung 50-150mm f/2.8 S ED OIS (Not Tested Yet on DXO Mark How convenient) can produce a lot details coupled the the NX1 Sensor www.imaging-resource.com/PRODS/samsung-nx1/FULLRES/Y_SAM1533.HTM you can see that even at 1:1 view judging at pixel level, there is a lot of details proving that samsung can produce lenses that can have more than 24 Perceptual-Mpix, thus using the sensor full potential ....... anyway a very informative video as always i learned a lot ..... keep up the good work
Tony Northrup I have an honest question regarding sharpness using a full frame lens on a DX sensor. The DX sensor takes the image from the center portion, theoretically sharpest portion of the lens. Why then does a Nikon 610 (24mp fx) have a higher perceptual megapixel count that an Nikon 7100 (24mp dx camera) when using a Nikon 85mm 1.8G. Shouldn't theoretically the Nikon 7100 have a higher perceptual MP (sharpness) count than the 610 fx camera?
Thank you Tony!!! The DxOMark website SUCKS(I mean suck at browsing it:-) ! They might have great data, but I cannot even use or make sense of it....I'm all ears as I start this video!
i still do not understand why 7d mark II has much lower score than 70D with the same lens. I thought they have the same sensor. Is 70D sensor better than 7d mark II?
At 12:24 you state “DXOMark is really smart at how they score this (a very untrue statement)…..there is a good reason why (the D810 score is twice that of the D5500, which is also the same sensor in the D7200 by the way) the D810 scores twice as high, since the sensor is about twice as big ….For any given ISO it is gathering twice as much light (not true point 2)…and its that total light that makes the biggest difference in your image quality (not true point 3)” ISO in digital is a time sampling RGB signal gain and has a native limit at which point signal amplitude is applied. No lens on earth “cares” or “knows” what sensor is underneath it, the angle of interception of the circle of rear element of projection (Field of View) changes, but this has no effect on luminal density (LUX) per square mm of sensor. Photosite GAIN (less gain, and higher noise on DX photosites, which are smaller) has been mitigated by SNR time-elimination algorithms. Radio astronomy (light is EM, radio is EM, EM is EM, regardless) has used this software ‘trick’ now for many years, its use is now standard in interferometry. As for “image quality”, this confuses sensor level SNR (noise and gain) without applying SNR compression althorithms used by Nikon and others which have allowed for low light high ISO performance on DX sensors such as the D7200. Image quality is also relational to pixel density. The DX sensor, per square MM, while having less gain, has more translational data, which is why nature shooters are using 800mm lenses (for example) on DX sensor cameras. Likewise the lack of smaller DX photosite gain has been mitigated by advanced time-compressional SNR algorithms in the digital processors AFTER the sensor and AD converters but before the DSLR buffer. To speak of “quality” is obtuse, and leaves out 1. Rendition 2. Pixel density, 3. SNR gain algorithms, and other criterion. This is why the next generation of all digital cameras are FF sensors with DX pixel densities (ala the new Canon 50MP sensor camera,…which happens to be a 62MP sensor, but down-sampled to 50MP to allow for faster buffer read-writes). This incorrect conclusion assume two things which are wrong: (sensor) size is relational to luminal density, however no sensor is gathering more light, rather has (or can have) higher GAIN due to photosite size as present on FX sensors. Digital camera sensors contain ‘microlenses’ above and around each photosite to enhance their light-gathering ability. These lenses are analogous to funnels which direct luminal capacitance into the photosite where the light would have otherwise been unused. However as is the case, larger photosites with greater gain have a downside (again, the same thing was ‘discovered’ in radio astronomy using large dishes), which is larger photosites gather more unwanted peripheral luminal noise not desired to be directly captured. As such, the statement that “DXOMark is really smart” is misleading along a wide spectrum of empirical facts relational to FF and DX sensors due to GAIN, SNR algorithms, translational data per square mm (pixel density) and luminal density per square mm which is unchanging between FF and DX sensors. Likewise, again, no lens projects more or less light due to FF or DX sensor being behind same. Angles of interception have no relational inverse-causation on luminal densities.
Pat Fitzpatrick Heya, Pat. your understandings are common, and I thought the same things a year ago before I started doing all this research. Here's some info about noise and sensor size, with real-world data and examples to support it: th-cam.com/video/0OtIiwbAZi8/w-d-xo.html If you want more, here ya go: th-cam.com/video/f5zN6NVx-hY/w-d-xo.html and th-cam.com/video/DtDotqLx6nA/w-d-xo.html and th-cam.com/video/6Im4W_9blhY/w-d-xo.html Regarding using smaller sensors for wildlife photography, here's some good info: th-cam.com/video/YDbUIfB5YUc/w-d-xo.html Regarding pixel density, here ya go: th-cam.com/video/_KYvp8PrCFc/w-d-xo.html If you still disagree with something, I just ask that you create real physical evidence by performing experiments, and send over the images that prove your point.
I also hope that DxOMark figures out a standard to measure auto-focus capabilities of different camera mounting different lenses. I can't be the only guy who consider auto-focus accuracy more than perceivable megapixel, right?
