Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
05-24-2016, 04:06 PM   #31
Pentaxian




Join Date: Sep 2013
Location: Sydney
Photos: Albums
Posts: 844
QuoteOriginally posted by Mountain Vision Quote
Seems to not be an issue with Sony sensors?
Looking at the D810, D7200, and 645Z graphs, it would appear so (good! that matches my general experience with the K-1 & K-3!)

http://www.photonstophotos.net/Charts/PDR.htm#Nikon%20D7200,Nikon%20D810,Pentax%20645Z

05-24-2016, 04:39 PM   #32
Senior Member




Join Date: Dec 2013
Location: Buffalo, NY
Posts: 257
Well that certainly is interesting...
Even more interesting is that on the Nikon D5 it starts out dipping on the 1/3 and 2/3 stops, but after ISO 400 the 2/3 stops do better than the full stops.

I'm going to guess it has more to do with the processor chip circuitry or algorithms than the sensor itself, but that sure is strange!
05-25-2016, 02:49 AM   #33
Veteran Member
Tan68's Avatar

Join Date: Sep 2014
Photos: Albums
Posts: 396
I think the comment is based (mostly) on old knowledge...
I didn't look at the link above to the info about Canon cameras but appears the instructor's comment may still be valid at times.

I remember hearing similar comments around the time of the K20. Quality was better to some small degree using full stops for ISO. I never checked into this, myself.

I did learn that quality was better exposing to the right even if I had to raise ISO. If I was constrained by shutter and aperture and the image would be poorly exposed so that I had to later brighten things with PP, I would be better off to increase ISO in the camera until I had exposed right (as described above including later decreasing brightness in PP) rather than increase exposure in image editing software. The rule to always use as low an ISO as possible was not true. This held true only to ISO 1600. Granted, the differences in quality were small but they were apparent. I never used over 1600 for any reason.. This is the opposite of the newer 'ISO-invariant' cameras.

As far as ISO invariant sensors now, I will still try to set ISO mostly correct in camera. The DPR ISO invariant tests still show better quality when ISO is set in camera versus using base ISO and lifting all at home in PP. I see a small difference and this is only for the cameras I have reviewed, of course, so your mileage may vary.

What ISO invariance does mean to me is that I don't have to worry so much with having ISO set precisely. Unlike with the K20, the K5 has enough additional latitude that I don't worry so much with exposing to the right and etc. I get a good exposure and adjust the brightness in camera by raising the ISO, protect highlights, and fine tune things at home. I will continue to do this with K-1. I will not use base ISO and lift brightness/exposure completely in PP because the sensor is considered 'ISO-less'..
05-25-2016, 04:00 AM   #34
Pentaxian




Join Date: Jul 2009
Location: Tromsų, Norway
Photos: Albums
Posts: 951
I tried to find out wich of cameras that have a dip in DR for the in between ISOs. Here are the list of cameras that have repetitive dips (at least two). The graph are too messy, but the list of camera names should be easy to read. I cant see a clear factor that they have in common that may explain the behavior. It occurs at all price levels. Its a pitty they only tested one Pentax camera (645Z).


Last edited by Simen1; 05-25-2016 at 05:31 AM.
05-25-2016, 04:47 AM   #35
Senior Member
rangercarp's Avatar

Join Date: Oct 2012
Location: Michigan, Kansas, New Mexico
Posts: 236
Original Poster
Thanks for all the replies. It is nice to get some definitive answers. The few old threads I found had only a handful of replies and did not reach any conclusions. It is a shame DXO did not test more Pentax cameras, but the graph does give us a good generalization of where the problem lies.
05-25-2016, 05:43 AM   #36
Pentaxian
Digitalis's Avatar

Join Date: Mar 2009
Location: Adelaide.
Posts: 8,704
I use 1/3rd stops for all the things, As a studio photographer I need to wring every ounce of Image quality out of my files so high precision exposure is the ticket for me. Some of the older Canon 1 series cameras used different methods of amplification to simulate different sensitivities, and in some instances intermediate ISOs such as 320Vs 400 would show marked difference in overall image noise and tonality.

