Pages: 1 ... 8 9 [10]   Go Down

Author Topic: Best ISO for IQ280  (Read 52946 times)

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600

As a hint, BClaff indicates that the ISO-s on the X-axis are rated and not measured. So my guess is that ISO on the IQ-250 is a bit like IQ-180 and on the P645+ it is in accordance with the saturation based ISO standard.

Erik, since you are an engineer, I suggest that you would benefit from learning about the ISO:12232(2006) standard and the related CIPA DC-004 standard, because you persist in the total misunderstanding of what ISO SSat measures and what it is not.  For one thing, it is almost completely unrelated to the function of a cameras's "ISO" settings!

DXO measures this, and sees some small variation in the actual exposure index (exposure level suggested by the camera's meter) at a given ISO setting from the EI specified in that ISO setting, but only a little, and not anything close to going all the way down to the SSat minimum EI level.  It is extremely unlikely that Pentax or any camera maker uses that ISO SSat measure of the lower extremity of a camera's exposure latitude to calibrate its ISO exposure index settings, for reasons that I have explained (with references) multiple times in this thread and elsewhere.

For one thing, a camera's "ISO" setting is only relevant to (a) light metering and auto-exposure modes [exposure index], and (b) sensitivity as manifested in the default final JPEG output.  So it has nothing to do with the intermediate processes such as how much gain ("raw levels per photon") is used in producing the raw file data: ISO and CIPA are completely silent on how a camera maker handles the intermediate steps between photo-site signal and JPEG levels.  The extreme of applying equal gain at all ISO settings (or at all settings from 200 up, or all from 400 up) and then scaling up appropriately in the raw-to-JPEG conversion is as consistent with the ISO and CIPA standards as always applying gain in direct proportion to the ISO setting, so that the same raw-to-jpeg mapping is used at all ISO settings.

Further, it would be a total engineering fail to handle high ISO settings with a Sony CMOS sensor with amplification that places the midtones at that SSat based bare minimum of three stops below the maximum raw level: that would risk clipping some highlights that the sensor itself handled fine, and would amplify noise up to far above the raw level spacing, so adding multiple useless bits of noise to the raw levels, and not gaining any noise handling advantage over using less gain.
« Last Edit: July 12, 2015, 04:48:19 am by BJL »
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3366

Erik, since you are an engineer, I suggest that you would benefit from learning about the ISO:12232(2006) standard and the related CIPA DC-004 standard, because you persist in the total misunderstanding of what ISO SSat measures and what it is not.  For one thing, it is almost completely unrelated to the function of a cameras's "ISO" settings!

DXO measures this, and sees some small variation in the actual exposure index (exposure level suggested by the camera's meter) at a given ISO setting from the EI specified in that ISO setting, but only a little, and not anything close to going all the way down to the SSat minimum EI level.  It is extremely unlikely that Pentax or any camera maker uses that ISO SSat measure of the lower extremity of a camera's exposure latitude to calibrate its ISO exposure index settings, for reasons that I have explained (with references) multiple times in this thread and elsewhere.

For one thing, a camera's "ISO" setting is only relevant to (a) light metering and auto-exposure modes [exposure index], and (b) sensitivity as manifested in the default final JPEG output.  So it has nothing to do with the intermediate processes such as how much gain ("raw levels per photon") is used in producing the raw file data: ISO and CIPA are completely silent on how a camera maker handles the intermediate steps between photo-site signal and JPEG levels.  The extreme of applying equal gain at all ISO settings (or at all settings from 200 up, or all from 400 up) and then scaling up appropriately in the raw-to-JPEG conversion is as consistent with the ISO and CIPA standards as always applying gain in direct proportion to the ISO setting, so that the same raw-to-jpeg mapping is used at all ISO settings.