You need to update this video. I went to DXOMark and their interface is completely different now. It seems to have been dumbed down quite a bit. The access to data is very limited. I could not access any information about third-party lenses. Their search filter had only a handful of brands, and none of them were third party ie Sigma, Tamron, etc.
I respect DXO and there figures, but like Tony said, it can only be used as a small part of your buying decision, if they say one sensor is better than in another camera, you still need to way the other features of the camera, and the price, and weight and focusing speed, and many other things in your buying decision, and same thing with lenses, maybe one lens wins buy one point in a DXO test, but the other lens has IS, or one lens is weather sealed, or focuses faster, or one lens is allot lighter, or theres a big price difference, there are so many variables in a buying decision other than the DXO tests, also, personally I always try to test a lens before I buy it, using my sd card at the camera store, and then I go home and look at the images on my computer and try to come to a conclusion, about sharpness, contrast, and focusing consistency, and also the quality of the bokeh, and if i like the lens, i go back and buy the exact one I tested, when I bought my 85 Sigma f1.4, i had tried 3 of them, and 3 Canon 85 f1.2 lenses, at three different stores, and the 3rd Sigma was the charm, it does everything great, and it was almost one third the price of the Canon at the time, and I came to the conclusion that it was better than the Canon f1.2, at any price because it is lighter, faster focusing, and just as sharp, but the first two Sigma 85's were not as consistent with the focusing accuracy, and the first two Canon 85's did allot more hunting than the 3rd one I tried, so all lenses of the same kind aren't equal, witch makes me wonder how many copies DXO mark test, anyway now that I just looked at DXO marks sharpness chart, the Sigma 85 actually slightly out scored the Canon L with sharpness, witch made me feel good about my decision, even though DXO wasn't part of my decision buying it, but so many people will buy an L lens just because of the red ring, and also the Canon 85 L is actually f1.4 T stops, but I don't think the difference between f1.2 and f1.4 matters anyway, personally i don't want to shoot shallower than f1.4, and the best sharpness is maybe from f4 to f8 anyway, but it all depends on the look you are going for, but overall I agree with Tony that DXO is a cool site
Hello, great work from you, like always... I use a lot their website and the new update it was quite a shock for me :). But... I'm very disappointed also... they updated the website but they didn't added new tests... I mean there are some lenses that are already on the market for years and they didn't tested them and probably won't ever. The 12-24 from Nikon it doesn't even exist on their website... Maybe the fact that they have the only website of this kind, puts them in position to get some money from brand to test their product... so maybe they have some kind of priority list... Anyway I'm disappointed of them... By the way... if you choose Lenses->Tokina->Any price->Any date... you get no results... For me, the old filter system and list, was more versatile and user friendly
thanx for the video . now im trying out the free trial.. but is a possible way to make a video on how to fix some photos like outdoors.. i will like to learn how to use it very good
8 ปีที่แล้ว
Muito obrigado pelos ensinamentos! Estou estudando a aquisição de uma nova camera e lentes e aprecio muito os seus vídeos, pois descrevem a ciência e os aspectos práticos de diversas combinações de equipamentos fotográficos. Utilíssimos.
📚 Buy Our Books on Amazon! 📚
📕Stunning Digital Photography: help.tc/s
📘Lightroom 6 Book: help.tc/l
📙Photoshop Book: help.tc/p
📗Buying Guide: help.tc/b
Tony I have been following you for almost a year. I have yet to say a word till now. But it’s time now. Tony, plainly you are simply the “Man”. Let me define this. The “Man” tells people just how it is, the pros and cons, doesn’t smooth over things that might god forbid “offend” people and most of all just tells it HOW IT IS WITH RESEARCH and SCIENCE to back it up! I don’t understand how anyone could want anything different or more and still be true to themselves. You are true to yourself and it’s easy to see. You do an incredible job and so I will always be with you learning what the “master” can teach us that are willing to listen.
How cool that they updated the website. I stopped going to the site out of frustration when trying to compare more than one camera or lens. Thank you for this video.
This is the best presentation to show you how to properly do your research before maybe buying the wrong camera or lens. Thanks lot Tony!!
This video is outstanding public service. Thank you.
This was super helpful. I have never gone beyond the overall score of the camera or lens and turns out most of the helpful info is in the details
Also, you know the internet is about data and you really give proper data. Good job Tony!
You blew my mind.
Fantastic video Tony! Very helpful in helping me understand DXOMark
Kit Hoong Tan Thanks!
This is a great guide to the testing data from DXOMark. Now begins an arduous re-evaluation of my equipment, followed by buyer's remorse and urge to buy new lenses. If only this were a simple world and the DXOMark score would answer all questions!
Thanks Tony for directing me to this excellent tool and showing me how to use it I am not sure how I missed this most information packed video and noting total score does not tell enough of the story... There is a lot of information in your movie it will take a few more hours/reviews and I feel it is worth the time.
I have talked with a lot of friends and family who have spent some serious coin on hardware and still only shoot jpeg how much of your teachings are directed towards or expecting some post processing of the RAW file to get the best results.