What the instructor said was a rather sweeping generalization- Sony sensors have entirely different amplifier design and technology from Canons in-house designed sensors, when it comes to IQ there is no singular universal approach that will produce equal results across all sensor architectures.
05-25-2016, 06:19 AM   #37
Loyal Site Supporter
WPRESTO's Avatar

Join Date: Aug 2010
Location: Massachusetts
Photos: Gallery | Albums
Posts: 25,175
So many replies I cannot read all of them. IMHO a 1/3 EV step change is about the smallest exposure change that makes a noticeable difference, almost anyone can detect a 1/2 EV change, and one full stop is a substantial change, sometimes too much. I generally select 1/3 EV steps, but I cannot always see the difference. I've always thought it was hilarious overkill that hand-held meters I used had 1/10 EV steps. That belies a wildly optimistic faith in the accuracy of shutter speed and aperture settings.
05-25-2016, 07:31 AM   #38
Pentaxian
Digitalis's Avatar

Join Date: Mar 2009
Location: Adelaide.
Posts: 8,704
QuoteOriginally posted by WPRESTO Quote
...belies a wildly optimistic faith in the accuracy of shutter speed and aperture settings.
I have often thought the same thing, even with advanced electronically controlled shutters from seiko the aperture mechanism in the lens or the film may not be up to speed, so there is always a margin of error wide enough to take up the slack. Hence 1/10th EV stops which are used as a method of compensating for such variables. The flash units I work with are also able to use 1/3rd exposure increments, which is becoming increasingly common and is very handy when you are delicately balancing the output of light sources.

05-25-2016, 07:38 AM   #39
Pentaxian
The Squirrel Mafia's Avatar

Join Date: Jan 2015
Posts: 822
I set my K-50 to shoot at 1/3rd stops as soon as I found out that it was capable of doing so. That was about a couple of weeks after ownership. Been shooting like that since then. No issues here & I can see the difference between ISO 4000, 5000, & 6400.
05-25-2016, 09:45 AM   #40
Site Supporter
stevebrot's Avatar

Join Date: Mar 2007
Location: Vancouver (USA)
Photos: Gallery | Albums
Posts: 27,125
QuoteOriginally posted by Simen1 Quote
There is some weird performance drop for the in between ISOs of some Canon and Nikon cameras. Click
QuoteOriginally posted by Alex645 Quote
Wow. This is a "smoking gun". I'd never seen this before, either as data or from anecdotal experience shooting both Pentax and Nikon. Thanks for sharing this....IF I shot Canon, that's a big deal.
QuoteOriginally posted by Digitalis Quote
Some of the older Canon 1 series cameras used different methods of amplification to simulate different sensitivities, and in some instances intermediate ISOs such as 320Vs 400 would show marked difference in overall image noise and tonality.
I underlined a portion of Digitalis' comment for emphasis because those few words are telling. There are many ways to capture, modify, and utilize the sensor capture in terms of sensitivity. My (very limited) understanding is those methods fall into a couple of broad categories:
  • Adjusting detector gain. This is similar to turning up the volume on the radio.
  • Tweaking analog/digital conversion (x amount of charge = 2 bits rather than 1)
  • Tweaking digital values during translation to RAW. Nasty as this might sound, this is a simple bump of values and is equivalent in some ways to brightening an image in PP.
All of the above is an attempt to normalize sensor response to something that approximates linearity (twice the light = doubling of the numeric value on output) and each of the above has its own trade-off in regards to quality. Note that the list above pertains to RAW capture. The processing to make an in-camera JPEG or TIFF introduces additional variables.

