Why use iso 800

[David Mullen ASC]

From ARRI's website FAQ:

Does ALEXAs exposure latitude change with different Exposure Index (EI) settings?

 

No, it does not. Most digital cameras have the greatest exposure latitude at a specific EI setting (often called the 'sweet spot'). Choosing an EI setting that is higher or lower than the optimal setting will result in a sometimes surprisingly large loss of exposure latitude. ALEXA is unique in that its exposure latitude of over 14 stops (as measured with the ARRI Dynamic Range Test Chart (DRTC)) stays constant from EI 160 to EI 3200.

As I suggested before, the reason many other cameras do change dynamic range at different ISO's is that they are trying to hide problems at the noise floor of the signal. And you'll find that the cleaner the signal from the sensor is in general, as part of the design, the wider the dynamic range because there are more usable stops at the bottom of the signal.

Digital sensors tend to have a hard clip at the top of exposure when the sensor photosite can't handle any more light/signal, it's just overloaded, but at the bottom, the problem is that detail just starts to fall into the noise floor making it unusable, so it's often the noise in the system that limits the dynamic range to some extent.

[David Mullen ASC]

 

So, ultimately, higher ISO = more noise and not

 

 

I guess reading about still photography, I always thought that you use higher ISOs in low-light situations and that the image is noisier when the ISO is higher not because of lower amount of light, but because of high ISO. Come to think of it, I even think that when they do a review of a camera, they shoot in not that lower a light and the image at high ISO levels is way noisier, though the noise gets lower with every new camera model.

 

It somehow really messed with a lot of what I thought I knew when you said that low ISO in daylight situations is a no-no, because the highlights will get blown out, when up to this point I thought the opposite would happen: the camera would be less sensitive to light and will not overblow the highlights. Perhaps it’s a still-camera thing and film cameras work differently. :ph34r: Though that shouldn’t be the case.

 

The sensor has a basic sensitivity that some call the "native" sensitivity -- the problem is more how you define the optimal exposure to determine this native sensitivity as an ISO value. Do you base it on where middle grey lands with an equal amount of highlight and shadow information? Truth is that all you can really measure is the sensor's total dynamic range and it's up to the manufacturer or you to decide which ISO works best in that range. I mean, when you use a Canon 5D, for example, do you know what the "native" ISO is for that sensor? Or do you just use at at different ISO ratings until you find one that works best for you?

 

A raw recording or signal is not a viewable color image, so it has to be processed into RGB from its original bayer-filtered signal. It's in the conversion to color that you can process the signal to different brightness levels that are labeled "ISO" settings, so if you chose a high ISO setting and expose, let's say, a grey card so that it looks normally exposed for that ISO setting according to your meter, in actuality you are giving the sensor less light than if you chose a low ISO setting. So when the raw signal is converted to color, the processor brings up the values to the level that they are correct for the ISO selected, so in a sense, the signal is being boosted when the sensor was underexposed so that the image looks bright enough for your high ISO setting.

 

The ISO setting is just processing -- you have more noise because you are "pushing" a weaker signal because the sensor was underexposed. So it's the amount of light that hits the sensor that determines the noise IF you have to process the signal to look "normal" in brightness (I mean, you could process the raw file to a lower ISO and the underexposed sensor would just give you a darker image, and thus it wouldn't get noisy until you tried to brighten it.)

 

You don't lose more highlight information at low ISO settings IF your camera's sensor naturally has a low sensitivity. For example, with many full-frame still cameras like the Canon 5D, the Nikon D600, etc. the maximum dynamic range and thus probably the "native" sensitivity of the sensor is rather low, like around 200 ISO. It's just that the signal is so clean that it can be rated at very high ISO's with acceptable noise.

 

I find that with my Nikon D600, for example, that it is better to shoot in raw mode at a low ISO (like 200 ISO) and underexpose because otherwise it tends to lose highlight detail too quickly but it has a lot of shadow detail that is usable at the low ISO rating.