According to your rather condescending analysis, Eric is not the only one totally confused by the Ssat standard. I am in company with him. :-[

I don't have access to the ISO documents, but must rely on the summary of them posted on Wikipedia. As I understand things, the Ssat standard was carried over from the ISO 12232:1998 standard, and it has nothing to do with JPEG output,  sRGB, or the camera light metering. It is defined by the exposure in lux-seconds necessary to result in a given sensor saturation. It is defined as Ssat = 78 lux-sec/Hsat, where Hsat is the exposure necessary to saturate the sensor. The factor 78 was chosen such that exposure settings based on a standard light meter and an 18-percent reflective surface will result in an image with a grey level of 18%/√2 = 12.7% of saturation. The factor √2 indicates that there is half a stop of headroom to deal with specular reflections that would appear brighter than a 100% reflecting white surface. This is standard used by DXO and with which you find fault.

A standard exists so that one may obtain the same results when one uses that standard when working with different cameras. If differing amounts of highlight headroom are allowed, when working with raw files, one would obtain differing amounts of sensor saturation and ETTR exposure would be complicated. This post by Jack Hogan is consistent with my interpretation of the Ssat standard, and I would appreciate your comments.

Regards,

Bill
« Last Edit: July 12, 2015, 09:30:26 am by bjanes »
Logged

dchew

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1017
    • Dave Chew Photography
Best ISO for IQ280
« Reply #182 on: July 12, 2015, 10:14:14 am »

Kudos and many thanks to BJL and others for providing some expertise here.

So despite this thread meandering long and far away from the OP’s original question and the many rants (mine included), I’ve learned some things (I think!).

- When set at 200 ISO, my back shoots the same raw data it does at 35-100. I did not know that, and I don’t know when it started doing that. For all I know it always has. In my defense I can’t remember the last time I seriously used an ISO setting above 100. More on that later…

- “Base ISO” (or I suppose any other camera ISO setting) is not some sacrilegious specification. Camera ISO setting is useful to the photographer for taking pictures, but not for comparing cameras without interpretation. Frankly I never used sensitivity graphs for making a buying decision. Some here think photographers use these ISO graphs for that purpose, which brings me to my rant.

To Edmund and Sunli:
I might agree that on a consumer level, DSLR’s are compared, and perhaps choices are actually made on which DSLR to purchase, based on specs. But I will never believe that is the case with MFD backs, especially those from Phase One.

So I, pretty emphatically, do not agree with these statements:
“ISO 35 is just marketing”
“Phase… understand that cameras are often bought on specs”
“These numbers in the specifications play a significant role for marketing.”
“Phase is getting customers based on this spec (who here apart from some experts would buy an ISO 35 fixed-ISO back)?”


Just look at one of the most important characteristics of their marketing materials. Successful marketing materials need several things and one of the most important is a call to action. After a potential buyer interacts with a brochure, website, ad, salesperson, whatever, what is it precisely you want that potential buyer to do? Do you want them to pick up the phone and call? Write a cheque? Add the product to their shopping cart? Take a test drive? What??

Go to B&H and add that IQ380 to your shopping cart – oh wait, you can’t do that can you? Why do you think that is? Do you really think it is because their dealer network is almighty powerful and won’t let them do it? Nope. If Phase thought they would sell more stuff through those retailers they would. Now think about their marketing strategy: Pricing, distribution network, product feature set, and call to action in their marketing materials.  It is pretty clear that the central call to action is The Demo.

Better yet, attend a PODAS workshop, which has been a reasonably successful marketing program Phase One “sponsors.” You arrive and after the usual introductions and logistics, you are handed a camera. For the next 45 minutes or so you fumble around taking random photos and learning where and what the switches and dials do while technicians run around helping. Then you pack everything up and head out to some fantastic location and start taking photos. From that moment on the whole workshop is about the photographs; it is not about the camera. People are taking photos, processing photos, eating, drinking and discussing photos. Nobody is running around whispering, “This camera is amazing because it will do this at that ISO.”

The marketing slight of hand at PODAS is not about some spec, it is that you are in a beautiful location with great light and could take stunning photos with an iPhone or any other camera. Again, it is about the demo and about the photographs.

BTW, I would buy a fixed ISO 35 camera, and essentially did (and paid a lot for it!). I'm serious that I cannot remember when I've ever used anything higher than ISO 100, and I’ve loved every minute of it. Do I wish it had Sony-sensor dynamic range? Of course! But Sunli, I don’t think saying the IQ180 has great dynamic range is a false statement. It is certainly not as good as other cameras, but almost all cameras have great dynamic range in my opinion. I opened the freezer the other day and pawed through a bunch of film I still have in there to get at something. Only one type of film in there, and it is ISO 50. Did I used to use other film? Sure I did but not very often. So for me, buying a “fixed ISO camera” wasn’t that much different than what I had been doing all along!