Are things looking complicated yet? I hope not too much so, because here are a few other mind benders, a few of which might clarify things and put some demons to rest.
  • The historic root is the ISO sensitivity of photographic film where density on an exposed negative is plotted against a luminance standard. The film speed is the point of x density above background (fog). As might be imagined, the actual response of various films varies quite a bit depending on processing, emulsion physical characteristics, and the nature of the response curve beyond the point at which speed is measured. (Sound familiar? See list regarding digital processing above.) Many serious film photographers settle on using the ISO rating as a starting point and set their meters to an Exposure Index (EI) representing how the film responds for the intended development process.
  • ISO sensitivity for digital capture is not standardized in the usual sense and is a fuzzy concept. There is a standard, but it is a little like the Pirates Code in "Pirates of the Caribbean"...advisory in nature. The intent is to provide a response that is functionally equivalent to using photographic film of a given ISO sensitivity. In truth, the values set on the digital camera are closer to an EI (see above) than traceable to an ISO standard. The camera simulates film sensitivity* (see Digitalis comment quoted above). As a result, there is considerable implementation variability between manufacturers and camera models including those that reputably use the same sensor.
  • Because of the above point, comparing ISO for camera A vs. camera B can be a little problematic since the numbers assigned for each pixel and interpreted by the RAW processor (oops...another point of variation...the curves are "bent" for effect...) reflect the intent of the engineering group as guided by the project managers and marketing.
  • Because of the above point, it is highly likely that some points on the sensitivity scale may result in non-linear image degradation as ISO ramps. How and why would depend on camera brand and the state of tech at the time the camera was made.
  • The notion of "base ISO" is a little artificial. In most of our minds, I think it means the sensitivity where minimum gain is applied during capture. The marketing departments as well as "engineers on display" at shows are a little evasive on this number, perhaps because it may not be that simple.
  • So-called ISO invariance is an outcome-based observation and is a reflection of how the manufacturer manages dynamic range. The term (Edit: ISO invariance) magically appeared about a year ago, I think as a covert effort by the marketing departments of certain brands whose cameras exploit intrinsic dynamic range for their sensitivity settings. In practical terms, some of the online claims (shoot low and deal with it in PP) are a little like saying that a high torque engine on an automobile makes transmission gearing obsolete. You can do it, but YMMV. Personally, I would not make ISO invariance a critical part of the decision process for purchase or as an indication of intrinsic capture quality.
  • There is nothing intrinsically magical about ISO setting numbers in "full-stop" increments starting (arbitrarily) at ISO 5. They are easier to remember, but that is all. Edit: It was brought to my attention that I was not clear here. What I intended is to say that the current convention for full stops is not due to any principle of physics or digital processing.
  • Splitting hairs at 1/3 stop vs. 1/2 stop vs. full stop intervals may be of value in much the same way as a film photographer shooting at a chosen EI rather than box speed. I routinely do so for my film photography. The problem is determining which values (at 1/3 increments) are optimal for your camera and sticking with them. Something tells me that an objective analysis would be difficult. Something also tells me that auto-ISO and/or TAv mode would not be useful with this approach.


Steve


* Perhaps this gives Fuji a certain edge (just kidding)

Last edited by stevebrot; 05-25-2016 at 11:16 AM.
05-25-2016, 10:27 AM   #41
Site Supporter
Alex645's Avatar

Join Date: May 2015
Location: Kaneohe, HI
Photos: Albums
Posts: 1,250
Steve,

Totally agree with your post with two exceptions:

a) "The term (dynamic range?) magically appeared about a year ago...." Or are you referring to ISO invariance? I first read about ISO invariance around 2011. I've been aware of the term dynamic range long before 2011.
b) "There is nothing intrinsically magical about ISO....full stop increments." Until yesterday, I would have agreed with that, but based on what was posted earlier in this thread, there appears to be evidence that is not true to a large degree to certain cameras like the Canon 70D and to a lesser degree other DSLRs as per the previous linked graph:

Photographic Dynamic Range versus ISO Setting

IF I owned a Canon 70D, I could not ignore that my dynamic range at ISO 159 on that chart claims to be worse than ISO 503, and as Simen says or the OP, using full stop ISO is certainly the way to go on that camera.