 

This article on exposure vs. brightness is useful:

http://www.dpreview.com/articles/8148042898/exposure-vs-brightening

[Tyler Purcell]

Well yes David. You are right that the over-all images dynamic range as a total number, doesn't "decrease" per say. However, the "effect" of going over or under the base ISO, IS a detractor to the over-all image. Lower the ISO and you have less highlight dynamic range. Increase the ISO and you have less black dynamic range.

 

So where the total number (lets say 14 stops) doesn't change. There is a significant difference in the imagers ability to capture the same dynamics as it's base ISO.

 

http://www.provideocoalition.com/alexa_iso_settings_the_least_you_need_to_know/

 

Still, that's the Alexa... the most advanced cinema camera on the market. In reality, most cameras and imagers don't function anywhere near the Alexa.

[David Mullen ASC]

This is where the real world meets with the specs on paper, because outside in sunlight in an open field, you are more likely to encounter extreme highlights, particularly if the sky is in the shot, wheres in a forest with patches of sunlight, it might be reversed, shadows can plunge many stops into darkness. And course your subject may be someone in a white outfit or a black outfit, etc. So you have to play it by ear (or eye) in terms of whether you'd be better off with a higher or lower ISO rating on the Alexa outdoors in sunlight.

 

But to some degree, one's sensitivity to clipping versus noise is personal, some have more of an issue with one than the other.

[Tyler Purcell]

Exactly, tests only say so much, in reality things can be slightly different.

 

The comments you made about filtration as well are so important. A lot of people don't realize what happens to an image with over filtration.

[Jan Tore Soerensen]

 

 

Hi Jan.. In this case you are "over exposing" on purpose.. to get a "cleaner" image when corrected in post.. i.e. simply letting more light fall onto your sensor..

Yeah, I just stated a fact that it didn't seem like he knew.

Edited May 31, 2016 by Jan Tore Soerensen

[Robin R Probyn]

Yeah, I just stated a fact that it didn't seem like he knew.

 

 

Ok yes sorry I see what you mean.. apologies my mistake sir..

[Robin R Probyn]

 

Up to this point I thought Robin was a woman. :ph34r:

 

 

haha no.. like Batman and Robin.. but I wear red tights ..

[Bruce Greene]

On my Canon 5d, when I raise the ISO setting, it seems that the RAW signal is amplified before recording it. This results in lower dynamic range as I raise the ISO setting and the highlights begin to clip. On an Arri Alexa, or RED camera, the RAW data is not changed as the ISO setting is raised, just the meta data and the preview are changed, preserving all the highlight data. Perhaps Canon believes that this enables them to minimize noise as one raises the ISO setting. To get around this, one needs to expose at a lower ISO setting, such as using the exposure for ISO 400 when setting the camera to ISO 200 and increasing the brightness in the RAW processor in post. As David suggested earlier.

 

So yes, the strategy of exposing RAW stills on a Canon 5d are different than when exposing on a Arri Alexa camera regarding ISO settings.

 

Just to complicate things, the Varicam 35 has two ISO settings: ISO 800 and ISO 5000. Each has a similar dynamic range. On this camera all other ISO settings consist of changes in brightness (or amplification if you will) to the preview and work similarly to changes in ISO on an Alexa.

[Alexandros Angelopoulos Apostolos]

Here is what I was talking about, the subsection ISO & Noise Performance:

 

http://www.imaging-resource.com/PRODS/nikon-d810/nikon-d810A5.HTM

 

I believe this is the same phenomenon Bruce Greene is talking about.

[John E Clark]

 

So, David, what happens at ISO 1600, which is one full stop above ISO 800? How much stops of dynamic range do you have for shadows and how much for highlights?

 

 

This chart is from the ARRI site to depict the difference in ISO settings and the amount of 'detail' above and below the middle grey value.

 

It is different for different manufacturers... but there are similar results.

 

[John E Clark]

 

I find that with my Nikon D600, for example, that it is better to shoot in raw mode at a low ISO (like 200 ISO) and underexpose because otherwise it tends to lose highlight detail too quickly but it has a lot of shadow detail that is usable at the low ISO rating.