I don’t sell cameras, but I strongly believe very few if any of these backs are sold because of some spec in an ad or brochure. The idea that the way Phase uses ISO settings is purely a marketing gimmick, and is therefore the only the reason they still exist (because no sane person would actually want or use their products if they new “the truth”) is 180 degrees away from everything I see in their marketing strategy. And everything I experience with the product.

Dave
« Last Edit: July 12, 2015, 08:21:40 pm by dchew »
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600

Bill, I have no substantial disagreement with your description of the SSat measurement procedure (just one quibble, below).  My complaint is that the intent and appropriate usage of this measurement is not what some people think it is.

Specifically, it is intended (and best used) as a measure of one extreme of the range of suitable exposure levels: the maximum exposure level and thus the minimum exposure index corresponding to a guideline for a minimum advisable amount of exposure headroom -- an amount that is fine in many common situations like uniformly lit scenes, but which can fail badly when the metering is dominated by a shaded subject but there are parts of the scene that are in direct sunlight and are far brighter.

My main claim is that is a bad idea to use this recommended minimum as an ideal or standard, and wrong to say that this is what the ISO standard intends: especially at higher ISO settings, where it increases the risk of blown highlights while giving no IQ advantage in noise or such over somewhat lower raw placements.  No surprise that almost every camera maker using modern high DR sensors allows more than this minimal highlight headroom!

Whatever the virtues of standardization, this extreme value would be a vey poor choice for camera makers to standardize on in light metering (and to repeat, the ISO never intended that!)  Also the standardization (to mid-tones at about 12.5% of maximum raw level) can be and effectively is handled with EXIF info for "level placement intent", which effectively specifies where the mid-tone placement when it deviates from "12.5%".

For photographers who practice careful ETTR placement, SSat (and the DXO measurement) is clearly a useful measurement to know -- just not one that should be embodied in the operation of the ISO dial setting or as a mandate prohibiting a camera maker from providing more than this somewhat minimal amount of highlight headroom it its raw files, to the detriment of anyone wanting to use default JPEG conversions some of the time.

I don't have access to the ISO documents, but must rely on the summary of them posted on Wikipedia.
Me too mostly, but I have also found some other useful discussion by industry experts in slides from seminar presentations and such; I posted them in another thread a long time ago, and will try to find them for you.

As I understand things, the Ssat standard was carried over from the ISO 12232:1998 standard, and it has nothing to do with JPEG output,  sRGB, or the camera light metering. It is defined by the exposure in lux-seconds necessary to result in a given sensor saturation. It is defined as Ssat = 78 lux-sec/Hsat, where Hsat is the exposure necessary to saturate the sensor. The factor 78 was chosen such that exposure settings based on a standard light meter and an 18-percent reflective surface will result in an image with a grey level of 18%/√2 = 12.7% of saturation. The factor √2 indicates that there is half a stop of headroom to deal with specular reflections that would appear brighter than a 100% reflecting white surface. This is standard used by DXO and with which you find fault.
My quibble is that DXO applies this to raw levels after possible amplification, so it is not always measuring sensor saturation, but instead at higher ISO settings is usually measuring where amplifier or ADC clipping occurs.  However, I think that the original objective of ISO:12232(1998) was indeed measuring saturation of the sensor's photo-sites, and this is probably what DXO's measurements still do at the lowest ISO setting, which is the main subject of debate in this thread.  So I would take DXO's measurement at a camera's lowest ISO setting as a measure of sensor saturation, and thus as a reasonable and useful measure of the sensor's base or minimum exposure index -- "base ISO" as it is sometimes called.  (But if instead you want a measure comparable to ASA/ISO film speed, the ISO 40:1 measure is it, as that Wikipedia article you mention explains.)
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Re: Best ISO for IQ280
« Reply #184 on: July 12, 2015, 11:00:48 am »

