Pages: 1 2 [3]   Go Down

Author Topic: DxO bashing - a bit of perspective  (Read 14331 times)

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: DxO bashing - a bit of perspective
« Reply #40 on: June 16, 2014, 05:15:20 am »

Hi,

For the audio comparison, there is a lot of hype in that area.

Regarding the DxO tests, they present a lot of info, but all of that info relates to noise. DxO mark essentially ignores resolution, for instance. I am pretty sure that the measurements are accurate.

Now, I would say noise is an important characteristic of cameras, but other factors may be more important, especially if noise related data is good enough. Lenses, AF and handling may have a significant role.

Best regards
Erik


Hi Erik,
Noise (of various types) has always been a huge nuisance in both photographic and audio reproduction. It's been one of the major factors that contributes either directly or indirectly to a clear sense of the difference between a live performance of music (without amplification and loudspeakers, such as a classical concert) and a recording of the same music.

The ideal situation for me, when listening to music through a well-adjusted hi fi system, is to be able to close my eyes and imagine, without distraction, that I have the best seat in the Concert Hall or Opera House.

It's misleading to claim that the DXO tests ignore resolution. Sensors have no resolution in themselves. The megapixel count describes the potential resolution of the sensor. Surely one doesn't need to be told that a 36mp sensor has a greater potential resolution than a 16mp sensor!

The actual resolution will always be determined in conjunction with a specific lens, and this is what DXOMark also test, producing the new metric 'Perceptual Megapixel' or P-Mpix.

Now one might sometimes disagree with the quantitative significance of those P-Mix results, but one can't sensibly disagree that the resolution of a recorded image is always a product of both sensor and lens. If one increases the potential resolution of either one of them, the resolution of the recorded result will also increase.
Logged

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: DxO bashing - a bit of perspective
« Reply #41 on: June 16, 2014, 05:37:35 am »

It's almost a cliche to mention it, but the lessons learned from the good (bad) old days of high-end audio are salutary.

Stereo Review magazine did in depth technical reviews of amps and such down to four decimal points. Their reviews were so influential that manufacturers actually designed some products to test well, with 0.005 THD. The problem was that there was little if any correlation between test performance and how an amp sounded.

The Absolute Sound magazine took the other approach, eschewing lab tests and just relying on the judgement of experienced listeners. In this case most observers noted that this was much more closely correlated with the user's listening experience that the lab tests. The take-away was that people listen to music not numbers. The reasons are now well understood, though they wern't at the time.

Michael
 

Strangely enough, people also choose which classical concert they go to by personal sampling and by word of mouth or reviews, not numbers :)

Edmund
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: DxO bashing - a bit of perspective
« Reply #42 on: June 16, 2014, 09:27:40 am »

More likely they feel they do not need to use numerical lab results. Most top chefs probably dont feel the need to use timers and temperature probes. They see, feel, taste.

The science is important, nobody credible would say we would have the current level of tools without the science. DxO tries to use a repeatable scientific methodology for comparisons. The deciding factor for everyone is judgement. Some people have it based on a variety of things. Some people lack judgement in a variety of things.

Hi,
This analogy doesn't quite make sense. If you wish to compare photography with cooking, I think the camera, lenses, other photographic equipment and editing programs would be analogous to ovens, grills, choppers, blenders and mixers etc. The raw, unprocessed food would be analogous to the  RAW files of the scenes or subjects that have been photographed, and the preparation, mixing, addition of various spices, and the cooking, would be analogous to the processing of the RAW images in Lightroom and/or Photoshop.

Quote
IMO DxO numbers are better at telling you what is good equipment than what is bad equipment.

Don't understand this. If a DR of 14EV is good, then by the same reasoning a DR of 7EV is bad.
Logged

Telecaster

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3686
Re: DxO bashing - a bit of perspective
« Reply #43 on: June 16, 2014, 03:17:16 pm »

Don't understand this. If a DR of 14EV is good, then by the same reasoning a DR of 7EV is bad.