And yes it is time I get your books so I can get the BIG picture:)
Phil...
PS :Thanks again for sticking your chin out as professionals you and Chelsea set the bar high and always include the avid armatures in your overviews talking us up not down:)
Excellent points about the sharpness in relation to a lens to body with AA
DxOMark actually is the reason why I mainly have Sigma primes today, and not Sony lenses. Sharpness Field maps rock. I even marked all my lenses with stickers that tell the interval of F-Stops the individual lenses have their sweet spots, so I know what F-stop range i have to work with when throwing them on my cam. Also Great tool for comparing and purchasing. Totally on the same page as you Tony. Cheers.
I found this place when I tried the DXO one camera attachment to the iPhone. Great info on the site. reminds me when you purchased a lens many year ago, you could select a lens based on the lens specs. Thanks Tony, you have made the site more meaningful now that I understand it more then before.
Great work tony
Your videos are great. Thanks for putting this together. Very helpful!
***** Thanks, Logan!
Thank you !! That was very useful. I kinda guessed what the sweet spots on my lenses based on trial and error. This is a great resource helping me get the best images with my lenses of choice. Thanks for explaining.
This was very useful for evaluating lenses for my upcoming purchase of a camera and lenses. Thanks much
VERY INFORMATIVE ENTHUSIASTIC !!!
Roger Cicala over at the extremely useful "Lensrentals" site has this to say about "Perceptual MegaPixels" (P-Mpix as DXOMark describes the idea): "...my major problem is it gives a number that takes into account: 1- lens performance at all apertures, 2- lens performance all across the field, 3- for a zoom, performance at all focal lengths, 4 - a formula of calculation that is not shared (but apparently is normalized down to a lowest common denominator, than multiplied up to a sensor in question, which is pretty insane), 5- a method of measurement that isn't shared and cannot be confirmed or denied by anyone else, and 6 - for lenses that work on multiple cameras and mounts the calculations are done on one camera and mount and therefore aren't valid.
So we have this number that equates to something but we aren't sure what, based largely on measurements that we definitely don't know what, and people are making purchasing decisions based on that.
More to the point, people are misinterpreting it to say things like 'this lens is rated at 16 megapixels so if I buy a 42 megapixel camera my images won't improve' when, in fact, they will."
I appreciate the work that goes into providing a free resource and a platform for discussing matters of interest. Keep up the good work.
Thanks Tony. You are the best ever
Good explanation at ~27 minutes about the dynamic range on ISO differences. I've been shooting with Canon and lately upgraded from 60D to 6D. I managed to check out the competition for the fun, eg. Nikon D600 and noticed on Dpreview.com that it performed better on dynamic range. This wasn't the case at higher ISOs which the sample images proved, and you finally explained why it contradicted with the dynamic range score. Thanks Tony!
Tommi Puuska Happy to help! I'm constantly learning new stuff about camera mechanics by digging deep into DxOMark, too.
Your video certainly cleared up DXO's site. At first you don't really know what you're looking at. Sometimes it's easier to watch someone go through it on a video than plowing into it with no previous experience.
This information is priceless. Thanks for sharing.
Thanks!
THIS is handy video of the year!
Great Info as usual
Great video Tony, really informative. DXO mark is great when you know how to use it.
Although you did cut yourself off at the end.
great explanation- thanks it's really helped me find my way round the DXO mark site and make decisions about new the value of new lenses
So helpful! Thank you!
The problem that I see is not that DxO mark is wrong but that there is variation between individual copies of lenses.. so its hard to tell if DxO just used good copies or if you'll just get a bad one. Tony Northrup
Another thing I didn't know. Cheers
Thanks. A great and very useful video.
In Advanced search the focal range selection works kind of logically. It picks equal or higher value available for minimum focal range. It do not search if the lens is able to work in specified focal range. Overall the change they made is nice.
This video was uploaded less than 10 min ago and some close minded person dislike it already
***** They can be nasty!
Thank you for this video. So "stuffed" with useful info. Thank you for teaching me to take a deep dive into the measurement data for lenses. Thank you for teaching me that the overall score is a "blend" of parameters where the weighting is subjective - will use the underlying data going forward. And the T-stop point is should be marketed - but maybe the vendors do not want the truth to be told.
Excelent video!
Here's a question about "lens" sharpness measurements... As someone below mentioned, DXOMark, when testing a lens, tests it on a body, and that's fine if you're looking at what you can expect given the body you have. What I wonder about is that if we're evaluating a lens by itself, is there any good way to evaluate the extent to which the lens itself vs. the sensor+processor is the limiting factor in being able to resolve a given resolution? I mean, you have the fundamental issue that if you have, say, a 16 Mpix sensor, then in practice, you are probably limited to ~13 P-Mpix even with flawless optics. Yet, when an M43 lens rates 7 or 8 P-Mpix vs. an FF lens on a D810, these two results are put side by side as if the scale is absolute in spite of the fact that you don't really know whether the M43 lens is limited by the M43 sensor or FF lens is "carried" by the high-res FF sensor. I can think of a number of ways in my head to resolve this variation in a more objective way, but of course, the detailed data that is necessary to do that isn't available on the DXOMark website. Alternatively, if DXOMark would actually do tests with adapted lenses, that might give some insight.