- When set at 200 ISO, my back shoots the same raw data it does at 35-100. I did not know that, and I don’t know when it started doing that. For all I know it always has. In my defense I can’t remember the last time I seriously used an ISO setting above 100. More on that later…
That seems to be so, and I see what this is puzzling and even bothersome to people who usually choose exposure settings manually based on the histogram or spot metering of highlihgt or some other careful strategy.  The only explanation I can offer is that:

a) These different low ISO exposure index settings (35, 50, 100, 200?) that use the same gain and so give the same raw levels when using the same exposure settings are primarily for convenience when using the in-camera meter.

b) They can also be of value with in-camera review, and when using live view: the same f-stop and shutter speed could be used with (i) ISO 50 with the light meter calling it one stop underexposure), (ii) ISO 100 ("on meter") or (iii) ISO 200 ("one stop overexposure") and give the same raw file, but the live view and rear-screen review would differ in brightness: using the ISO setting for which the light meter agrees with your exposure settings probably gives the most convenient review/preview image.
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3366
Re: Best ISO for IQ280
« Reply #185 on: July 12, 2015, 11:18:16 am »

Go to B&H and add that IQ380 to your shopping cart – oh wait, you can’t do that can you? Why do you think that is? Do you really think it is because their dealer network is almighty powerful and won’t let them do it? Nope. If Phase thought they would sell more stuff through those retailers they would. Now think about their marketing strategy: Pricing, distribution network, product feature set, and call to action in their marketing materials.  It is pretty clear that the central call to action is The Demo.

Another take on this: you can't buy a Phase One camera from B&H for the same reason you can't buy a Rolex at Walmart. One does not buy a Rolex on specs. A $200 Seiko has better specs than a $5000 Rolex. The electronic movement of the Seiko is much more accurate than the mechanical movement of the Rolex, but people with the money buy the Rolex for prestige and pride of ownership. If you are at a business meeting or cocktail party with a Rolex on your wrist, that advertises that you have arrived and are among the elite and would not deign to buy a cheap watch from Walmart, the store that caters to the unwashed masses. You would be shopping at Tiffany's.

Admittedly, an IQ 180 does have some advantages over the Nikon D810, and the parallels of this analogy are not exact. However, the price:performance of the Phase over the Nikon is not favorable.

Bill
Logged

DucatiTerminator

  • Newbie
  • *
  • Offline Offline
  • Posts: 46
Re: Best ISO for IQ280
« Reply #186 on: July 12, 2015, 12:08:57 pm »

Assuming the D810 is the Seiko, and the Phase a Rolex, you still can't buy a D810 at Walmart either. ;)

Alvin


Sent from my iPhone using Tapatalk
Logged
La gallina vecchia fa buon brodo

Paul2660

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3978
    • Photos of Arkansas
Re: Best ISO for IQ280
« Reply #187 on: July 12, 2015, 12:10:36 pm »

Another take on this: you can't buy a Phase One camera from B&H for the same reason you can't buy a Rolex at Walmart. One does not buy a Rolex on specs. A $200 Seiko has better specs than a $5000 Rolex. The electronic movement of the Seiko is much more accurate than the mechanical movement of the Rolex, but people with the money buy the Rolex for prestige and pride of ownership. If you are at a business meeting or cocktail party with a Rolex on your wrist, that advertises that you have arrived and are among the elite and would not deign to buy a cheap watch from Walmart, the store that caters to the unwashed masses. You would be shopping at Tiffany's.

Admittedly, an IQ 180 does have some advantages over the Nikon D810, and the parallels of this analogy are not exact. However, the price:performance of the Phase over the Nikon is not favorable.

Bill

Hi Bill,

Good points.  Back in 2008, it was easy for me to make the Phase One decision, as they were the only "affordable" solution for 39MP, single frame non-uprezed images.  The 60MP and 80MP backs still hold this especially the 80MP as far as the kings of resolution.  I personally have yet to find any software uprez solution that can get close the the native 60 or 80MP resolution from a Phase One.  Do I need 100MP, no can't justify that, not now with my current business.  To me improvements that the 35mm DSLR cameras have made since 2008 are very impressive and I feel that we will continue to see such price performance.