Good and Bad are the wrong terms IMO. Too absolutist. More DR is More, less is Less. There are plenty of photographic situations where seven stops of DR is sufficient. Or maybe my photographic style features high contrast with crushed shadows and/or blown highlights. In such cases having 14 stops available is a luxury—great if I like starting with spatial/tonal-rich source data and then decimating it in post—but not a necessity.

When I look at a DxO evaluation it's to see whether or not a particular piece of gear will do what I want it to, or maybe if it has extra capabilities I might want to take advantage of. This by itself doesn't tell me whether or not I'll actually like the gear in use. To find this out I've gotta put it through its paces.

Also, sometimes gear will behave in interesting ways when pushed beyond its intended limits. Guitar amp distortion, anyone? This happens with electronic cameras too. I like the tonal distortion my Leica M8.2s produce when overexposed. I suspect this causes the sensor to record enough infrared to affect chroma response, as in the attached photo (which was taken two days ago, not in autumn...those leaves, in the visible spectrum, are intensely green).

-Dave-
« Last Edit: June 16, 2014, 03:27:05 pm by Telecaster »
Logged

Fine_Art

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1172
Re: DxO bashing - a bit of perspective
« Reply #44 on: June 16, 2014, 05:02:12 pm »

Hi,
This analogy doesn't quite make sense. If you wish to compare photography with cooking, I think the camera, lenses, other photographic equipment and editing programs would be analogous to ovens, grills, choppers, blenders and mixers etc. The raw, unprocessed food would be analogous to the  RAW files of the scenes or subjects that have been photographed, and the preparation, mixing, addition of various spices, and the cooking, would be analogous to the processing of the RAW images in Lightroom and/or Photoshop.

Don't understand this. If a DR of 14EV is good, then by the same reasoning a DR of 7EV is bad.

To me the light entering the lens would be raw food. The lens is processing, like cutting, peeling, cooking. The sensor is arranging for plating, the software is spices. I agree the analogy is not great.

AFAIK DxO does its measurements on one copy they get. That is not scientific. I know for a fact lenses they say are poor resolve well to 14MP APS-C, they also have good OOF rendering characteristics. Therefore I think DxO may get a bum copy of a lens. If they rate something well it is unlikely the copy is far above other samples. If they rate something badly they may have a defective product.
Logged

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: DxO bashing - a bit of perspective
« Reply #45 on: June 16, 2014, 07:21:41 pm »

To me the light entering the lens would be raw food. The lens is processing, like cutting, peeling, cooking. The sensor is arranging for plating, the software is spices. I agree the analogy is not great.

AFAIK DxO does its measurements on one copy they get. That is not scientific. I know for a fact lenses they say are poor resolve well to 14MP APS-C, they also have good OOF rendering characteristics. Therefore I think DxO may get a bum copy of a lens. If they rate something well it is unlikely the copy is far above other samples. If they rate something badly they may have a defective product.

My guess is production line pass/fail diagnosis is now probably DxO's main line of business, except it concerns camera phones. I am not sure that Nikon, Canon and co. enjoy the idea of end users being able to test lens samples in the same way.

The phone industry is probably now the driving factor behind the ISO standard test procedures and sensor development.

I participated in a camera phone quality meeting at their headquarters, and feel reluctant to bash them too hard because they are earnest people who genuinely care about image quality, and their catering and coffee is superb. Their developers get unlimited free Nespresso Pro!

Edmund
« Last Edit: June 16, 2014, 07:28:42 pm by eronald »
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: DxO bashing - a bit of perspective
« Reply #46 on: June 16, 2014, 10:12:09 pm »

Good and Bad are the wrong terms IMO. Too absolutist. More DR is More, less is Less. There are plenty of photographic situations where seven stops of DR is sufficient. Or maybe my photographic style features high contrast with crushed shadows and/or blown highlights. In such cases having 14 stops available is a luxury—great if I like starting with spatial/tonal-rich source data and then decimating it in post—but not a necessity.

When I look at a DxO evaluation it's to see whether or not a particular piece of gear will do what I want it to, or maybe if it has extra capabilities I might want to take advantage of. This by itself doesn't tell me whether or not I'll actually like the gear in use. To find this out I've gotta put it through its paces.

