Is Dynamic Range the new albatross?

[Bootstrap]

It used to be "noise" and megapixels and now it's DR.

As I'm not a techie, and that's said with all respect to those who are, I don't get it.
Is there a problem with the "DR" in this image?



Please disregard the "other" problems but as far as DR is concerned, is there a problem?
Perhaps I'm missing something.
Thanks.

[any bets on how fast this thread becomes "Hot?]

--
Troll Whisperer
Bill Turner

 

[JohnMcL7]

Unless you've faked the EXIF on that picture and bought another camera without announcing it to the forum (very unlikely) then it's not particularly relevent - the 'dynamic range' issue is centred around the E-410/E-510 sensor, not the E-330/L1.

John

 

[Bootstrap]

I was under the impression that it was generally a 4/3 problem rather than just E-410/510 specific.
I appreciate your comment.
--
Troll Whisperer
Bill Turner

 

[Vision3]

You raise a good point... it always has to be something with some folks. Finding fault.

Couple this with the cultural obsession with keeping up with the "Joneses" (currently being played by Canon's 5D, MkIII, 40D & Nikon's D2/D3/D300) and you end up with a miserable paranoid few who can't seem to enjoy anything but bashing their own tools.

Every now and then I ask myself, if OLY were the only manufacturer of digital SLRs, would I be happy with the results I'm getting and the answer to that is a resounding YES. Absolutely. OLY dSLRS (in my case, E-510, E-330, E-1 (two), and an E-500 I've sold recently) take wonderful pictures and when done correctly, most don't require much PP. And you know what else, I don't need to bash Canon or Nikon to feel good about my stuff either.

Finally, as to the admittedly overdue E-3 (speaking strictly from a marketing standpoint, OLY has taken way too long to get this thing to market... but again, the reasons are known and they, more than anyone, know that it has cost them, to some degree, market share), if history is our guide, it will BE A GREAT CAMERA. Whatever the specs are (and anymore, I don't care), I just want the d*mn thing here and available for purchase (I also want that 12-60, which quite possibly may become the best walkabout lens anywhere).

So, the short answer is YES it is, but more importantly, those of us invested in the 4/3rds technology don't care, because a). the product(s) work(s) and they work well, and b.) we know what we're doing (I think often the cause of some of these posts is just plain ignorance and the unwillingness to admit it).

So, until Oct. 17th, I'm just going to keep shooting my OLY gear (and making good money at it by the way), producing quality images, and generally having fun.

After that, I'll queue up for the E-3 and do my best to ignore those with inferiority complexes who feel compelled to bash OLY's latest and extol the virtues of "fill in the blank". Why? If you need to ask, then reread the post.

Good shooting...

--
Vision3 (aka., Rich)

 

[Bootstrap]

Thanks for the reasoned reply.
--
Troll Whisperer
Bill Turner

 

[Jay Turberville]

... and you have dynamic range hysteria stew.

First you start with some basic theory about pixel pitch, add some simple math, now toss in a few tests for tonal response and call them dynamic range tests and viola - you gotcha some stew!

So lets unravel it a bit. The pixel pitch is a significant issue. But it isn't the whole story. My 2/3" Coolpix 8400 gives up only one stop of DR to my E-300/330/500. Simple pixel pitch theory math says there should be a two stop difference.

The change in pixel pitch from a 7.5Mp to a 10Mp 4/3" sensor just using simple pixel pitch math should result in around 1/3 of a stop difference. DPReview "measured" twice that. Of course, DPReview says that the Panasonic DMC-L1 has about a half stop more DR than an E-330 that uses the same sensor. That should be a big clue that DPReview isn't really measuring DR.

But there is very good reason to believe that the pixel pitch math shouldn't even be applied to the E-410/510. Panasonic says that they kept the photodiode the same size in the 10Mp L10 that presumably uses the same sensor. So it should have almost exactly the same DR as the E-330 and L1.

Then Olympus committed the sin of letting people make JPEGs with the noise reduction disabled. Oops!!

BTW, I've measured the L1 and it seems to have about 1/10 or so less DR than the E-330. Maybe Olympus is processing the signal a bit more cleanly. I notice pattern noise at high ISO in the L1, so I wouldn't be surprised.

I'll re-extend my invite to anyone in the general vicinity of Phoenix, AZ that I'd be happy to meet up and run some DR tests on one of the new cameras with the 10Mp sensor. That would give an objective comparison to the existing E-300/500/330 line.

So that's the DR hysteria stew as I see it. Lots of steam, some smoke and a rather foul smell.

I was under the impression that it was generally a 4/3 problem rather
than just E-410/510 specific.

Actually, I think you are right. The current concerns seem to revolve around those cameras and projections about the E-3.

