Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
11-05-2020, 10:35 PM   #31
Senior Member




Join Date: Aug 2009
Location: Broomfield, Colorado
Photos: Gallery | Albums
Posts: 152
Original Poster
QuoteOriginally posted by stevebrot Quote
Is there a glossary of terms? I am old school where dynamic range is the range of light brightness that may be rendered with detail (i.e. at least two tones). In practical terms, this would be the range of light bounded by clipped highlights on the high end and clipped shadow on the low end. Since we are talking digital capture, that means the range of 1 to ((2^14) -1) in the RAW data, the boundaries being zero and clipped white.
I don't think the definition of Dynamic Range has changed, it's just being measured in a more detailed manner over the camera's range of ISO Settings. It reveals any weaknesses in a specific model, especially if the sensor itself is adding extra noise which degrades Dynamic Range.
Here's a nice interactive tool for comparing different DSLRs...

Photographic Dynamic Range versus ISO Setting

---------- Post added 11-05-2020 at 11:00 PM ----------

DXOMARK evaluates a DSLR sensor's low-light performance with the "sports" rating.
Here's a link to their article explaining a little more about it.

DXOMARK camera sensor testing protocol and scores

---------- Post added 11-05-2020 at 11:02 PM ----------

Here's results for a Pentax K-1

Pentax K-1 - DxOMark

11-06-2020, 11:00 AM   #32
Otis Memorial Pentaxian
stevebrot's Avatar

Join Date: Mar 2007
Location: Vancouver (USA)
Photos: Gallery | Albums
Posts: 42,007
QuoteOriginally posted by BB_Zone28 Quote
I don't think the definition of Dynamic Range has changed, it's just being measured in a more detailed manner over the camera's range of ISO Settings.
OK, so what is meant by RAW Dynamic Range and how is it measured?


Steve
11-06-2020, 11:04 AM   #33
Loyal Site Supporter
Loyal Site Supporter
rogerstg's Avatar

Join Date: Oct 2007
Location: Rhode Island
Photos: Albums
Posts: 1,168
QuoteOriginally posted by BB_Zone28 Quote
Do you remember the HUGE Pentax 6x7cm SLR camera???
Yes, but I was referring to the pixel size, not camera size. You raise good points though.
11-06-2020, 11:33 AM   #34
Site Supporter
Site Supporter




Join Date: Dec 2015
Location: Goldsboro North Carolina
Photos: Gallery | Albums
Posts: 2,860
The dynamic range vs. ISO chart above shows me that my KP has about the same DR at ISO 640 that it does at ISO 100, better than it has at ISO 400. Interesting....

11-06-2020, 11:52 AM - 1 Like   #35
Senior Member




Join Date: Aug 2009
Location: Broomfield, Colorado
Photos: Gallery | Albums
Posts: 152
Original Poster
QuoteOriginally posted by rogerstg Quote
Yes, but I was referring to the pixel size, not camera size. You raise good points though.
I was referring to both the pixel size and how it can impact the physical sensor size in order to achieve higher image resolutions, and thereby the ergonomic impact on the physical camera size in order to house the larger sensor.

Last edited by BB_Zone28; 11-06-2020 at 12:07 PM.
11-06-2020, 12:06 PM   #36
Otis Memorial Pentaxian
stevebrot's Avatar

Join Date: Mar 2007
Location: Vancouver (USA)
Photos: Gallery | Albums
Posts: 42,007
QuoteOriginally posted by Apet-Sure Quote
The dynamic range vs. ISO chart above shows me that my KP has about the same DR at ISO 640 that it does at ISO 100, better than it has at ISO 400. Interesting....
Prevailing wisdom is that 640 is somewhere around the point at which the accelerator chip kicks in. What I have found interesting is that PDR varies by image format.