Also, sometimes gear will behave in interesting ways when pushed beyond its intended limits. Guitar amp distortion, anyone? This happens with electronic cameras too. I like the tonal distortion my Leica M8.2s produce when overexposed. I suspect this causes the sensor to record enough infrared to affect chroma response, as in the attached photo (which was taken two days ago, not in autumn...those leaves, in the visible spectrum, are intensely green).

-Dave-

Of course, Dave. I agree. This is why numbers are so much more useful than ordinary words when attempting to describe such characteristics in an objective manner. If the DR of a P&S camera is found to be 7EV at ISO 80, and the methodology is sound, and the term DR is well defined, then it's an indisputable fact that the DR is 7EV. Whether that level of performance is good or bad, or good enough, or not too bad, is always a matter of opinion and can be the subject of endless debate.

Without the benefit of DXO test results, or similarly objective results, one could waste an awful lot of time and money testing cameras for oneself. I'll always remember the exaggerated opinions expressed about the performance of the Nikon D3 when it was first released, at a time before DXO had made its test results available. The hype was ridiculous. Because the D3 had an unprecedented high ISO setting of 25,600, it was assumed that the high-ISO performance of the camera was streets ahead of any other camera on the market. I recall claims of up to 2 stops better high-ISO performance than any other DSLR then available.

I was using a Canon 5D at the time and took the trouble to visit a Nikon store to do a few test comparisons. Underexposing the Canon at ISO 3200 to simulate the higher ISO settings up to 25,600, my results indicated that between ISO 1600 and ISO 25,600, the D3 had no more than 0.5EV better DR. Now that's worth something, and is a noticeable improvement, but is not particularly significant, and a long way from the hype of 2 stops improvement.

When I first reported these findings on this forum, the debate and disputes continued because DXO had still not made its test results available. A couple of years later the DXOMark site appeared, showing comparisons between the Canon 5D and the Nikon D3. I admit I was surprised to discover how closely my own results, using relatively crude methods, matched those of DXO.

Making use of the performance limitations of equipment to create an unusual artistic effect might have some merit. The banding that's very noticeable in the deep shadows of shots from the 5D can be used to create a tapestry effect, by severely underexposing the entire scene.  ;D
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: DxO bashing - a bit of perspective
« Reply #47 on: June 16, 2014, 11:49:40 pm »

Hi,

In general, I feel that test are far more reliable than user experience. The main reason is controlled conditions and the possibilty to go back and redo if in doubt. My own experience is that the more factors are equalised the more the equipment performs as expected. Say, shooting a white balance card and use it to for white balance makes colours much more similar, shoot a color checker and equalize black and white fields, global contrast will be similar. Create a DNG colour profile for both cameras, even less difference will be visible.

Regarding DxO, I feel their data is usable. For me it is mostly about the lens tests, as I am essentially happy with my cameras regarding noise characteristics. DxO says that it takes five points on DxO mark for a visible difference, so I can see if a new generation of sensor gives a significant improvement or not by a glance.

Regarding the DR stuff, I feel that it is not the more important stuff in life, and it is quite hard to measure in real life, as it is often reduced by lens flare. In practical photography I see that there is a significant advantage in reproduction of darks between my late generation Sony cameras and my P45+ back that is using a Kodak sensor developed around 2005. This corresponds to DxO data.

Getting back to lenses, I look at the curves. I feel it is a good info, and it is useful to decide if a lens is an useful upgrade. For instance, I have Sony 24-70/2.8 lens that has weak corners at 24 mm and drops of axis resolution at 70 mm. The new Sigma 24-105/4 "Art" semt to be an attractive alternative, but looking at DxO I found out that it is probably not better than the Sony lens.

DxO makes their measurements to be able to support sensor/lens combinations with their quite reespected raw converter. So their results go into their raw converter, which does quite a few lens corrections.

Personally, I much enjoy Roger Ciala's writing over at LensRentals. They test tens of thousands lenses each year anf thus also see statistical variations. I also like Rogers explorative mind, insights on optics and scientific/statistical approach. LensRentals has probably the most extensive lens testing facility outside the lens manufacturing business and they are glad to share their findings.