--
Jay Turberville
http://www.jayandwanda.com

 

[Jonathan Demarais]

That image is fine. The problem or albatross is lack of DR. Not only does it restrict photographic opportunities, it makes you watch your exposures like a hawk, like you were back shooting slide film! Of course, instant images and histograms help...

 

[smokeyjoe]

First off, it 'became' a problem when people started referring to it as 'dynamic range', as opposed to what it was/should still be called- 'exsposure lattitude.'

That'll tell you something right there: people who were new to photography didn't know what to call it, so they made up a 'new' term. They also brought with them their photographic inexperience which compounded things further.

Not knowing how to expose a photograph properly, has got to be one of the biggest.

I hear people talking about things like 'clipped highlights,' and when they post examples of their stuff the exposure is way off!

FWIW, Bill, your shot has plenty of exposure lattitude. I, however, would have given it about a third to a half a stop more exposure. .......but that's just me.

Cheers!
Joe
--
http://www.flickr.com/photos/58453489@N00/

 

[SirSeth]

... is that limited dynamic range is a digital sensor problem. Meaning that all cameras with a digital sensor have significantly poorer dynamic range than film which has much less dynamic range than we see with our eyes. The only thing I see giving even acceptable dymamic range without blending layers/dodging/burning is the FujiFilm DSLRs. The S5 is slow and has huge files, but it's a real leg up on DR compared with Nikon/Canon/Pentax/Olympus/Sony crowd. Check out the pictures in this thread.

http://forums.dpreview.com/forums/read.asp?forum=1020&message=24709284

So yes, I think the train in your picture, Bill, is much darker than it should be compared with the sunny sides of the pictures if the standard is dynamic range as seen by our eyes. But that said, there are other things that make a good picture besides DR and I have no problem applying blending techniques to overcome some limitations. The MP race is slowing, noise is handled really well now even by Oly, and Oly has busted through the dust problem since the beginning. Live view and stabilization are nice new features available in almost every system. But focus on image quality improvements needs to be more than just low noise at high ISO. So I hope the "hysteria" concerning DR is addressed by more than just FujiFilm in the future. Of course I would love it if the E-3 had a FujiFilm sensor.

Cheerio,
Seth

--
What if the hokey pokey really is what it's all about?

--
wallygoots.smugmug.com
wallygoots.blogspot.com

 

[Rlinnell]

I have an E-1, E510 and recently sold my E500. I honestly haven't noticed any real world difference between the E510 and the Kodak sensors. I highlight the term "Real World" photography because I haven't taken pictures of indoor mini model scenes and then tried to determine if the shadow on a paper clip has "clipped" or not. I will leave those kind of comparisons for others.

I agree with Phil's review and others that the E510 may have a tad less range of say a 5D or even the E-1 but it's not significant. Like I said, has never caused me any issues in the months I have had the E510. I just watch the highlights and the histogram a little closer.

 

[Jay Turberville]

... is that limited dynamic range is a digital sensor problem.
Meaning that all cameras with a digital sensor have significantly
poorer dynamic range than film which has much less dynamic range than
we see with our eyes.

You really should specify which film when making such sweeping generalities.

The only thing I see giving even acceptable
dymamic range without blending layers/dodging/burning is the FujiFilm
DSLRs.

If you aren't dodging and burning, then you aren't exploiting the DR in your capture medium very well.

The S5 is slow and has huge files, but it's a real leg up on
DR compared with Nikon/Canon/Pentax/Olympus crowd. Check out the
pictures in this thread.

Yep. And for those scenes that can use more than 9 plus stops of DR, its a great tool.

So yes, I think the train in your picture, Bill, is much darker than
it should be compared with the sunny sides of the pictures if the
standard is dynamic range as seen by our eyes, and I hope Oly works
on this problem next.

There is no "standard" DR as seen by our eyes. Our brains selectively and actively "dodge and burn" bright and dark areas constantly - which is why even when you have a high DR capture medium, you still usually need to dodge and burn to get pleasing results.

In Bootstrap's image, if you bring up the tones in the train, the sky gets brighter and loses the nice blue. That's not DR, that's a tone issue. Here is a version with a brighter train. No DR was increased in the making of this modification.



--
Jay Turberville
http://www.jayandwanda.com

 

[Ross]

Yes, and the albatross is laughing its head off! It remembers Kodachrome and the five stops of exposure lattitude that made decades of National Geographics Magazine the showcase of low DR pictures. It is a good thing they are now going digital so we can see more of that fine flat modern look of HDR.
Here is to the compulsion to open every shadow to see what is within.
--
Bob Ross
http://www.pbase.com/rossrtx

 

[PDXguy]

It seems the Oly DR issue has been brought to light more frequently since the intro of the 410/510. Perhaps because of Phil's concern after reviewing the 410? Perhaps Oly's rabbit in the hat will be a solve for this...perhaps that HDR patent talk a few months back will come into play? If not the E-3...the E-5?