Also, in regards to your initial point about "I got this one picture super sharp"... it does depend on the subject. Certain subjects -- say, a very fine checkerboard pattern, it isn't even possible in principle to resolve more than 4 Mpix on a 16Mpix sensor because of the RGB mosaic and so on, regardless of how good the lens is. Note that 4Mpix here refers to the checkerboard having 4-million black&white blocks visible in the image. And even that result relies on being able to align the checkerboard cells perfectly with the photosites on the sensor and so on, which is basically impossible in practice. Also, the nature of aliasing is such that even if a 4 Mpix checkerboard resolves perfectly, a 3.99 Mpix checkerboard will not because of those same aforementioned alignment issues.
By comparison, a super-sharp photo of someone's cat can look super sharp with some given lens because of the fact that this sort of extreme contrast between neighboring pixels doesn't often exist. It seems that the MTF measurements used to relate to P-Mpix measures at the limits of resolvable contrast. So even if a given MTF shows that you can only get about N P-Mpix when looking at variances over the entire dynamic range of the sensor (hypothetically, say, 12 stops for some sensor X), many real-world images might only have details with 3-4 stop variance between pixels, which means that the same MTF results that got you some N P-Mpix will still be able to resolve that comparatively "low contrast" detail at a higher resolution. So it's often not that straightforward to relate that figure to the real world results exactly.
great info. just great.
LOL @ "look I took this picture of my cat..." Great stuff as usual, Tony! Another excellent vid.
Congrats! You blow my mind! Thanks!
Hi Tony , I really appreciate the amount of time and detail you go to, no one else does this at this level, I entirely agree with your summing up of how to evaluate CAMRA and lenses using ISO sharpness and dynamic range , thanks once again to you and Chelsea and the team , dave newby
I know you're getting a lot haters, but just know that you're doing something right. Great quality videos (literally) both spoken, explained, and visually. Always room for improvement, but I know you know that and I appreciate you're insight. Better than a college professor and more content, more often!
Devon Wayne Thanks!
I use DxO software for image processing. Great software.
Amazing
I agree that DxO marks are very accurate and useful. HOWEVER, it can make you 2nd guess your camera and lens combination, and convince you that you have crap for optics. To get over that I let my lens do the talking and got sharp, crisp results. I don't have any L lenses, but I do have top of the line Sigma and Tamron. So I think DxO marks are based on consistent performance also. Whereas I can get sharp images with my lenses, just not at a consistent rate as L lenses
Thank you so much for this video, Tony! It was incredibly useful.
***** You're welcome!
in regards to focal lenght filter: it filters out all lenses which have higherfocal length than set by filter, ie olympus zuiko 75-300 won't be included in results when you'll set your filters between 100-300, but canon 200-400 will be shown in results. weird, i know :)
amazing, so much to learn from this video. Thanks a lot!
Hugely useful and opened a whole new window on photography for me. It also helped me to avoid buying a particular upgrade by revealing that was no significant optical difference between three generations of a particular lens. As a very ex-electronics engineer understanding decibels was the least of my worries!
2 years later still very comprehensive and useful video. I trust you guys and
Your assessment when it comes to photography! Thank you
Helpful. Sure wish you would return to this format vs. the one that's more fun & games, i.e. entertainment. Please read this as positive feedback and thanks for all the work you've done.
Great info - learned a lot :-)
Thanks!
Excellent video Tony.
This is a very informative channel, one of the best in youtube... but there's way too much focus on the equipment. A great photographer is the one who knows how to see the light, to compose a photography and only then who uses his camera the correct way... some types of photo styles might rely more on the hardware (like for sports) but for most types the hand who shoots the photo matters way more
Most of our videos are not about equipment... Just check our channel page. Ironically, people who think we're only about gear are the ones who choose to only watch gear videos.
Do this video explanation again during your weekly LIVE show, when everyone has had a few to drink. It would be a blast to try and understand it plastered.
Tony Northrup Hey Tony, I think a lot of people who are complaining about color depth are actually seeing color data lost due to poor dynamic range performance. Lost data is that, simply lost data. That said, I can tell the difference in Jpeg files from camera or lightroom -vs- RAW files from the camera. Even exported to 16bit Tiff there is noticeably more color range. This is not the cameras fault though, its due to Jpeg being 8bit. Note, Jpeg standard was updated to 12bit recently, but that update has not made it to Lr or our cameras at this time.
Tony,
Just started really getting into your videos, which are excellent. They are really opening a whole world to me, especially this one regarding DxOMark and how to use it to compare cameras and lens. Also, how to find a sweet spot for a given lens is very valuable.
However, I am confused about the Vignetting rating which shows a negative EV rating like -1.5 EV. This is said to mean the corners are 1.5 EV darker than the center. This seems backwards to me. A smaller EV number is gotten by using a slower shutter speed and/or a wider aperture opening. That should result in a brighter image, not darker. So I would say the corners are +1.5 EV darker. Where am I going wrong? This is not a big deal, but I am trying to get it right.