A problem with formal tests is that they are focused on in focus sharpness, but most images will have significant areas more or less out of focus, as plane of absolute focus is very thin. So in many cases, out of focus rendition may be even more important than rendition in absolute focus. This is an area there many new lenses can shine, with circular aperture for instance, while many designs show weakness with longitudional chromatic aberration and double contours caused by over corrected spherical aberration. Photozone covers the bokeh stuff pretty well in their tests.

Best regards
Erik


Of course, Dave. I agree. This is why numbers are so much more useful than ordinary words when attempting to describe such characteristics in an objective manner. If the DR of a P&S camera is found to be 7EV at ISO 80, and the methodology is sound, and the term DR is well defined, then it's an indisputable fact that the DR is 7EV. Whether that level of performance is good or bad, or good enough, or not too bad, is always a matter of opinion and can be the subject of endless debate.

Without the benefit of DXO test results, or similarly objective results, one could waste an awful lot of time and money testing cameras for oneself. I'll always remember the exaggerated opinions expressed about the performance of the Nikon D3 when it was first released, at a time before DXO had made its test results available. The hype was ridiculous. Because the D3 had an unprecedented high ISO setting of 25,600, it was assumed that the high-ISO performance of the camera was streets ahead of any other camera on the market. I recall claims of up to 2 stops better high-ISO performance than any other DSLR then available.

I was using a Canon 5D at the time and took the trouble to visit a Nikon store to do a few test comparisons. Underexposing the Canon at ISO 3200 to simulate the higher ISO settings up to 25,600, my results indicated that between ISO 1600 and ISO 25,600, the D3 had no more than 0.5EV better DR. Now that's worth something, and is a noticeable improvement, but is not particularly significant, and a long way from the hype of 2 stops improvement.

When I first reported these findings on this forum, the debate and disputes continued because DXO had still not made its test results available. A couple of years later the DXOMark site appeared, showing comparisons between the Canon 5D and the Nikon D3. I admit I was surprised to discover how closely my own results, using relatively crude methods, matched those of DXO.

Making use of the performance limitations of equipment to create an unusual artistic effect might have some merit. The banding that's very noticeable in the deep shadows of shots from the 5D can be used to create a tapestry effect, by severely underexposing the entire scene.  ;D

Logged
Erik Kaffehr
 

Fine_Art

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1172
Re: DxO bashing - a bit of perspective
« Reply #48 on: June 17, 2014, 12:34:07 am »

DR is significant. I pay close attention to how the shaded undersides of trees look in bright sun. You can see detail easily that may be hard to capture on camera.

Try this simple test: Walk with your colorchecker passport from a well lit room to a very dark room at night. Can you still see the colors? Yes. Can you see the small shifts of warmer/ cooler white tone? Yes, very easily. If all this detail vanishes in your images the camera, or maybe the output medium, needs better tonality in darks.
Logged

hjulenissen

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2051
Re: DxO bashing - a bit of perspective
« Reply #49 on: June 17, 2014, 03:05:42 am »

AFAIK DxO does its measurements on one copy they get. That is not scientific.
If they are being accurate about _what_ they are doing, I am reluctant to claim that they are "not scientific" based on them reviewing only one sample. You would have a much stronger case for calling them "not scientific" on the basis that their reports are not peer-reviewed and published in a reputable journal. But then, how much scientific literature is being discussed here?

I agree that it is valid to question how representative a single sample of a lens is. It seems that the measurement that cause most irritation on discussion forums, though, is the camera DR measurement (especially among Canon owners). Do you also think that there is sample variation in interchangeable-lens cameras?

-h
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: DxO bashing - a bit of perspective
« Reply #50 on: June 17, 2014, 03:33:41 am »

I agree that it is valid to question how representative a single sample of a lens is.

While true, it still is a sample of the larger population (like the one you might buy). Besides, I believe that DxO have in the past have updated their database, when it became obvious (the benefit of publishing it for comparison to other tests) that they had tested an outlier.

Quote
It seems that the measurement that cause most irritation on discussion forums, though, is the camera DR measurement (especially among Canon owners).

