Posts for: bclaff
Jul 3, 2018 12:28:55 #
bclaff wrote:
DxOMark determines something they call "Measured ISO".
At PhotonsToPhotos a good place to look is in the sort-able table at underneath the DxOMark Derived Measured ISO chart.
Note the Factor and Stops columns; most cameras are at at least -3 rather than 0. And follow the links under "Further Reading".
At PhotonsToPhotos a good place to look is in the sort-able table at underneath the DxOMark Derived Measured ISO chart.
Note the Factor and Stops columns; most cameras are at at least -3 rather than 0. And follow the links under "Further Reading".
Correction ! -1/3 not -3
Jul 3, 2018 11:24:18 #
Cdouthitt wrote:
Just curious, where are you seeing/reading that Olympus places their middle gray at a different place in the raw data when compared to other camera brands?
DxOMark determines something they call "Measured ISO".
At PhotonsToPhotos a good place to look is in the sort-able table at underneath the DxOMark Derived Measured ISO chart.
Note the Factor and Stops columns; most cameras are at at least -3 rather than 0. And follow the links under "Further Reading".
Jul 3, 2018 10:48:17 #
Cdouthitt wrote:
The thing is, I've been able to see quite a big difference between the original EM1 and the new EM1ii, especially in bright situations (dropping the highlights and pulling up the shadows)...so frankly I believe the data as I have seen the difference with my own eyes. Now is it as good as they say it is...I can't say as I don't have any APS-C gear to test it against in the same conditions.
Sure, the E-M1 Mark II does have more PDR than the E-M1.
But comparisons against other brands and sensor sizes is obfuscated by their choice of where to place middle gray in the raw data.
Jul 3, 2018 08:14:04 #
Cdouthitt wrote:
Was just reading an article about dynamic range vs ISO...and found it interesting that the EM1ii was holding it's own against larger sensors. Link to follow. I know I've personally noticed an uptick in performance in dynamic range since upgrading from the original em1...but I really didn't expect it to be pretty much on par with other APS-C sensors.
Your Mileage May Vary (YMMV).
As someone is sure to point out some brands, like Olympus, place middle gray at a different place in the raw data.
This doesn't mean that the dynamic range as measured is wrong, just that middle gray is lower than for most other cameras.
On The Other Hand (OTOH), middle gray will be in the expected place in JPEG files.
And ... who knows where middle gray will be in your image requiring high dynamic range?
Some people feel PhotonsToPhotos ought to use "Measured ISO" as the x-axis on these charts.
But because of variations, much as metering, I've determined that is no more useful than using the manufacturer stated ISO.
The Laws of Physics dictate that in general sensor area is going to control the maximum possible Dynamic Range (DR); so MFT can't reach the performance of APS-C, etc.
Apr 3, 2018 14:52:18 #
burkphoto wrote:
Given our current limited understanding of physics...
Your faith in future technology conflicts with the laws of physics, not just our current understanding.
Ultimately, even if read noise is brought down to zero; we are limited by photon noise, and we're not all that far from that limit now.
The "ideal" lines on the PhotonsToPhotos Photographic Dynamic Range (PDR) chart are drawn at the limit of photon noise.
In the near term the most improvement is in computational imaging which is currently being implemented in mobile phone cameras.
Apr 3, 2018 14:46:57 #
amfoto1 wrote:
...
Image "noise" essentially occurs when the wrong color is recorded by one or more individual pixels on the sensor.
...
Image "noise" essentially occurs when the wrong color is recorded by one or more individual pixels on the sensor.
...
Pixels don't record color. They record (in a linear fashion) the amount of light that gets through a color filter.
Color is calculated later in a process called demosaicing.
Inaccuracy (noise) in one pixel can affect the demosaicing and result in a less accurate hue (color).
Apr 3, 2018 12:41:32 #
Yes, the temperature of the sensor and the immediate surroundings have an effect on dark current.
(Extra electrons that are mistaken for signal.)
This is why some astro-photographers use cooled camera systems.
(Extra electrons that are mistaken for signal.)
This is why some astro-photographers use cooled camera systems.
Apr 3, 2018 12:37:11 #
Ron Dial wrote:
Noise is caused by heat across the camera sensor. The higher the ISO, the greater the heat, and the greater the noise. Camera Manufacturers have software that tries to deal with noise but it is an ongoing problem
Sorry but heat is not a measurable part of the issue.
Most noise that people find objectionable/noticeable arise from not gathering enough light.