 

 

The Wife (D600) and I (D7300) shot the Daughter's wedding in San Francisco City Hall earlier this year. Fortunately we did shoot RAW+JPEG, and the JPEGs were pretty useless as the highlights were pretty blocked up. The RAW processed via Lightroom looks great...

Edited May 31, 2016 by John E Clark

[David Mullen ASC]

Here is what I was talking about, the subsection ISO & Noise Performance:

 

http://www.imaging-resource.com/PRODS/nikon-d810/nikon-d810A5.HTM

 

I believe this is the same phenomenon Bruce Greene is talking about.

 

 

The raw section further down is more informative. Again, whenever you are talking about conversion to JPEG, you have a lot of image processing going on.

 

The raw section points out what Bruce is talking about, which is that some camera's raw is more processed (less raw) than others. ISO is clearly being "baked" (through signal amplification) into the raw recording with these DSLR's compared to raw recording on a Red camera or the Alexa. And with the Alexa, the Arriraw files are 12-bit log files even though they haven't been converted to RGB (raw on the Sony F65, for example, is 16-bit linear data).

 

The article also shows what I mentioned, that the actual native or base sensitivity of some of these high pixel count sensors is on the low side, so the only reason you are able to shoot at very high ISO ratings is that the noise is low enough to make underexposed information usable.

[cole t parzenn]

 

From ARRI's website FAQ:

Does ALEXAs exposure latitude change with different Exposure Index (EI) settings?

 

No, it does not. Most digital cameras have the greatest exposure latitude at a specific EI setting (often called the 'sweet spot'). Choosing an EI setting that is higher or lower than the optimal setting will result in a sometimes surprisingly large loss of exposure latitude. ALEXA is unique in that its exposure latitude of over 14 stops (as measured with the ARRI Dynamic Range Test Chart (DRTC)) stays constant from EI 160 to EI 3200.

As I suggested before, the reason many other cameras do change dynamic range at different ISO's is that they are trying to hide problems at the noise floor of the signal. And you'll find that the cleaner the signal from the sensor is in general, as part of the design, the wider the dynamic range because there are more usable stops at the bottom of the signal.

Digital sensors tend to have a hard clip at the top of exposure when the sensor photosite can't handle any more light/signal, it's just overloaded, but at the bottom, the problem is that detail just starts to fall into the noise floor making it unusable, so it's often the noise in the system that limits the dynamic range to some extent.

 

 

 

Could it be that Alexa footage just looks cleaner, since we're almost always seeing it downsampled? That would be unique, I was led to believe.

[David Mullen ASC]

It probably helps a little but many digital cameras have even more pixels than the Alexa's sensor does so are downscaled even more to get to 2K or HD... but don't necessarily look cleaner.

 

It's a tradeoff because more pixels in the same sized sensor mean smaller photosites and less sensitivity, and thus more noise in low light, and yet more pixels also means there is more downscaling and you can use more noise reduction if you are going to end up with a smaller sized final version.

[Robin R Probyn]

"Could it be that Alexa footage just looks cleaner, since we're almost always seeing it downsampled? That would be unique, I was led to believe."

 

​Definitely not unique .. even the lowly Fs7 .. about the price of an Alexa VF cable... down samples from a "full" 4K sensor .. as does the F5/55

 

As far as i know . ISO is like gain in ENG camera,s... native ISO is 0 db gain.. no amplification of the signal.. high ISO is not part of the exposure.. its just applying electronic amplification to the signal.. and will add noise.. the camera,s noise level will determine how much you can get away with..

Edited June 1, 2016 by Robin R Probyn

[Alexandros Angelopoulos Apostolos]

That’s the thing: I am only now finding out that it’s a gain or something far less precise than I thought it would be. I thought it was a very precise physical quantity, expressed, perhaps, via the application of the photoelectric effect: a light falls onto a sensor and the sensor measures the intensity of the current. The sensitivity of the sensor to light is expressed in ISO, either as a direct or calculated quantity. But now I find out it’s not even the same method that the various camera makers use to determine this. :ph34r:

 

I’m glad Tyler mentioned a potentiometer. :D :lol:

[Robin R Probyn]

Its the same as applying gain to an audio signal.. the volume goes up .. and the noise /hiss.. its not changing the actual recording levels.. just amplifying them ..but alot of camera,s these days you can push the ISO(gain) alot and still have an acceptable picture..