It is what it is, and it's almost exactly the same as what I measure on the models that I tested.

Quote
Do you also think that there is sample variation in interchangeable-lens cameras?

There is always sample variation, but the real question is how large is it, and where do we draw the line of calling it still acceptable (as it will drive cost, up or down). Then, should we be inclined to purchase it, we can still try and find a copy that scores relatively high, by conducting proper tests if one knows how. Of two copies, there is most likely one better than the other (maybe the DxO copy), but maybe not on all aspects, so we also need to make personal trade-offs.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

Fine_Art

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1172
Re: DxO bashing - a bit of perspective
« Reply #51 on: June 17, 2014, 03:44:06 am »

No, I don't think DR varies like lenses. The quantity of lens variability is shown by the need for focus adjust on high end cameras. What governs DR? Full well capacity? That would be based on the volume of silicon, the purity of the chip? Probably not a lot of variability. Manufacturer process quality differences yes.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: DxO bashing - a bit of perspective
« Reply #52 on: June 18, 2014, 02:00:00 pm »

Hi,

Full well capacity and readout noise are the major factors.

Best regards
Erik


No, I don't think DR varies like lenses. The quantity of lens variability is shown by the need for focus adjust on high end cameras. What governs DR? Full well capacity? That would be based on the volume of silicon, the purity of the chip? Probably not a lot of variability. Manufacturer process quality differences yes.
Logged
Erik Kaffehr
 

Hulyss

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 734
    • H.Bowman
Re: DxO bashing - a bit of perspective
« Reply #53 on: June 19, 2014, 07:14:32 am »

For me, there is a huge gap between scoring or charting cameras and photography in general.

Scoring and charting a camera (or lens) is scientific purpose. I agree that the photographer need to know his gear specificity but hell ... that will not make him a better photographer, at all.

DxO imho is aiming at a specific new age species of photoconsumers who stay more time on internet than shooting. The capacity of a camera is subjective as bokeh, in a certain limit. Old cameras still run properly, old mfdb backs still run properly.

So DxO is a sort of marketing tool, after all. They even might have some brand preferences so they are not that impartial.

Gear technology is a deviance in photography. The photographer have so much infos coming trough his head that he is always sort of stressed.

"Have I made the good choice ??" ... "So my camera is not that good after all ??" ... "should I return it and jump on this another brand ??" ... "DxO is saying that my lens is crap but my photos are selling well; what should I do ??" ...

In this world, a lot of ppl do not feel secure; the world want you to feel insecure. Brands try to make you feel secure by many sneaky and clever ways (some ppl are paid for that). Those kind of sites just put even more doubts in their heads and it result in a form of stress and frustration, endless debate on whatever is good or not.

Meanwhile you may notice that the most rewarded photographers on this planet do not give a flying buck at all of this, unless they are paid to do so.

For me, Michael reviews are completely sufficient because it come from an experimented photographer. Photographers who have a load of flight hours also review things with wisdom.

DxO just look like a freakin bot compared to real and uninterested photographer review.

So, at the end, it is amazing how internet and ppl who know his power can remote a whole population of humans. Even if the world is getting more and more comfortable in developed country, this intra-species behaviour is sort of alarming.  You have the shepherds and the goats.

P.S: the not so funny fact is that some humans like being goats and know they are goats. So that lead to a big question: Are humans prone to obedience and submission, after all ?
« Last Edit: June 19, 2014, 07:27:56 am by Hulyss »
Logged
Kind Regards -  Hulyss Bowman | hulyssbowman.com |

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: DxO bashing - a bit of perspective
« Reply #54 on: June 21, 2014, 10:52:39 am »

Scoring and charting a camera (or lens) is scientific purpose. I agree that the photographer need to know his gear specificity but hell ... that will not make him a better photographer, at all.

This notion doesn't make much sense to me. You agree on the one hand that the photographer needs to know the specifications of his gear, yet claim on the other hand that knowing such specifications will not make him a better photographer at all.

One of these statements must be false.

Quote
Gear technology is a deviance in photography. The photographer have so much infos coming trough his head that he is always sort of stressed.