To the folks that have really applied the technical knowledge to this post, I do want to thank you.  I rarely print out a post but this one I did.  A ton of good information and I do appreciate the time it took for you to post it.  I have been trying to get my hands around the ISO settings of the Phase One backs, and never found a good technical doc.  But this post has it in spades and backs it up with a lot of great graphs and images.  

It will help me in the future with my photography with the Phase backs I use.  

Thanks
Paul Caldwell
Logged
Paul Caldwell
Little Rock, Arkansas U.S.
www.photosofarkansas.com

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3366
Re: Best ISO for IQ280
« Reply #188 on: July 12, 2015, 12:38:03 pm »

Assuming the D810 is the Seiko, and the Phase a Rolex, you still can't buy a D810 at Walmart either. ;)

Alvin


Sent from my iPhone using Tapatalk

Really? They (Walmart) advertise the D810 online for $2996.95. It would be shipped from OneCall, but it is still available by ordering from Walmart. However, you are mis-stating the intent of my analogy, where B&H is the mass-market option corresponding to Walmart and Tiffany is the exclusive retailer corresponding to Capture Integration or other Phase One dealers. I don't think you can buy an IQ180 from Tiffany.  :)

Bill
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3366

Bill, I have no substantial disagreement with your description of the SSat measurement procedure (just one quibble, below).  My complaint is that the intent and appropriate usage of this measurement is not what some people think it is.

Specifically, it is intended (and best used) as a measure of one extreme of the range of suitable exposure levels: the maximum exposure level and thus the minimum exposure index corresponding to a guideline for a minimum advisable amount of exposure headroom -- an amount that is fine in many common situations like uniformly lit scenes, but which can fail badly when the metering is dominated by a shaded subject but there are parts of the scene that are in direct sunlight and are far brighter.

My main claim is that is a bad idea to use this recommended minimum as an ideal or standard, and wrong to say that this is what the ISO standard intends: especially at higher ISO settings, where it increases the risk of blown highlights while giving no IQ advantage in noise or such over somewhat lower raw placements.  No surprise that almost every camera maker using modern high DR sensors allows more than this minimal highlight headroom!

Whatever the virtues of standardization, this extreme value would be a vey poor choice for camera makers to standardize on in light metering (and to repeat, the ISO never intended that!)  Also the standardization (to mid-tones at about 12.5% of maximum raw level) can be and effectively is handled with EXIF info for "level placement intent", which effectively specifies where the mid-tone placement when it deviates from "12.5%".

For photographers who practice careful ETTR placement, SSat (and the DXO measurement) is clearly a useful measurement to know -- just not one that should be embodied in the operation of the ISO dial setting or as a mandate prohibiting a camera maker from providing more than this somewhat minimal amount of highlight headroom it its raw files, to the detriment of anyone wanting to use default JPEG conversions some of the time.
Me too mostly, but I have also found some other useful discussion by industry experts in slides from seminar presentations and such; I posted them in another thread a long time ago, and will try to find them for you.
My quibble is that DXO applies this to raw levels after possible amplification, so it is not always measuring sensor saturation, but instead at higher ISO settings is usually measuring where amplifier or ADC clipping occurs.  However, I think that the original objective of ISO:12232(1998) was indeed measuring saturation of the sensor's photo-sites, and this is probably what DXO's measurements still do at the lowest ISO setting, which is the main subject of debate in this thread.  So I would take DXO's measurement at a camera's lowest ISO setting as a measure of sensor saturation, and thus as a reasonable and useful measure of the sensor's base or minimum exposure index -- "base ISO" as it is sometimes called.  (But if instead you want a measure comparable to ASA/ISO film speed, the ISO 40:1 measure is it, as that Wikipedia article you mention explains.)

Thanks for the clarification and additional information. DXO does not publish their raw data and how they obtain some of their results is not defined publicly, and
the "ISOs" above base ISO obviously do not measure sensor saturation but merely clipping in the ADC. Setting the ISO on the camera to a value above base does not change the sensor sensitivity.