2/3 inch 3CCD camera,s no one would ever want to go above +6 (2 stops) .. +9 for alien landing shots only..

[Robin R Probyn]

Changing the ISO is not changing the sensitivity of the camera.. to do that you would have to change the sensor ..!

[John E Clark]

That’s the thing: I am only now finding out that it’s a gain or something far less precise than I thought it would be. I thought it was a very precise physical quantity, expressed, perhaps, via the application of the photoelectric effect: a light falls onto a sensor and the sensor measures the intensity of the current. The sensitivity of the sensor to light is expressed in ISO, either as a direct or calculated quantity. But now I find out it’s not even the same method that the various camera makers use to determine this. :ph34r:

 

I’m glad Tyler mentioned a potentiometer. :D :lol:

 

I believe the current ISO Standard for digital cameras is ISO 12232:2006, and is listed as 88 CHF from the ISO site directly (My exchange calculator indicate's that's about $89...) so, I'm not going to be quoting directly from that document...

 

But... the following excerpt is part of a review found here:

 

http://www.dpreview.com/forums/thread/2838786

 

 

The 'bottom line' is manufacturers can pretty much give any number based on their 'evaluation' of image quality. There are more quantitative tests... but that may cause the ISO values for a camera to be less than desired for marketing purposes...

 

This leads to pretty much a lazy faire for the buyer to beware... As noted in the text below, some of the standards relate only to JPEG encoding or sRGB representation, and do not include RAW or other types of processed (or nonprocessed) image data...

 

I suspect that camera producers like ARRI, who are trying to make the transition from Film film to Digital film as 'easy as possible' will give Digital Exposure Indices which approximately match that of Film film. Since I'm not privy to the engineering of the various companies... this is just a conjecture...

 

But I will say, when I was at NAB a few years ago, and asked at the Big Name Brand DSLR manufacturers, no one at the booth could answer the question of 'how is the ISO values of your camera determined' (most often I was shown how to 'set' an ISO value for a given camera). In several cases I was referred to Japan for such an engineering question.

 

Then there was Black Magic, where I was able to talk to the lead of their sensor section... and we did have a good conversation about the difficulties of picking a number for exposure in the modern digital world.

 

So for now, I use the following method to determine my personal EI for my cameras.

 

1) For 'log' type representation, such as the Black Magic Pocket in RAW mode, use the following

 

Take a reading of a 18% grey card, and then evaluate the resulting image waveform display. Adjust the EI value to achieve a 38.4% IRE(Blackmagic specific other manufacturers products may vary...) reading.

 

2) For Rec. 709 or similar. Use 45-50% IRE for the 18% grey card.

 

Use a step wedge chart to determine the number of stops above and below the middle grey value that can be distinguished, yielding the effective dynamic range of the camera at that EI setting.

 

 

 

 

---- The quoted material from the standard, as included in the above referenced forum post...

 

The ISO 12232:2006 standard

 

The ISO standard 12232:2006[15] gives digital still camera manufacturers a choice of five different techniques for determining the exposure index rating at each sensitivity setting provided by a particular camera model. Three of the techniques in ISO 12232:2006 are carried over from the 1998 version of the standard, while two new techniques allowing for measurement of JPEG output files are introduced from CIPA DC-004.[16] Depending on the technique selected, the exposure index rating can depend on the sensor sensitivity, the sensor noise, and the appearance of the resulting image. The standard specifies the measurement of light sensitivity of the entire digital camera system and not of individual components such as digital sensors, although Kodak has reported[17] using a variation to characterize the sensitivity of two of their sensors in 2001.