"Have I made the good choice ??" ... "So my camera is not that good after all ??" ... "should I return it and jump on this another brand ??" ... "DxO is saying that my lens is crap but my photos are selling well; what should I do ??" ...

On the contrary! DXO doesn't use such terms as 'crap'. They use numbers. Once the photographer understands the significance of the numbers, he is more able to make an informed choice and free himself from all stress regarding the performance of his equipment, and whether or not he has made a good choice. If a new camera appears on the market which, according to DXO, has higher specifications with regard to DR, for example, than the camera one is currently using, and one knows from experience that the DR of one's current camera is always acceptable, or at least very rarely unacceptable, then one should able to resist any hype and sales pitch from any source praising the new camera for its excellent DR capabilities.

Furthermore, if one is actually dissatisfied with the DR of one's current camera, or any other specification which DXO test, one has the opportunity to check for oneself, from the DXOMark comparisons, whether or not the new camera has a worthwhile increase in DR, or other specification, which would justify the purchase of a new camera. Worthwhile improvements might be Color Sensitivity greater than 1 bit higher; DR greater than 0.5EV higher, and SNR greater than 6dB higher.
Logged

Hulyss

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 734
    • H.Bowman
Re: DxO bashing - a bit of perspective
« Reply #55 on: June 21, 2014, 12:18:32 pm »

This notion doesn't make much sense to me. You agree on the one hand that the photographer needs to know the specifications of his gear, yet claim on the other hand that knowing such specifications will not make him a better photographer at all.

One of these statements must be false.

On the contrary! DXO doesn't use such terms as 'crap'. They use numbers. Once the photographer understands the significance of the numbers, he is more able to make an informed choice and free himself from all stress regarding the performance of his equipment, and whether or not he has made a good choice. If a new camera appears on the market which, according to DXO, has higher specifications with regard to DR, for example, than the camera one is currently using, and one knows from experience that the DR of one's current camera is always acceptable, or at least very rarely unacceptable, then one should able to resist any hype and sales pitch from any source praising the new camera for its excellent DR capabilities.

Furthermore, if one is actually dissatisfied with the DR of one's current camera, or any other specification which DXO test, one has the opportunity to check for oneself, from the DXOMark comparisons, whether or not the new camera has a worthwhile increase in DR, or other specification, which would justify the purchase of a new camera. Worthwhile improvements might be Color Sensitivity greater than 1 bit higher; DR greater than 0.5EV higher, and SNR greater than 6dB higher.


There is no false statement.

All FF cameras since 2008 have enough DR for all major professional work (weeding, products, landscape...). The thing is to understand light and lightning and reading the camera's user manual. Knowing the specification of his gear mean knowing how to use it at the maximum, and knowing very well how to PP specifically behind. If consumers was focussing on those two very points, trust me, ppl would not buy that much gear. Market just want us to buy more, that the only reality. The reality is also that gear become better and better even in other sectors of this market. I finished a weeding one half hour ago and I did it with the Fuji XT-1 only and some good lenses. The D700 and 800 stayed in the car and I KNOW that this little inexpensive camera did the job extremely well.

I think that photographers, if they focus enough, can output outstanding results out of "obsolete" cameras. The camera is just a dead tool. the photographer's skills are the most important. It is why your numbers tell me nothing. Now I have a check to bring to my bank ;)
Logged
Kind Regards -  Hulyss Bowman | hulyssbowman.com |

barryfitzgerald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 688
Re: DxO bashing - a bit of perspective
« Reply #56 on: June 22, 2014, 09:01:15 am »

There is no false statement.

All FF cameras since 2008 have enough DR for all major professional work (weeding, products, landscape...). The thing is to understand light and lightning and reading the camera's user manual. Knowing the specification of his gear mean knowing how to use it at the maximum, and knowing very well how to PP specifically behind. If consumers was focussing on those two very points, trust me, ppl would not buy that much gear. Market just want us to buy more, that the only reality. The reality is also that gear become better and better even in other sectors of this market. I finished a weeding one half hour ago and I did it with the Fuji XT-1 only and some good lenses. The D700 and 800 stayed in the car and I KNOW that this little inexpensive camera did the job extremely well.