Thanks again!
PS: I am an old engineer and IT guy so I do understand math - at least I thought so!!
It has nothing to do with shutter speed. It just means 1,5 stops darker. You are not compensating anything.
Color depth is a real issue. I know this is an old video but I've ran into issues concerning color depth doing macro work and some pet photography. Now I think it really comes into play with a camera lens combo from what I found. I won't mention manufacturers as I didn't go into why it was an issue but merely changed to another brand which resolved the issue right away.
Thanks so much for posting this video. This was really useful and saved me a lot of time..
Tony, you have made a valued tool "DxOMark" more comprehensible and therefore of greater value to me and I'm sure many others. The mark of a good teacher is to make a complicated idea or concept understandable to others. Here is where you excel in communicating in ways that are tangible and clear in your TH-cam videos. You demonstrate the potential that the internet promised but so rarely delivers. I continue to learn from your generous sharing of our knowledge. Very Kind, thankyou!
I really like the DXO website, but I didn't know that they now had a better interface.
Good overview Tony - thank you.
Only two things:
1.
When comparing T-stops as in your example, 4.2 to 5.1, it is not 1 T-stop. The next T-stop to 4.2 would be 5.9. (see typical stop-row: 2.8 - 4.0 - 5.6 - 8.0). T5.1 compared to T4.2 is only about half a T-stop - even if the numbers difference is almost 1.
2.
Distortion:
It is true that distortion is auto-corrected in post process, but it should be mentioned that distortion-correction affects the sharpness in a negative way. Similar as vignetting it does not influence the center, but it does at the borders.
Do you know if the distortion-correction is already considered by DxO-mark when looking to p-mpix?
Thomas Feldbauer 1. Yes, and I wrote a title over the video to explain that mis-speak.
2. Yes, I mentioned that. I'm sure DxOMark doesn't factor in distortion correction.
Very informative...It is very nice to be able to understand DxOMark better.
Thank you Tony.
*you would think that if someone has a problem with the information that you provide us, that they wouldn't bust your balls own your channel, but rather set up their own test on their own channel and try to prove you wrong.?
Wow, the way Tony teaches is a perfect match for how I learn. I feel like I want him to keep throwing more and more at me in some sort of teacher -student challenge until I call UNCLE! I had an economics teacher in college that elicited the same response from me. He was some guy named Lester from MIT. What's fun is that I'm a highly analytical person but at the end of the day I just want to take great shot.
DXO website has gone from bad (as in pretty confusing) to worse with their recent make-over. Small wonder Tone pulls up two browser windows to compare 2 lenses. That used to be easily done within a single window on that site.
When you filter down say cameras and want to look at the best re ISO, you get the first three places in a little box right in the graph. But ... all the rest are given as per their launch year (how silly is that) and *not* with their ISO indicated but their general sensor mark, which Tony rightly says is rather superfluous.
I really wish they introduced a more sensible design at DXO, it could be such an invaluable site.
On DXOMARK the D5's low light score is even lower than the D810. However, even when viewed at the SAME SIZE, the D5's pictures are significantly cleaner. In fact, for high ISO, the D5 is a very, very good camera for that. Probably even better than the A7SII. Practically noise-free at 12800, excellent at 25600, very usable at 51200, and usable at up to 102, 400.
F-stop is based on the size of the aperture, not the front element, by definition. (If that were true, you couldn't adjust the f-stop of a lens without swapping out glass.)
doubledeej right, it's technically the iris size relative to the focal length. I was just trying to convey that f/stop describes the light coming into the lens while t-stops describe the light that exits the lens.
Tony please. This frightens the hell out me! I watched this video and then looked at DxOMark tests to see what lenses would be best for my D500. That is, to find lenses where the PMP equalled the actual MP of my camera. A sharp intake of breath and a Gerry mouse style gulp later, I find that unless I have a few grand to spare on each lens I need, I may as well use an empty Coke bottle with a bayonet mount!.. Ok, just a fun response from an English guy with a weird sense of humour..
Great stuff. When you said you were bringing it out I was thinking more post-summer!
I don't think you've gone into enough detail to convince those who dislike it for whatever reason despite saying to totally disregard the ballsed up subjective formulas but appreciate the effort gone to here to help people past the fluff and onto useful info. Those who are scared by the numbers latch on to the idea that it's garbage and those who could understand them if they looked analyse the rubbish seen in clickbait articles then on page one of the site and (quite fairly) never give it a chance beyond that.
A couple of tripped over words that will no doubt be latched on to and perhaps a couple of opinions that will be mixed in with the facts you're presenting here (I do agree with ISO holding the most value in sensor rating but disagree a little in what you're saying about print ISO being whole story when you yourself promote cropping (though I obviously see where you're coming from)). Not sure the disclaimers for photog>>>gear and you shouldn't be pixel peeping are going to make it through to enough people either but we can be optimistic.