In a previous exchange with you I did a careful reevaluation of my D800e at base ISO of 100. Since ISO settings are usually in increments of 0.33 there will well be some rounding error. The real base ISO may be 100 ± 0.33 EV steps. I do not have a calibrated light source and must assume that the camera meter is according to ISO 2720 with a Κ of 12.5, which I understand is the case. A target exposed according to the light meter with the camera set at ISO 100 results in 10.3% sensor saturation, which would allow 0.8 stops of highlight headroom as compared to the 0.5 EV specified in the Ssat standard. The EXIF indicated that the REI standard was employed, so Nikon can use any headroom allowance the desire, just as Phase One can use 1.5 EV without any trickery or dishonesty. IMHO, the REI standard is not a standard at all.

DXO claim that they use 78 lux-sec in their calibration, but is this really so? For the D800e they rate measured ISO at 73 with a manufacturer value of 100. The difference is 0.45 EV, which would be on top of the already allowed 0.5 EV for a total of 0.95 EV. What numerator did they use for the ISO calculation?

Regards,

Bill
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Re: Best ISO for IQ280
« Reply #190 on: July 12, 2015, 01:31:04 pm »

Bill, just two comments:

1) Yes, the ISO and CIP standards actually specify that all values should be rounded to numbers in the standard sequence . . . 25, 32, 40, 50, 64, . . . (so that for example the minimum ISO setting of 32 for the IQ280 is exactly what is prescribed for the exact value of 29 reported by DXO.)

2) REI sounds very slippery, but there is a clause in there that seems to require it to be the same as SOS when the camera is in a "traditional (i.e. simple)" metering mode like center-weighted.  So the flexibility in REI might only be for "non-traditional" modes like multi-zone metering, for which SOS seems inapplicable.

So my hopeful guess is that any camera like the D800e that has both "traditional" and "non-traditional" metering modes does use SOS for the former, but has to describe its method as REI due to doing things differently in those exotic modes.
« Last Edit: July 12, 2015, 04:22:59 pm by BJL »
Logged

Ken R

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 849
Re: Best ISO for IQ280
« Reply #191 on: July 12, 2015, 10:10:31 pm »

This guy explains it pretty well I think:

HERE
Logged

DucatiTerminator

  • Newbie
  • *
  • Offline Offline
  • Posts: 46
Re: Best ISO for IQ280
« Reply #192 on: July 12, 2015, 10:58:49 pm »


Really? They (Walmart) advertise the D810 online for $2996.95. It would be shipped from OneCall, but it is still available by ordering from Walmart. However, you are mis-stating the intent of my analogy, where B&H is the mass-market option corresponding to Walmart and Tiffany is the exclusive retailer corresponding to Capture Integration or other Phase One dealers. I don't think you can buy an IQ180 from Tiffany.  :)

Bill

Relax, Brotha, it was meant as a joke, hence the emoticon. You can buy almost anything at rakuten.com, but most items are fulfilled by other vendors/dealers. I suspect this is not dissimilar to items Walmart does not fulfill themselves.

That said, no misstatement intended.. I really didn't read too much into your analogy or I would have pointed out that there is a much larger disparity between Walmart and Tiffany merchandise/brands vs. that of B&H and CI/DT/etc.

I apologize if I misunderstood your reply, but I was sent to this forum to learn from some very skilled and artistic knowledgeable photo enthusiasts. So far it seems that there are a lot of egos hellbent on proving themselves right.

Alvin


Sent from my iPhone using Tapatalk
Logged
La gallina vecchia fa buon brodo

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11288
    • Echophoto
Re: Best ISO for IQ280
« Reply #193 on: July 13, 2015, 01:19:07 am »

Hi,

Yes, those articles are excellent. Pretty much of my initial understanding is coming from that.

This is another article often referred to, which I feel is more clear: http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/

Much of the later discussion relates to different ISO standards. Saturation based ISO is a solid standard, but there are a lot of variants not based on sensor saturation but on acceptable noise levels in JPEG or TIFF images.

The problem with those standards is that they give a lot of leeway for setting quite arbitrary values. For instance, this link compares the Pentax 645Z with the Phase One IQ-250 : http://home.comcast.net/~NikonD70/Charts/PDR.htm#Pentax%20645Z,Phase%20One%20IQ250 , a screen dump is attached.