I think that photographers, if they focus enough, can output outstanding results out of "obsolete" cameras. The camera is just a dead tool. the photographer's skills are the most important. It is why your numbers tell me nothing. Now I have a check to bring to my bank ;)



Well there is sufficient and not quite enough as far as DR goes. I personally detested the early digital era where we got a similar effect to the early days of chemical photography..white out skies with nothing there. Situation has improved a lot, but not if you're shooting Canon who are lagging behind on DR, you can get by with what's on offer, but they are not up to other makers in that areas.

I could spend a long time counting Wedding photographers who have struggled with DR issues by using Canon. Technique can help but it is one of the reasons I didn't move to Canon (the other being personal handling preferences) None of Canon's FF bodies have DR even up to the better APS-C models, and the margin is quite notable too. You don't always need big DR, but there are times it is very useful.
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: DxO bashing - a bit of perspective
« Reply #57 on: June 22, 2014, 08:03:17 pm »

There is no false statement.

Okay! Let's examine this statement of yours, "the photographer needs to know his gear specificity". The immediate question that follows from such an assertion is, "Why does the photographer need to know his gear specificity ?" I presume we're not talking here about basic procedures in the manual such as how to turn the camera on and off, and how to insert the batteries.

Now according to you, it doesn't make him a better photographer in any way. If this is also true, then one can't help wondering, "Why would anyone, whose goal is to become a better photographer, waste his time learning his gear specifications?" Surely there must be more important things to do.

Your answer to this appears to be, "Knowing the specification of his gear mean knowing how to use it at the maximum, and knowing very well how to PP specifically behind."(I'm  presuming that last part of your sentence means, "specifically, knowing how to post-process very well afterwards".)

Does knowing how to use the gear to its maximum, that is knowing how to use the gear to its full potential, not make one a better photographer? Are you really claiming that an understanding of how to get the most from one's equipment has nothing to do with being a better photographer? If so, perhaps we should define more clearly what constitutes being a 'better' photographer.

As regards post processing, surely that is another, separate, skill which has to be learned in addition to the purely photographic skills of selection of subject, lighting, and the taking of the shot.  Apparently, Henri Cartier Bresson never did any post processing. He was a good photographer, wasn't he? I get the impression that most people using cameras rely upon the post processing built into their cameras. However, I don't.

Quote
All FF cameras since 2008 have enough DR for all major professional work (wedding, products, landscape...).

Well, I don't consider myself to be a professional photographer, so perhaps I'm not qualified to refute that statement. Also, I'm so modest that I would hesitate to claim that my standards must be higher than those required for major professional work.

However, I find such a comment very strange when you include categories such as landscape. I can understand that DR specifically may not be not a concern when taking photos of products in a studio, or in any situation where the photographer can create his own lighting and pre-arrange the composition.
I can also appreciate that portraits of people in general, such as those taken at a wedding, may not require the camera to have an exceptional DR. What's usually more important in those circumstances is SNR at 18% grey.

By contrast, the real dynamic range of many landscape scenes can be huge; far, far greater than the capabilities of the Nikon D800E, for example. We might think that 13 or 14 stops of DR in a camera is very impressive, and it is impressive compared with other, lesser cameras, but not when compared with the real scene. Those last 3 or 4  stops of DR from the D800E, say between the 10th and the 14th stop, will appear quite degraded compared with what the eye saw in the real scene.

The real, practical advantage of a camera with a  DR rating of 14 EV, compared to one with a DR rating of say 11 EV, is that the 14 EV camera has a usable DR of maybe 10 or 11 EV, whereas the 11 EV camera, using the same standards of noise-free acceptability, may have a usable DR of only 7 or 8 EV.

Quote
I finished a weeding one half hour ago and I did it with the Fuji XT-1 only and some good lenses. The D700 and 800 stayed in the car and I KNOW that this little inexpensive camera did the job extremely well.