I have to agree with Bill here, I really don't see much of a problem with DR, in particular with the E-330 or even E-500. Admittedly, it could be better with the E-1. I feel that knowing the limitations of the tools you use will make you a better craftsman, resulting in a better end product because of learned technique with the tool at hand. God I knows I have my own road to walk.

BTW...nice shot, Bill

Unless you've faked the EXIF on that picture and bought another
camera without announcing it to the forum (very unlikely) then it's
not particularly relevent - the 'dynamic range' issue is centred
around the E-410/E-510 sensor, not the E-330/L1.

John

--
Regards,
Steve

 

[Vision3]

As one of a handful who give the forum class & distinction, you've earned, at least in my book, a certain amount of deference with regard to your posts which are, without exception, meaningful, thoughtful, & timely. Wish there were more like you...

If you're ever around San Antonio, you'll be welcome.

--
Vision3 (aka., Rich)

 

[Clark666]

http://forums.dpreview.com/forums/read.asp?forum=1019&message=24862856

I plan to buy an E3 if it has features that I like. I am absolutely certain that it will have flaws that I will have to work around.
--
Olympus E510, 40-150, 50-200
c-50,
Canon Elan 7, XTi 28-90, 100-300, 50 f/1.8, sigma 10-20
http://www.flickr.com/photos/75673106@N00/

 

[Timskis6]

AND the MP count is too low.

=) Couldn't resist.

But seriously, you're definitely on to something. First it was MP, then noise, and now DR. The arguments will always still be there, but they tend to shift from one specification to another when the previous is exhausted.

What will it be next?
--
Tim
'Be the change you wish to see in the world.' -Mahatma Gandhi
http://www.flickr.com/photos/timskis6/

 

[A Aquacelli]

It's not just about not having blocked shadows or blown highlights. It's about a nice range of tones, especially in darker areas. Smaller sensors are stretched to the limit in this regard, as evidenced by the greens in your grass and the red in your caboose.

 

[silentstorm]

I 2nd that.

I've seen many pics floating around the web & noticed the DR for 410 & 510 limiting compare to my E1. I'm not at all keen to get either of those.

What I've noticed was that if you preserved the highlights, you compromised with the mids & blacks. When you boost it in post, the mids & blacks get ugly. The other way round, you lose the highlight details but retain the mids & blacks. This compromise is not what I'm willing to make.

My E1 has many dead pixels now despite the mapping feature. Addition to it, the E3 is nowhere to be found. I'm not too sure if Oly is going to get my continual support now. :-S I certainly hope the E1 successor isn't plague by this limiting DR.

btw, anyone has info on what is the pixel pitch size for the 10MP NMOS sensor used in E410/510?

 

[Rlinnell]

I have to agree with Bill here, I really don't see much of a problem
with DR, in particular with the E-330 or even E-500. Admittedly, it
could be better with the E-1.

This I this is the first time I have seen anyone claim that either the E300 or the E500 had more"DR" than the E-1. I have found the opposite to be true. I have found the E-1 to have the most latitude of any of the Oly E cameras I have owned (E-1, E500 and E510). I have however become VERY fond of the E510 for it's other strengths. I have grown to love this camera!

Regards,
Steve

 

[John Cozijn]

Dynamic range for digital cameras can actually be quite precisley specfied as a property of the sensor expressed in the formula 20xlog (Saturation/Noise), with the values in parentheses being the number of photo-electrons.

We actually know what these numbers are for the e-1 and the e-300/e-500 (since Kodak published their specs), which are 67db and 64db respectively (and I expect all DSLRs to be within the 60-70db range).

But of course that's not the sense in which people are now using the term - it's been redeployed to mean the number of usable stops a camera can reproduce - from blocked shadows to blown highlights.

There are a number of problems with this usage, not the least of which is how to measure it. The most serious attempt to find an objective measure is Imatest, but like Phil's test procedure it is based on analysing a step-wedge. But that so completely abstracts from actual photography that I regard the results as of academic interest only. And what is "usable"?

The nature of the noise (luminance or chroma), how the image is exposed (under or over), and indeed the quantity/quality of the light one is working with seem to play a much larger role in the real world. And to go one step further, the issue is not so much the capability of the camera to encompass a wider luminance range but the craft of the photographer in capturing the range that will work for the image in question.

The bottom line is that with DSLRs through all price brackets the difference in exposure range is probably no more than 1 stop in RAW, with a median value that is somewhere between transperancy and negative (6 to 9 stops, respectively).

As for the Olympus cameras, my impression is that the nMOS sensors are a tad more intolerant of underexposure than the Kodak sensors, but it seems to me the important issue is the quality of what is happening in the captured range rather than fretting over the fringes.

Cheers, John