 

The Recommended Exposure Index (REI) technique, new in the 2006 version of the standard, allows the manufacturer to specify a camera model’s EI choices arbitrarily . The choices are based solely on the manufacturer’s opinion of what EI values produce well-exposed sRGB images at the various sensor sensitivity settings. This is the only technique available under the standard for output formats that are not in the sRGB color space. This is also the only technique available under the standard when multi-zone metering (also called pattern metering) is used.

 

The Standard Output Specification (SOS) technique , also new in the 2006 version of the standard, effectively specifies that the average level in the sRGB image must be 18% gray plus or minus 1/3 stop when exposed per the EI with no exposure compensation . Because the output level is measured in the sRGB output from the camera, it is only applicable to sRGB images—typically JPEG—and not to output files in raw image format. It is not applicable when multi-zone metering is used.

 

The CIPA DC-004 standard requires that Japanese manufacturers of digital still cameras use either the REI or SOS techniques. Consequently, the three EI techniques carried over from ISO 12232:1998 are not widely used in recent camera models (approximately 2007 and later) . As those earlier techniques did not allow for measurement from images produced with lossy compression, they cannot be used at all on cameras that produce images only in JPEG format.

 

The saturation-based technique is closely related to the SOS technique, with the sRGB output level being measured at 100% white rather than 18% gray. The saturation-based value is effectively 0.704 times the SOS value .[18] Because the output level is measured in the sRGB output from the camera, it is only applicable to sRGB images—typically TIFF—and not to output files in raw image format. It is not applicable when multi-zone metering is used.

 

The two noise-based techniques have rarely been used for consumer digital still cameras. These techniques specify the highest EI that can be used while still providing either an “excellent” picture or a “usable” picture depending on the technique chosen.

---

Edited June 2, 2016 by John E Clark

[Alexandros Angelopoulos Apostolos]

Its the same as applying gain to an audio signal.. the volume goes up .. and the noise /hiss.. its not changing the actual recording levels.. just amplifying them ..but alot of camera,s these days you can push the ISO(gain) alot and still have an acceptable picture..

2/3 inch 3CCD camera,s no one would ever want to go above +6 (2 stops) .. +9 for alien landing shots only..

 

:)

 

Changing the ISO is not changing the sensitivity of the camera.. to do that you would have to change the sensor ..!

 

Yep, yep. I got it. :) I got it on the previous two pages. When I wrote “The sensitivity of the sensor to light is expressed in ISO, either as a direct or calculated quantity.” it was a close follow-up of the previous line. It was meant to state how I thought things worked. But then I destroyed the statements from that line and the one preceding it with the one stating that manufacturers don’t even have the same method of determining ISO levels.

 

 

I believe the current ISO Standard for digital cameras is ISO 12232:2006, and is listed as 88 CHF from the ISO site directly (My exchange calculator indicate's that's about $89...) so, I'm not going to be quoting directly from that document...

 

But... the following excerpt is part of a review found here:

 

http://www.dpreview.com/forums/thread/2838786

 

 

The 'bottom line' is manufacturers can pretty much give any number based on their 'evaluation' of image quality. There are more quantitative tests... but that may cause the ISO values for a camera to be less than desired for marketing purposes...

 

Yep. That’s exactly what I meant when I stated that there are several, I’d say baffling (but, I guess, that’s the best we can do now), methods according to which manufacturers can determine their ISO levels in a camera. I read about exactly the same things you pointed out to in your nice post, the ISO standard and the five methods.

 

Thank you so much for posting about your personal method of determining EI! :) Makes so much sense. They’re a great thing to know.

[Tom Yanowitz]

gobbledygook = ˈɡɒb(ə)ldɪˌɡuːk,-ˌɡʊk/

nouninformal

language that is meaningless or is made unintelligible by excessive use of technical terms.

example : 99% of topics dealing with ISO in the digital world are utter gobbledygook.

ISO is an index that is supposed to help us link a camera sensitivity and a given exposure. It has no units. It has more to do with the DP's tastes than the camera.