Now, both cameras use the same sensor but show very different ISO dependency. The sensors use internal on chip raw conversion and there is probably very little difference between the two. The two curves meet at minimum ISO. This essentially means that Phase One uses a different ISO with additional highlight protection, while Pentax probably is very close to saturation based ISO. Also note that Phase One stops at 6400 ISO while Pentax goes up to 256000 ISO. How can that be?

Some consideration to put things in perspective:

Signal / Noise is essentially the square root of captured photons. If we assume 12.5% reflections exposure would be at least three stops below saturation.

If we assume that each pixel can hold 65000 electron charge, corresponding to 65000 photons (or so):

SNR = sqrt(65000 * 0.125) = 90

Now, add 1.5 stops of highlight protection:

SNR = srqt(6500 * 0.125 * 0.35) = 53 which still is a very good value.

The different JPEG/TIFF based ISO standards see to use SNR 40 as a limit.

Another way to see it is that the IQ-250 has 1.68 times the area of a 24x36 sensor, so it can give same SNR as a 24x36 sensor with 0.75 EV less exposure.

Saturation based ISO gives a lot of usable information for determining exposure. A more lax implementation/interpretation of ISO makes it much less usable.

Best regards
Erik


This guy explains it pretty well I think:

HERE
Logged
Erik Kaffehr
 

voidshatter

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 400
    • www.500px.com/voidshatter
Re: Best ISO for IQ280
« Reply #194 on: July 13, 2015, 02:20:55 am »

Hi Erik,

The IQ250 is 16-bit while the 645Z is 14-bit. Could it be the reason why the IQ250 can implement the extra highlight headroom at ISO 200?
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Roger Clark on what the ISO setting on a camera does, and "ISOless range"
« Reply #195 on: July 13, 2015, 06:10:30 am »

This guy explains it pretty well I think:

HERE
Yes, Roger Clark is a good source, and that looks like a good discussion of what the ISO setting on a camera does (adjusts gain, in conversion to raw and/or later in conversion to JPEG).

Some small quibbles with his Myth #1:
- the ISO setting does affect what the ISO and CIPA standards call "Output Sensitivity", which is the ratio between how much exposure the sensor gets and the final (JPEG?) output levels.
- the ISO setting can effect what I will call the overall "system sensitivity" in the sense of the ISO 40:1 or 10:1 measures related to SNR at a given signal level: this happens in what Clark calls the "camera electronics limited" lower ISO settings.
But I agree with his intended point that the ISO setting does not effect the sensitivity of the photo-sites themselves.


I would like Roger Clark to get or test some non-Canon cameras!  I would be interested to see his analysis of them.
Logged

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: Best ISO for IQ280
« Reply #196 on: July 13, 2015, 08:23:41 am »

Maybe using a single ISO number to characterize complex behaviour is not so smart

Edmund
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Re: Best ISO for IQ280
« Reply #197 on: July 13, 2015, 09:18:25 am »

Maybe using a single ISO number to characterize complex behaviour is not so smart.
Agreed: maybe that's why the ISO standard 12232(2006) defines at least four "numbers", measuring different aspects of sensor and camera performance.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11288
    • Echophoto
Re: Best ISO for IQ280
« Reply #198 on: July 13, 2015, 12:56:39 pm »

Hi,

Here in Sweden I can buy both Phase One and Hasselblad on mail order.

Scandinavian Photo is like B&H : https://www.scandinavianphoto.se/kategori/1008167349/mellanformat/digitalt-bakstycke#brand=551,5661440&filter=&category=1008167349&page=0

Goecker used to be a store for professionals: http://webshop.goecker.se/web/guest/shop/-/schcategory/id/13197

Best regards
Erik


Really? They (Walmart) advertise the D810 online for $2996.95. It would be shipped from OneCall, but it is still available by ordering from Walmart. However, you are mis-stating the intent of my analogy, where B&H is the mass-market option corresponding to Walmart and Tiffany is the exclusive retailer corresponding to Capture Integration or other Phase One dealers. I don't think you can buy an IQ180 from Tiffany.  :)

Bill
Logged
Erik Kaffehr
 
Pages: 1 ... 8 9 [10]   Go Up