With the video coming up to an hour I understand why you may have edited out some more detail which people will stumble upon by clicking through the measurements tab. Would be really curious to see stats for how many got through to near the end. Also value you not breaking it down into 3 separate videos etc.
I'd argue that even the ISO score loses a lot of its value when they combine it to that 'sports' rating. Would be nice to see that broken down into colour, luminance and test shots but that's asking a lot of them when you do pixel crops vs print crops too. Besides, DP review tends to have a relatively accurate way to compare these subjective things if you're really looking at a new body.
Also along with that focus breathing vs effective field of view it's also worth mentioning that even on some high priced primes that distortion often adjusts as you move through the focal range. Would be nice to see an extra few measurements on that field map tab to click through. (Though I admit it's not really fair to be asking anything of them considering what people pay to use it - nada)
I can't see them ever moving away from things like their overall score because that's what the media use in their sensationalist clickbaity articles as you said, which drives more people to DXO mark (even if only long enough to look at the front page and rightfully decide those first few scores are pretty garbage)
Anyways, good job and much respect for spending the time on a video that won't generate nearly as many clicks as some of the others.
James R Thanks, James. For the record, my goal isn't to convince people to trust DxOMark if they don't now--there's definitely no way I'm going to change those peoples' minds. I just want to help everyone else understand it, because their website is hella confusing.
Tony Northrup we are in agremant on the confusing bit ...you did a good job of exsplaning it
This is the single most useful camera *tech* video on the web. Spot on. Explains perfectly what is unique and useful (a lot) and not useful (also a lot) on DxO.
FWIW, you can see that Leica doesn't lie about their ISO performance like the others. Leica T at 3200 indicated is 4100 ISO while the D7200 is at 2300. I've never seen a camera reviewer bring up the "this one goes to eleven" problems. You can't find that info anywhere else.
(Suggest use 'Select' and "Compare camera"/"Compare lens" to overlay 3 of each on same chart. New menu requires another step, 'View comparison selection')
Screen matters for video where you don't downscale your 1080p videos, however it will be more useful when you shoot 4k and then downscale to 1080p.
smurfpiss This is true--unfortunately they don't test video quality.
I like the focus breathing idea. Great vid. Thanks.
Most helpful video I've seen in years! Before seeing it I was always frustrated when trying to navigate through DXOMARK. I don't understand why they don't they improve their website.
I don't have a problem with what DXOMark measures. My issue is that there are things that they don't measure and may not be measurable, like micro-contrast and color rendition. Looking at DXOMark you would think that a Sigma Art lens was just as good as (or better than) a Zeiss lens, but the Sigma lenses lack that Zeiss pop and color rendition.
Tony, just a few comments while I watch your video ...
The DxO color depth IS useful. It is DxO's way to measure color noise. All other metrics are black & white only. W/o this metric, vendors could have very wide color Bayer filters which produce a lot of color noise and possible bad CRI and rock DxOmark. DxO prevents this by including the color depth metrics.
Why does that require them to show it on the front page?
Attaching cameras to the same body??WTF?...Great video by the way!
miguelao1 Attaching the same lens to different bodies is what he meant.
***** Yeah, I'm bound to say the wrong word somewhere in an hour-long technical video :)
Tony Northrup LOL, an innocent slip-up. However, got me thinking ... maybe you could with a macro reversing ring.
Tony Northrup lol...1 mistake in an hour-long video is quite acceptable !!! Keep up the good work!
Omesh Singh Camera A+light_source+Camera B
Maybe not that crazy ,after all ...if you want to measure ISO with exactly the same conditions and time stamp!
I really wish DxOMark had more lenses to compare as well as more lens/body tests. They have a really incomplete list. I know it isn't easy to test as many lenses as they have but there was an website that I can't seem to access outside of waybackmachine called photonotes that had way more lenses and comparisons you could actually look at yourself. The comparison page doesn't work on waybackmachine but it was a really awesome website that seems to have died.
theduckfu actually the site photonotes is just down because of internet issues. It should be back later check it out they have a ton of EOS samples to compare lenses. It is incredibly exhaustive.
Also I prefer looking at actual samples at different apertures then a 9 or 10 sharpness rating. I don't know how the 9 or 10 is quantified and how the graphs actually look on performance. I do know a lens can very greatly being wide open vs being stopped down a few pegs.
Random trivia:
85mm f/1.4G on D800: 22 P-MPix
85mm f/1.4G on D800E: 30 P-MPix
Camera von Theuns Yeah, that anti-aliasing filter makes a big difference! I did a bunch of research on this for this: th-cam.com/video/Ww6QGpryrLM/w-d-xo.html
I've watched many of the Northrup videos and I believe this one to be the most educational, overall, than any other video. It may be focused on dxomark but in explaining the site Tony teaches about many important aspects of photography. Right on!
I was debating about upgrading my Tamron 150-600 G1 to a G2. The data from Dx0Mark and how to interrupt from this video has helped me decide it is not worth it. And actually, if sharpness is the only consideration, G1 is better.
I think this is a great presentation, love the clarified detail and recommendations, a great service to photographers and DXO.
ps
What do you think of their DXO software?
Thanks this was helpful.
