Dynamic Range - Tests

[bigmack]

OzRay,

I was incorporating Sugnet and Hickeys observations from above posts as well as yours .My general impression from other data I've seen as well.
Regards.

 

[Gary Eickmeier]

The cameras meter records as 18% any single tone that is placed in
front of the lens (or thereabouts, as there can be slight
variations from camaera to camera). I could have used white, grey,
black or any variation in between.

I used the paper towel because of the texture, which enabled me to
clearly see when the DR had been exceeded (ie the ability to still
record detail). I found that when I used a sheet of white paper, it
was much harder to delineate where things ended.

The ideal test target would be a calibrated grey scale card that
ranged from pure black to pure white in clearly defined steps. I
had no way of getting one, so I followed Michael Reichmann's guide
using a single colour.

You are saying two different things here. Either a single tone is correct or a grey scale is correct. I think the latter.

Your test would be fine if the subject were a grey cat on a grey card in diffuse lighting. But even then, the eyes might blow out long before your test would reveal it.

That doesn't help me when my subject is a black bride in a pure white dress standing next to a groom in a black tux.

Gary Eickmeier

 

[WarrenKK]

I agree with Gary somewhat. If you are using a white towel, what is the range of values in that towel? At the darker or lower range, you are seeing detail in the highlights of the towel, and at the brighter range, you are seeing detail in the shadow areas. This will give you a false idea of the dynamic range. As it is impossible to capture 8 stops of values on paper, a better test might be to have two values on stop apart (ie. two gray cards) next to each other. Expose at different shutter speeds and when you can't distinguish the grays apart from each other, you should have your dynamic range.

Ozray's dynamic range test doesn't seem right to me. He is taking a
single textured sheet and exposing it up and down the scale. But a
subject with a single color or shade is not the problem here. When
we are trying to calculate exposure, we are confronted with varying
tonalities in the image, not a single one. We need to know how much
we can go over and under and still keep detail in both the
highlights and the shadows.

Perhaps a better test would be a very light textured surface
alongside a very dark textured surface. You then go up and down the
scale and see how many stops will hold detail in BOTH surfaces that
is usable.

Gary Eickmeier

--
WarrenKK

PetPeeve: posting a 800x600 to show anything other than how soft my lens is or why I need new glasses.

 

[OzRay]

It doesn't really matter, because you are adjusting your exposure to left and right until there is no discernable change in the tones at either end of the exposure range.

Try it, use anything that is of an even colour and evenly lit and go through the exposure process and you'll see that the results are effectively the same.

All you are doing is over and under-exposing until you cannot see any difference being made by changes in exposure. That's your dynamic range. It's everything between pure black and pure white (some include the pure black and white in the range).

Keep either shutter speed or aperture the same (just for ease of calculating over/under-exposure) and make sure you have enough adjustment to go the full range of whatever you decide to be the variable.

Cheers

Ray

You are saying two different things here. Either a single tone is
correct or a grey scale is correct. I think the latter.

Your test would be fine if the subject were a grey cat on a grey
card in diffuse lighting. But even then, the eyes might blow out
long before your test would reveal it.

That doesn't help me when my subject is a black bride in a pure
white dress standing next to a groom in a black tux.

Gary Eickmeier

--
http://www.australianimage.com.au

 

[Danijel Turina]

I also did tests once and my conclusion was that E1 has around 10 stops of DR, however it is difficult to say how it compares to film, because film has a different response curve. Generally, I'd say E1 can produce results similar to negative film such as Superia, but I prefer to compress the curve a bit and get results similar to Sensia slide, which means contrasty, but with delicate pastel tonality. This is where E1 works best; its default tonality looks very much like Kodak E100G, or E200 - rich, but delicate portrait tonality, with enough contrast and saturation to give it a 3-dimensional look, but not go overboard. The guys in Kodak seem to have designed the sensor to emulate their slide emulsions. The Sony R1 I tested recently emulates a different emulsion - everything about it screams "velvia" - contrasty, highly saturated, very punchy, just right for vivid nature shots. Of course, you can set E1 up to emulate this effect, but every camera seems to have a certain mode where it works best, and for E1 it is portraiture. It absolutely rules at portraiture, it emulates my favorite portrait slide film and I love it.

--
Danijel Turina, http://www.danijel.org/

 

[SalothSar]

Can you perform the same exact test under the same exact conditions with a 35mm film camera? Maybe some store will let you borrow one.

Or here's an idea, buy a cheap, plastic, disposable camera from the department store and take a picture of a grey test chart.

I think you'll be in for a surprise.

---

Think of it this way. Our camera is barely capable of making a 11x17. The future of Olympus will give us the E-7XL or whatever that has things we haven't even dared to dream of yet. Just ten years ago Digital was out of reach for almost everyone. And technology doesn't stop simply because we're used to it.

But I have a feeling that you will have distinct shades on the film than you would with the digital. And maybe with these 16 shades you will see that film easily handles this, while digital lags behind. The test is which camera will handle all of these shades the best; film or digital.



If you wanted to take it a step farther, you could use this chart, and place 1/2 half in shadow & 1/2 in light.



---

I know that you want to be proud of your cameras & that's great. But the fact of the matter is that when I'm doing a wedding, digital still isn't cutting it yet. I have to blend any image that has the bride's gown in it because it's so white & the detail gets lost so easily. If I expose for the gown then everything is thrown into shadow. If I expose for the guy's dark suits then her dress is a white blank nothing. There is not very many stops of dynamic range.

The Fuji S3 adds 2 more stops of DR, that's it. Two. If our camera has 10 stops of DR then why would Fuji spend so much money for just 2 stops. No. The E-series is good. Things will get better. But right now. No. I'm sorry; I'm just not convinced by this test of yours.