It is VERY different from the camera sensitivity itself, which can be quantified with precise units (electrons per lux.s, volts per lux.s...)

There isn't any definitive standard for ISO currently. The most popular one being the REI, which is the contrary of a standard really (it allows each manufacturer to determine ISO numbers how the please).

Hence there certainly isn't any native ISO for any camera.

What do people exactly expect to be carried in the word 'native' anyway.

The optimal ISO depends on the scene, it's light level, it's dynamic range, how clean you want your shadows etc

More on all of this in my controversial topic How the alexa really works.

http://www.cinematography.com/index.php?showtopic=70643

[John E Clark]

 

What do people exactly expect to be carried in the word 'native' anyway.

The optimal ISO depends on the scene, it's light level, it's dynamic range, how clean you want your shadows etc

 

 

I don't use the term 'native' much, but how I interpret that is some what like the basic 'manufacturer's recommended ISO' of Film film. Kodak rated Tri-X still film at ASA/ISO 400... I shot it at 200 with appropriate development. At the time that I was worried about such things I also used densitometers on the negative, and printed any number of test strips to find my 'personal' ASA/ISO/EI value... which was determined to be 200...

 

When confronted with any new camera I pretty much do the same... using digital tools these days.

 

In the Film film days, one was somewhat required to use a 'fixed' ISO... within some limits... one could push or pull processing to get effective changes in the response of the negative to the exposure, but there were some side effects that had to be accepted... push... more grain... I tended not to pull on processing, since that often lead to processing not being uniform on the negative, given that I hand processed my still film...

 

These days, with the variable ISO there is sort of a tendency to just make that adjustment on a per scene basis... but one can also say that the discipline of using a fixed ISO is better... and just light appropriately...

 

On the other hand when shooting with more available lighting or 'low power' lighting, ISO adjustment may be required.

[Thad Donovan]

I’m looking for help in understanding something about native ISO. I’ve seen native ISO described as “the highest ISO before the signal is amplified” (or something close to that). If that’s the case, then wouldn’t a high native ISO be preferred?

For example, the Blackmagic Ursa Mini G2 shows a native ISO of 3200. The Arri apparently has native ISO of 800. Does that mean that the Arri amplifies the signal at 1600 ISO (for example) and therefore is not as desirable from a noise perspective as a higher native ISO? Whereas the BMUMPG2 would NOT amplify the signal at 1600 ISO, so is thereby cleaner in that regard.

Something tells me I’m wrong on that, but I just don’t know why.

Thanks for any clarifications!

 

[Tyler Purcell]

25 minutes ago, Thad Donovan said:

I’m looking for help in understanding something about native ISO. I’ve seen native ISO described as “the highest ISO before the signal is amplified” (or something close to that). If that’s the case, then wouldn’t a high native ISO be preferred?

From my understanding, the imagers and electronics are balanced for a specific ISO, where the camera will have the best dynamic range and color science. When you increase the ISO, you're adding amplification to the circuit and you're also having to add more noise reduction to help combat the noise. 

A higher base ISO like the A7S series cameras, isn't necessarily a good thing unless you shoot in the dark. To get the A7S at base ISO (10,000) into a range that works outside in direct sunlight, you need A LOT of ND filtration. This of course changes your color balance and it's going to affect the image negatively one way or another. Where a camera that has a base ISO of 800, would only need a bit of ND filtration to work outside in direct sunlight. 

Honestly, the best solution is double ISO imagers. They have two preamp circuits, one focused on around 400 ISO and one focused on around 12,000 ISO. This means, BOTH ISO's are optimal with low-noise and it also means you don't really need much filtration when shooting outside and adjusting the camera to a lower ISO, won't be as detrimental. 

I'm very happy that double ISO is making it into cinema cameras. I just hope Arri picks up on the concept because it will deliver far better images dynamic range wise due to it's wider working ISO range. It wouldn't surprise me if we see triple ISO imagers at some point, it's just a matter of fitting extra components into the imagers die. The problem is that higher end imagers aren't mass produced in quantities great enough to afford the development of these technologies.