Steve

Last edited by stevebrot; 11-06-2020 at 12:15 PM. Reason: clarity
11-06-2020, 12:45 PM   #37
Site Supporter
Site Supporter




Join Date: Dec 2015
Location: Goldsboro North Carolina
Photos: Gallery | Albums
Posts: 2,860
QuoteOriginally posted by stevebrot Quote
Prevailing wisdom is that 640 is somewhere around the point at which the accelerator chip kicks in.
I see. Is there a good resource to read up on this mysterious chip? Maybe it's only mysterious to me.

11-06-2020, 01:31 PM   #38
Otis Memorial Pentaxian
stevebrot's Avatar

Join Date: Mar 2007
Location: Vancouver (USA)
Photos: Gallery | Albums
Posts: 42,007
QuoteOriginally posted by Apet-Sure Quote
I see. Is there a good resource to read up on this mysterious chip? Maybe it's only mysterious to me.
It is mysterious to us all, but it is that chip that provides the KP's and K-1ii's high ISO capabilities. Theory is that it works by doing NR on the sensor output before sending the data to the image processor. This has led to claims that it eats detail, but evidence has not been compelling. It would be nice if Ricoh would provide a setting to allow it to be switched off.


Steve
11-06-2020, 04:25 PM - 2 Likes   #39
Moderator
Loyal Site Supporter




Join Date: Apr 2014
Location: Baltimore
Photos: Albums
Posts: 3,383
QuoteOriginally posted by stevebrot Quote
It is mysterious to us all, but it is that chip that provides the KP's and K-1ii's high ISO capabilities. Theory is that it works by doing NR on the sensor output before sending the data to the image processor. This has led to claims that it eats detail, but evidence has not been compelling. It would be nice if Ricoh would provide a setting to allow it to be switched off.


Steve
What I'd say is that evidence has been...confusing. In some cases it's done a great job, and seems to actually extract a bit more detail. In others it seems to mush things out slightly.

My personal impression is that for coarser details (like leaves in a landscape) it's great, but for extremely fine detail (like fine fur) not as good.
11-06-2020, 11:42 PM   #40
Senior Member




Join Date: Aug 2009
Location: Broomfield, Colorado
Photos: Gallery | Albums
Posts: 152
Original Poster
QuoteOriginally posted by stevebrot Quote
... What I have found interesting is that PDR varies by image format.
Stating the obvious, I would expect to see variations due to RAW files being 12 bit or 14 bit and compressed file formats being 8 bit.

Also as more processing power becomes available, interpolation algorithms including dynamic range from RAW to compressed formats along with other performance trade offs tend to steadily improve if the past is any indicator of things to come.

Last edited by BB_Zone28; 11-07-2020 at 12:18 AM.
11-06-2020, 11:58 PM   #41
Otis Memorial Pentaxian
stevebrot's Avatar

Join Date: Mar 2007
Location: Vancouver (USA)
Photos: Gallery | Albums
Posts: 42,007
QuoteOriginally posted by BB_Zone28 Quote
Stating the obvious, I would expect to see variations due to RAW files being 12 bit or 14 bit and compressed file formats being 8 bit.
I don't follow. The KP is 14-bit as is the K-1ii in both FF and crop mode and the site uses RAW exclusively for their analysis, yet the K-1ii in crop mode is deemed by photons-to-photos as having lower dynamic range than in FF mode (similar to the KP, despite lower pixel density). Does that mean that if I do a non-destructive crop (not applied until rendered) in Lightroom, that the resulting final image (say exported to 16-bit TIFF) has less DR than if it were not cropped?

DxO does something similar with their FF sensor ratings, giving them a higher rating solely on the basis of format.


Steve

Last edited by stevebrot; 11-06-2020 at 11:59 PM. Reason: clarity
11-07-2020, 10:46 AM - 1 Like   #42
Senior Member




Join Date: Aug 2009
Location: Broomfield, Colorado
Photos: Gallery | Albums
Posts: 152
Original Poster
QuoteOriginally posted by stevebrot Quote
I don't follow. The KP is 14-bit as is the K-1ii in both FF and crop mode and the site uses RAW exclusively for their analysis, yet the K-1ii in crop mode is deemed by photons-to-photos as having lower dynamic range than in FF mode (similar to the KP, despite lower pixel density).
You ask some really probing technical questions!