Can't comment on the performance of the Fuji XT-1. DXOmark haven't shown their test results yet. Also, I'm not sure a camera is a suitable tool for weeding purposes. I prefer to put on some rubber gloves and just pull the weeds out. (Sorry! Couldn't resist.  ;D  )

Quote
I think that photographers, if they focus enough, can output outstanding results out of "obsolete" cameras.

Are you implying that a similar degree of focus and dedication using a modern camera will produce results that are no more outstanding? Since you've already agreed that a photographer needs to know the specifications of his camera in order to be able to use it to its full potential, do you not think that a modern camera used to its full potential will produce, at least some of the time, even more outstanding results than obsolete equipment in the hands of the same photographer?

Quote
The camera is just a dead tool. The photographer's skills are the most important. It is why your numbers tell me nothing.

I'm afraid not, unless perhaps you believe in magic. I haven't yet come across any stories on the internet of any photographer being able to take a photograph with his skills alone, without a camera. I do believe that having a camera is the first essential, wouldn't you agree?  ;)

Also, I see a contradiction in your statement that the numbers tell you nothing. Didn't you previously state that the photographer needs to know the specifications of his camera? Perhaps you do not consider numbers such as '30dB SNR at ISO 3200' a specification.  Such numbers seem quite specific to me.

I'll give you an example of how useful and meaningful the DXO numbers can be. Let's say I'm considering purchasing a Sony RX100 MkII. I'm a bit undecided because I've been disappointed in the past when using P&S cameras, because the noise, DR and high-ISO performance leave much to be desired. How will this RX100-II compare with some of the DSLRs I've used in the past? What can I expect?

The last Canon DSLR I bought was the 15mp 50D, with which I was reasonably pleased. How would the basic, fundamental, RAW image quality of the RX100 II compare with that of the Canon 50D with a roughly equivalent lens attached, such as the EF-S 17-55/F2.8? DXOMark is able to tell me. Checking their charts, I find the following information which I've been able to interpret from the numbers, because I'm such a skilled photographer.  ;)

(1) Color Sensitivity of the RX100 is noticeably better than that of the 50D at base ISO, but slightly worse, to an insignificant degree, at ISO 3200.
(2) Tonal Range of the RX100 at all ISOs is very slightly worse, but to an insignificant degree.
(3) SNR at 18% grey is also very marginally worse, but to an insignificant degree.
(4) Dynamic Range of the RX100 II is actually significantly better than the that of the 50D at base ISO, by one whole stop. However, at ISO 3200 the DR of the RX100 is marginally worse by a less significant 1/3rd of a stop.
(5) The built-in zoom lens of the RX100 II gets a P-Mpix rating for sharpness of 6, whereas the 17-55/F2.8 on the 50D gets a slightly higher P-Mpix rating of 7. Distortion and chromatic aberration is also better in the Canon lens.

Conclusion. The images from the RX100 II, on balance, will be marginally worse, technically, than those from the Canon 50D used with EF-S 17-55/F2.8 lens (a lens I used to use with my 50D), but not to a significant degree and not as much as I expected. I've gone to the trouble of making this comparison because I'm considering getting the latest Mk3 version. However, I shall wait for the DXO tests of the RX100 Mk3 before I make a decision. The improvements in the Mk3 might result in basic image quality being at least the equal of the Canon 50D with 17-55/F2.8 lens, in all respects, and the wider aperture of F1.8-F2.8, of the Mk3, could result in lower noise as a result of sometimes being able to use a lower ISO than I would in the same circumstances with the 50D.

Quote
Now I have a check to bring to my bank


Ah! Now I understand your definition of a good photographer. The more money you earn, the better the photographer you are.  ;)
Logged

hjulenissen

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2051
Re: DxO bashing - a bit of perspective
« Reply #58 on: June 23, 2014, 10:44:55 am »

No, I don't think DR varies like lenses. The quantity of lens variability is shown by the need for focus adjust on high end cameras.
The need for micro-focus adjustment just shows that there is lens variation wrgt off-sensor PDAF. It does not (in my view) prove that there is lens variation when doing manual focus.

Of course, we all (?) agree that there is some variation in lens properties (even when manually focussed).

-h
Logged
Pages: 1 2 [3]   Go Up