Thanks Tony - great video. I always wonder how far we should take these scores.
Taking say the 6D as an example, the highest sharpness lens with that body is rated at 17 which I guess is broadly expected for a 20 MP camera. Does that mean that there is no point mounting say a 50mm 1.4 over a 50mm 1.8 from a pure sharpness perspective? I guess the argument is academic for the 6d because the speed difference is enough to justify changes but if you take the same logic and apply it to the A7. It being a 24 MP camera would indicate its unlikely to score much above 24 so if there any point in buying the expensive 55mm 1.8 which is a much sharper lens over the 50mm 1.8. According to DXOMark the 50mm is already going to give a resolution higher than what the camera body itself is likely to support (26 when mounted on the A7R II). Are A7 users better off buying the 50mm and saving money? Obviously if you have an A7R its a different question.
T&C... Just let the haters hate. Don’t give them energy by reacting to them. I think it’s really cool that you do stuff like this. :)
BTW. Still waiting for that EM-1 video. 😂✌🏼😇
I did not see the Opteka 500mm / 1000mm f/6.3 Telephoto Mirror Lens for Nikon shown on dxomark.
Tony, could you please review this lens
Thank you.
DXOMark also doesn't include the capacity of the sensor to produce details for a given lens, if you take for example the D5300 with nikon 35mm 1.8G DX the Perceptual MPix are 13 MPX but only 12 MP mounted on the D7100 (The D5300 Sensor can resolve more details this should be included on the final score), the gap should be more obvious on camera like Samsung NX1 with their 28Mpix BSI Sensor a lens like the Samsung 50-150mm f/2.8 S ED OIS (Not Tested Yet on DXO Mark How convenient) can produce a lot details coupled the the NX1 Sensor
www.imaging-resource.com/PRODS/samsung-nx1/FULLRES/Y_SAM1533.HTM
you can see that even at 1:1 view judging at pixel level, there is a lot of details proving that samsung can produce lenses that can have more than 24 Perceptual-Mpix, thus using the sensor full potential .......
anyway a very informative video as always i learned a lot ..... keep up the good work
Tony Northrup I have an honest question regarding sharpness using a full frame lens on a DX sensor. The DX sensor takes the image from the center portion, theoretically sharpest portion of the lens. Why then does a Nikon 610 (24mp fx) have a higher perceptual megapixel count that an Nikon 7100 (24mp dx camera) when using a Nikon 85mm 1.8G.
Shouldn't theoretically the Nikon 7100 have a higher perceptual MP (sharpness) count than the 610 fx camera?
Thank you Tony!!! The DxOMark website SUCKS(I mean suck at browsing it:-) ! They might have great data, but I cannot even use or make sense of it....I'm all ears as I start this video!
i still do not understand why 7d mark II has much lower score than 70D with the same lens. I thought they have the same sensor. Is 70D sensor better than 7d mark II?
At 12:24 you state “DXOMark is really smart at how they score this (a very untrue statement)…..there is a good reason why (the D810 score is twice that of the D5500, which is also the same sensor in the D7200 by the way) the D810 scores twice as high, since the sensor is about twice as big ….For any given ISO it is gathering twice as much light (not true point 2)…and its that total light that makes the biggest difference in your image quality (not true point 3)” ISO in digital is a time sampling RGB signal gain and has a native limit at which point signal amplitude is applied.
No lens on earth “cares” or “knows” what sensor is underneath it, the angle of interception of the circle of rear element of projection (Field of View) changes, but this has no effect on luminal density (LUX) per square mm of sensor. Photosite GAIN (less gain, and higher noise on DX photosites, which are smaller) has been mitigated by SNR time-elimination algorithms. Radio astronomy (light is EM, radio is EM, EM is EM, regardless) has used this software ‘trick’ now for many years, its use is now standard in interferometry. As for “image quality”, this confuses sensor level SNR (noise and gain) without applying SNR compression althorithms used by Nikon and others which have allowed for low light high ISO performance on DX sensors such as the D7200. Image quality is also relational to pixel density.
The DX sensor, per square MM, while having less gain, has more translational data, which is why nature shooters are using 800mm lenses (for example) on DX sensor cameras. Likewise the lack of smaller DX photosite gain has been mitigated by advanced time-compressional SNR algorithms in the digital processors AFTER the sensor and AD converters but before the DSLR buffer. To speak of “quality” is obtuse, and leaves out 1. Rendition 2. Pixel density, 3. SNR gain algorithms, and other criterion. This is why the next generation of all digital cameras are FF sensors with DX pixel densities (ala the new Canon 50MP sensor camera,…which happens to be a 62MP sensor, but down-sampled to 50MP to allow for faster buffer read-writes).
This incorrect conclusion assume two things which are wrong: (sensor) size is relational to luminal density, however no sensor is gathering more light, rather has (or can have) higher GAIN due to photosite size as present on FX sensors. Digital camera sensors contain ‘microlenses’ above and around each photosite to enhance their light-gathering ability. These lenses are analogous to funnels which direct luminal capacitance into the photosite where the light would have otherwise been unused.