I love my Evolt. I think it is a wonderful camera. But I don't think that it has very much dynamic range; certainly nothing extraordinary. Fuji forum has done such tests before & come up with 11, 13, 15, 17 stops and it's ridiculous.

---

Here's the only site I could find quickly enough to satisfy me and prove my point.

http://www.photoreview.com.au/Articlexasp/13f37675-e17f-4011-8d82-9943db73c239/Default.htm

It states that a human can see 11-stops. This is absurd. On a bright sunny day, I can see lots of information in the shadows under tree leaves, a bright blue sky, all the details of a wedding dress, and all the details of a groom's black suit, all at once. Until a camera can compete with the human eye, I'll be blending images to compete.

The point is that you shouldn't believe everything you read even if it is written by someone famous or for whom you have great respect. Common sense tells us that our camera is no different in regards to Dynamic Range than anything else out there. It's a great camera for the time, but DR isn't really a strength. It certainly isn't a weakness either by comparison of today's current technology; but compared to 2010, it's a $50 camera.

I'm not trying to bum you out. If I'm wrong & the film camera utterly fails, then I'm ready to eat my words. But to say that it does without actually testing it does nothing of real value. It's not a good test. And to believe that our camera is better than average to the point that you're delighted is a little like waking up in New Orleans, but hey at least its sunny today.

Oh well, I guess I've said enough. I've rubbed my body with vasoline so that like a prize fighter I could more easily take a shot. So you may now commence with the tomato and rotten vegetable and fruit throwing.

 

[Gary Eickmeier]

I basically agree. But the new Sony R1 has a circuit called AGCS that can compress the histogram or expand it to meet the capabilities of the imager. This new idea will blow this discussion and this problem out of the water.

Gary Eickmeier

 

[OzRay]

I still have film cameras, but there's no need to do such a comparison. What I did was valid, but many forget that digital is linear and film is not. That's why film can, if correctly exposed and processed, give you better results in shadows and highlights. It's different technology.

My test was not to prove which camera was better, but to prove that digital had better than 5 stops dynamic range, which was being touted as the maximum at the time. The test methodology stated by Michael Reichmann is quite sound, it provides results exactly the same as if I were to photograph the test charts that you displayed.

A camera may perhaps never be able to compete with the human eye (never say never), but again we're talking different technology. The eye constantly adapts to lighting conditions in a way that film and digital sensors cannot. Don't forget that the brain also manipulates what you see and adjusts things in a way that a camera cannot. The brain in fact post-processes a lot of what you see and if it were possible to print out the same scene viewed by three different people, you get three different results. The results would also vary depending on the mood of the person, state of health etc.

As for printing capability, it depends on what you are printing. I generally print nothing smaller than 13"x19" and have printed up to 24"x36" from my E1. The prints at 13"x19" can be viewed pretty much with your nose glued to the print and they do not break up. The 24"x36" certainly had to be viewed from further away. Granted, the E1 can't resolve very fine detail like leaves in a distant landscape, but even film has its limitations.

Digital is not perfect by any means and the E1 is not the perfect camera by any means; however, for me, the 'entire' photography experience that the E1 gives me is what's important. I want to go out and shoot and experiment and photograph things I've never considered before, digital motivates me to do that.

For me, there's no joy in the film photography; it's delayed, it's uncertain, it's inflexible, it's expensive (once you get over the DSLR price hump), it's just a pain. I really have no interest whatsoever in film photography and if I had to go back to film, I'd again give up photography.

Cheers

Ray

--
http://www.australianimage.com.au

 

[Dillon James]

That's just a tone curve. Lots of cameras can do that. But wait until you see the noise in the shadows.

I basically agree. But the new Sony R1 has a circuit called AGCS
that can compress the histogram or expand it to meet the
capabilities of the imager. This new idea will blow this discussion
and this problem out of the water.

Gary Eickmeier

 

[Dillon James]

I'm pretty sure the S3 can give me an additional three stops, using ACR or s7raw.

http://www.pbase.com/dillonjames/image/48536720/original

 

[Dillon James]

For me, there's no joy in the film photography; it's delayed, it's uncertain, it's inflexible, it's expensive (once you get over the DSLR price hump), it's just a pain. I really have no interest whatsoever in film photography and if I had to go back to film, I'd again give up photography

Ah, but the smell. The smell of roll film in the morning....

(which movie)

 

[andy mclean]

It is a suprise to see this thread redevlop after so long. I remember shooting text to check DR and concluding it was somewhere around 10 stops or there abouts for the E-1. Many at the time expressed bisbelief, some incredulity and I got hacked off that no-one seemed to believe me, despite owning the camera themselves.

Please correct me if I am wrong, but it seems to me that that the above figs have to be correct; for a JPEG 2^8 = 256 tones split into the 'old zones' as follows;
Zone 1 = 1 tone, black.

Zone 2 = 2 tones, Zone 3 = 4 tones etc up Zone 7 which will be comprised of 128 tones. Zone 8 being white.

With RAW you have 2^12 indivdual tones covering 12 zones in ye olde speak.

The real question is whether you can discriminate detail in the lower zones with so few tones present to express that detail in amongst sensor noise and whether you have access to a good enough printer to express the full tonal variation present. Not to mention the ability of the papers and inks to reflect light covering the full tonal range.

Maybe I am speaking nonesense here, but please correct me if I am!

Best wishes-
Andy

 

[Dan]

Danijel,

I'm interested in your settings for your E-1 and portraits, ie RAW or Jpg, sat, contrast, sharpness, and anything else?

If shooting RAW what engine are you using?

Thanks - I appreciate your response!

Dan