I misunderstood your previous comment about different formats being about file formats, not changing image capture size using the same file format.
That said, variations in DR between FF and cropped captures from the same sensor does seem both surprising and perplexing. Also, you may be on to something about how cropped and FF sensors/capture modes ultimately get evaluated for the benchmarks of the testing site.

I can only guess (aka my 0.02) it has something to do with what pixel data is read, how it's processed, and when it's processed. What's happening with noise or signal bleed from adjacent pixels if they aren't read?? Or if adjacent pixels are processed together as one pixel to theoretically reduce its signal to noise, is the IP algorithm doing something different that affects IQ when it also has to be optimized for speedy output??? One could even question the effectiveness of the subject content and lighting being captured to produce images for either laboratory or real world performance evaluations.

Last edited by BB_Zone28; 11-07-2020 at 11:04 AM.
11-07-2020, 11:39 AM   #43
Otis Memorial Pentaxian
stevebrot's Avatar

Join Date: Mar 2007
Location: Vancouver (USA)
Photos: Gallery | Albums
Posts: 42,007
QuoteOriginally posted by BB_Zone28 Quote
Also, you may be on to something about how cropped and FF sensors/capture modes ultimately get evaluated for the benchmarks of the testing site.
It has been a long-standing criticism of DxOMark methodology. They used to have an explanation on their site, but that seems to be gone. IIRC, the explanation was lame, but that may just be my memory. It is enough to note that FF always outperform crop sensors there and at photons-to-photos, even those with the same or similar pixel densities.


Steve
11-07-2020, 05:24 PM   #44
Veteran Member




Join Date: May 2016
Photos: Gallery | Albums
Posts: 2,722
Can I shoot at iso100 and keep cranking the exposure till I get to iso1mil?)
11-07-2020, 09:14 PM - 1 Like   #45
Pentaxian




Join Date: May 2009
Location: Somewhere over the rainbow
Photos: Gallery
Posts: 1,528
QuoteOriginally posted by BB_Zone28 Quote
That said, variations in DR between FF and cropped captures from the same sensor does seem both surprising and perplexing. Also, you may be on to something about how cropped and FF sensors/capture modes ultimately get evaluated for the benchmarks of the testing site.
What is the variable that people are missing is how we determine the floor at which the DR is limited and this is noise. Now you have to remember that noise is bandwidth dependent what this means is how closely we are going to view it.
The very same can be said about how much noise is visible as you increase the iso in your camera depending on how much we are going to zoom in as look at the photo will determine how much noise your are going to see.

When we are using only the cropped portion of the sensor your are basically zooming in more to makeup for what was cropped away. This will then appear as more noise just the same as iso 1600 on fullframe looks as noisy as 800 on cropped.

A very simple test than can be done is to take an image of a white piece of paper underexpose it by 10 stop and then lift the shadows by 10stop. Now take that image and print it in 8-10, then take the very same image and crop it by a factor of 1.5 and then print that also in a 8-10 and view them side by side.

This should also allow you to see that is more that the total amount of light that is captured ( via larger use of the sensor) that determines the amount of noise and DR you can capture.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
camera, dxomark, iso, iso 1 million, k-1, light, million, pentax, photography, pm, post, range, sensor, technique, terms
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Nature "Million Bells" I Believe. Tonytee Post Your Photos! 4 11-03-2020 02:27 AM
Theta360 leak exposes 11 million photos, user data interested_observer Ricoh Theta 360° Cameras 2 06-03-2019 12:43 PM
For Sale - Sold: Camera Bag: Crumpler 6 million Dollar Home 7.62lew Sold Items 2 01-09-2019 07:42 PM
Noisy ISO Club! paperbag846 Pentax DSLR Discussion 23 12-15-2010 03:50 PM
ISO digression from 135 club thread Class A Pentax SLR Lens Discussion 33 05-13-2010 04:40 PM



All times are GMT -7. The time now is 02:07 AM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top