However as is the case, larger photosites with greater gain have a downside (again, the same thing was ‘discovered’ in radio astronomy using large dishes), which is larger photosites gather more unwanted peripheral luminal noise not desired to be directly captured.
As such, the statement that “DXOMark is really smart” is misleading along a wide spectrum of empirical facts relational to FF and DX sensors due to GAIN, SNR algorithms, translational data per square mm (pixel density) and luminal density per square mm which is unchanging between FF and DX sensors. Likewise, again, no lens projects more or less light due to FF or DX sensor being behind same. Angles of interception have no relational inverse-causation on luminal densities.
Pat Fitzpatrick Heya, Pat. your understandings are common, and I thought the same things a year ago before I started doing all this research.
Here's some info about noise and sensor size, with real-world data and examples to support it: th-cam.com/video/0OtIiwbAZi8/w-d-xo.html
If you want more, here ya go: th-cam.com/video/f5zN6NVx-hY/w-d-xo.html and th-cam.com/video/DtDotqLx6nA/w-d-xo.html and th-cam.com/video/6Im4W_9blhY/w-d-xo.html
Regarding using smaller sensors for wildlife photography, here's some good info: th-cam.com/video/YDbUIfB5YUc/w-d-xo.html
Regarding pixel density, here ya go: th-cam.com/video/_KYvp8PrCFc/w-d-xo.html
If you still disagree with something, I just ask that you create real physical evidence by performing experiments, and send over the images that prove your point.
Another fantastic video, Tony. Thanks so much. Incredibly helpful.
I also hope that DxOMark figures out a standard to measure auto-focus capabilities of different camera mounting different lenses. I can't be the only guy who consider auto-focus accuracy more than perceivable megapixel, right?
+Aaron Yang Agreed--megapixels don't matter at all when your picture is out of focus.
You need to update this video. I went to DXOMark and their interface is completely different now. It seems to have been dumbed down quite a bit. The access to data is very limited. I could not access any information about third-party lenses. Their search filter had only a handful of brands, and none of them were third party ie Sigma, Tamron, etc.
I respect DXO and there figures, but like Tony said, it can only be used as a small part of your buying decision, if they say one sensor is better than in another camera, you still need to way the other features of the camera, and the price, and weight and focusing speed, and many other things in your buying decision, and same thing with lenses, maybe one lens wins buy one point in a DXO test, but the other lens has IS, or one lens is weather sealed, or focuses faster, or one lens is allot lighter, or theres a big price difference, there are so many variables in a buying decision other than the DXO tests, also, personally I always try to test a lens before I buy it, using my sd card at the camera store, and then I go home and look at the images on my computer and try to come to a conclusion, about sharpness, contrast, and focusing consistency, and also the quality of the bokeh, and if i like the lens, i go back and buy the exact one I tested, when I bought my 85 Sigma f1.4, i had tried 3 of them, and 3 Canon 85 f1.2 lenses, at three different stores, and the 3rd Sigma was the charm, it does everything great, and it was almost one third the price of the Canon at the time, and I came to the conclusion that it was better than the Canon f1.2, at any price because it is lighter, faster focusing, and just as sharp, but the first two Sigma 85's were not as consistent with the focusing accuracy, and the first two Canon 85's did allot more hunting than the 3rd one I tried, so all lenses of the same kind aren't equal, witch makes me wonder how many copies DXO mark test, anyway now that I just looked at DXO marks sharpness chart, the Sigma 85 actually slightly out scored the Canon L with sharpness, witch made me feel good about my decision, even though DXO wasn't part of my decision buying it, but so many people will buy an L lens just because of the red ring, and also the Canon 85 L is actually f1.4 T stops, but I don't think the difference between f1.2 and f1.4 matters anyway, personally i don't want to shoot shallower than f1.4, and the best sharpness is maybe from f4 to f8 anyway, but it all depends on the look you are going for, but overall I agree with Tony that DXO is a cool site
Woow, this video was super awesome and opens my eyes to choose the best lens. Thank you so much Tony.
Hello, great work from you, like always... I use a lot their website and the new update it was quite a shock for me :). But... I'm very disappointed also... they updated the website but they didn't added new tests... I mean there are some lenses that are already on the market for years and they didn't tested them and probably won't ever. The 12-24 from Nikon it doesn't even exist on their website... Maybe the fact that they have the only website of this kind, puts them in position to get some money from brand to test their product... so maybe they have some kind of priority list... Anyway I'm disappointed of them...
By the way... if you choose Lenses->Tokina->Any price->Any date... you get no results... For me, the old filter system and list, was more versatile and user friendly
Great video :) I agreed with all your thoughts :)
mwsi7128 Thanks!
thanx for the video . now im trying out the free trial.. but is a possible way to make a video on how to fix some photos like outdoors.. i will like to learn how to use it very good
Muito obrigado pelos ensinamentos! Estou estudando a aquisição de uma nova camera e lentes e aprecio muito os seus vídeos, pois descrevem a ciência e os aspectos práticos de diversas combinações de equipamentos fotográficos. Utilíssimos.
Anten