Again, I really get your other points, but I don't quite understand those two specific statements.
05-25-2016, 10:57 AM   #42
Site Supporter
stevebrot's Avatar

Join Date: Mar 2007
Location: Vancouver (USA)
Photos: Gallery | Albums
Posts: 27,125
QuoteOriginally posted by Alex645 Quote
a) "The term (dynamic range?) magically appeared about a year ago...." Or are you referring to ISO invariance? I first read about ISO invariance around 2011. I've been aware of the term dynamic range long before 2011.
ISO invariance. I first read the term about about a year ago. Before I commented, I did a quick Google search intending to find who coined it, but could not find anything on the first five pages older than 2015. Most of the search results concerned performance of Fuji and Sony product. Apparently the concept is now trendy in much the same way as "total light" and various equivalence topics have been in the past.

QuoteOriginally posted by Alex645 Quote
b) "There is nothing intrinsically magical about ISO....full stop increments." Until yesterday, I would have agreed with that, but based on what was posted earlier in this thread, there appears to be evidence that is not true to a large degree to certain cameras like the Canon 70D and to a lesser degree other DSLRs as per the previous linked graph
I think you may have mis-read my intent. It is an historic accident that the sequence is as it is. We could just as easily be thinking of full stops as 12, 24, 48, 96, 192, 384, etc. If the manufacturers have breakpoints in their algorithms at the historic ASA/ISO film speed marks, it is by convention and not a matter of physics. From the user perspective, there may be a quality reason to choose a particular speed(s) with a particular camera, but to formulate a full-stop rule from a 5-base sequence for digital photography in general, as with the instructor in the original post, is simply superstition and magical thinking.


Steve

(...the 70D allows setting to ISO 159? )

Last edited by stevebrot; 05-25-2016 at 11:16 AM.
05-25-2016, 11:16 AM - 1 Like   #43
Pentaxian




Join Date: Sep 2013
Location: Sydney
Photos: Albums
Posts: 844
QuoteOriginally posted by stevebrot Quote
(...the 70D allows setting to ISO 159? )
Well the judges have to deduct a point because of the 'canon' logo on the front....
05-25-2016, 11:39 AM   #44
Site Supporter
stevebrot's Avatar

Join Date: Mar 2007
Location: Vancouver (USA)
Photos: Gallery | Albums
Posts: 27,125
QuoteOriginally posted by robthebloke Quote
Well the judges have to deduct a point because of the 'canon' logo on the front....


I remember a discussion a few years back where the suggestion was made that comparison reviews should normalize ISO-specific noise evaluation for Nikon products to a different value due to ISO inflation by the brand.


Steve
05-25-2016, 12:39 PM   #45
Moderator PEG Judges
Kerrowdown's Avatar

Join Date: Aug 2008
Location: Highlands of Scotland.
Photos: Gallery
Posts: 32,180
QuoteOriginally posted by rangercarp Quote
ISO: 1/3 or full stops?
I've always worked with everything set to thirds... as I felt it gives me more control, whether it actually does or not is another question, but I feel good about it all.
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!
adjustment, adjustments, analog, camera, dslr, error, ev, exposure, in-camera, iso, noise, photography, pixel, pixels, post, pp, value, values
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
K-1 Full res JPGs versus D810 ISO 12800, ISO 51200 beholder3 Pentax News and Rumors 103 03-23-2016 01:24 PM
Full size high ISO K-3 JPEGS Ayoh Pentax K-3 34 10-19-2013 02:35 AM
1/3 vs 1/2 EV stops cbope Pentax K-5 29 08-17-2012 06:29 AM
does any pentax film camera meter in 1/3 stops? Gooshin Pentax Film SLR Discussion 21 03-14-2009 10:04 PM
1/2 or 1/3 EV stops? Nanthiel Photographic Technique 24 09-03-2008 09:40 PM



All times are GMT -7. The time now is 06:29 PM. | 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