Apr 3, 2018 10:26:26 #
TriX wrote:
If you really want to understand the mechanism that various components play in the noise of an image, the manner and which ISO adjustments in-Camera are made, and the effects of underexposure or ETTR, let me suggest the following read: http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p2.html
That is a great article but the original now has many broken links.
Emil gave me permission to put up a copy at PhotonsToPhotos where I repaired all the problems:
Noise, Dynamic Range and Bit Depth in Digital SLRs by Emil Martinec
Jan 6, 2018 00:00:24 #
blackest wrote:
bclaff = Bill Claff
Now I feel kinda stupid arguing with the fella that put together the data ..
How far wrong am I?
Now I feel kinda stupid arguing with the fella that put together the data ..
How far wrong am I?
I thought your initial post was pretty good and I was just clarifying that human vision and digital cameras are fundamentally different including on what dynamic range means.
Who I am and what I do gives me experience but what I say and how I back it up is what matters. I try to communicate well but sometimes I fail.
I'm quite sure you aren't stupid; just as I'm equally sure we are all, including myself, ignorant in some areas.
I try to be a little less ignorant every day.
Jan 4, 2018 16:41:35 #
burkphoto wrote:
No, that's not the point at all. The point is that sometimes it matters and sometimes it doesn't. Just use the right tool for the job.
...
Most of the masters of the last 100 years used very good, but not always the best, equipment. Chance favors the prepared mind.
...
Most of the masters of the last 100 years used very good, but not always the best, equipment. Chance favors the prepared mind.
It favors those who understand the capabilities and limitations of their equipment; that includes things like dynamic range.
Jan 4, 2018 12:24:29 #
burkphoto wrote:
I wasn’t writing about photographers... I was writing about the viewers of their work. Average Joe and Mary don’t care whether you got a stop more noise than the next guy.
That's a self-selected sample. They only see the photos that were "keepers" and for some photographers dynamic range and low light performance is what makes the difference.
burkphoto wrote:
Yes, if you only come out in dim light, or you make huge prints from high ISO images, DR and high ISO performance certainly matter. But a one or two stop difference at the margin is rarely a showstopper. Less than ideal? Surely. Will you and I notice? Yeah.
A stop is a huge difference to me and to many others.
burkphoto wrote:
However, I would never carry three times the weight or spend three times as much for something I can rent for the short period of time (<5%) that I might need it. The rest of the time, I’ll go with efficiency.
It would be a different story if 95% of my work were low light, or gigantic prints to be viewed from abnormally close range.
It would be a different story if 95% of my work were low light, or gigantic prints to be viewed from abnormally close range.
Once again you put forward your own personal experience and preferences as if they are universal; they are not.
Jan 4, 2018 12:08:41 #
BebuLamar wrote:
DXO measured dynamic range in EV and EV is logarithmic. It's log base 2.
It's a logarithm of a linear measure.
Jan 4, 2018 09:08:06 #
burkphoto wrote:
... However, I'm continually amazed at the emphasis folks on this site place on noise, dynamic range, and other technical nuances. In actual, real-world usage, there is seldom much difference to notice unless you are a highly critical, seasoned photographer. The rest of the world (about 7 billion of us) doesn't know or care.
...
...
I'm continually surprised at sweeping hyperbole based on personal observation.
If you don't photograph in low light or need high dynamic range than naturally you don't care.
But I can assure you that a significant percentage of photographers; wildlife, landscape, astro, etc. do care (and for a reason).
BTW, there are far fewer than 7 billion people using a digital camera; another example of hyperbole.
Jan 2, 2018 11:24:32 #
bclaff wrote:
It's important to realize that such sensor testing reports dynamic range in linear space.
But the final image, and human vision, operate in a more logarithmic fashion.
So any comparisons would be "apples to oranges".
But the final image, and human vision, operate in a more logarithmic fashion.
So any comparisons would be "apples to oranges".
blackest wrote:
isn't that the purpose of the F stop ? each stop is double the value of the one before. so they are log values.
Maybe what I have is an over simplification but it seems fair.
Maybe what I have is an over simplification but it seems fair.
Your response doesn't have to do with my warning.
Dynamic range reported by PhotonsToPhotos, DxOMark, etc. is linear dynamic range.
Dynamic range that you see is logarithmic.
The raw linear data goes through a power function and tone curve to make an image for you to view.
So linear dynamic range cannot be compared to human vision dynamic range in a meaningful way.