ISO is totally FAKE. Seriously.
Vložit
- čas přidán 18. 02. 2019
- Check out our #1 PHOTOGRAPHY BOOK on Amazon: help.tc/s
OUR STORE: sdp.io/sdpbook coupon 'CZcams'
ISO is fake for a bunch of reasons:
1) It's just a post-processing instruction; generally, it doesn't physically change anything about the picture taking process, unlike shutter speed and aperture.
2) While the term "ISO" refers to a real standard, camera manufacturers don't comply with that standard, like, at all. Many cameras are more than a full stop off the proper ISO. For that reason, light meters simply won't work as you expect them to. Using the same settings on multiple cameras will yield very different results.
3) High ISO could be eliminated completely; it's just an arbitrary limit to a multiplication problem.
4) Low ISO could be eliminated by simply using image averaging, allowing us to get long exposures in-camera without using an ND filter. This would also reduce noise in good conditions.
RECOMMENDED PHOTOGRAPHY GEAR:
STARTER CAMERAS:
Basic Starter Camera ($280 used at Amazon): Canon T3 help.tc/t3
Better Starter Camera ($500 at Amazon): Nikon D5300 help.tc/d5300
Better Travel Camera ($500 at Amazon): Olympus OM-D E-M10 Mark II help.tc/em10ii
LANDSCAPE CAMERAS:
Good ($550 at Amazon): Sony a6000 help.tc/a6000
Better ($1,400) at Amazon: Nikon D5500 help.tc/D5500 & Sigma 18-35mm f/1.8 help.tc/s35
Best ($3,150) at Amazon: Pentax K-1 help.tc/K1 & Pentax 24-70 f/2.8 help.tc/p24
PORTRAIT CAMERAS:
Beginner ($950 at Amazon): Canon T6i help.tc/t6i & Canon 50mm f/1.8 help.tc/c50
Better ($3,000 at Amazon): Nikon D610 help.tc/d610 & Tamron 70-200 f/2.8 help.tc/t200
Best ($5,300) at Amazon: Nikon D810 help.tc/d810 & Nikon 70-200 f/2.8E help.tc/n200e
WILDLIFE CAMERAS:
Starter ($1,100 at Amazon): Canon 7D help.tc/7D & Canon 400mm f/5.6 help.tc/c400
Great ($3,200 at Amazon): Nikon D500 help.tc/d500 & Nikon 200-500 f/5.6 help.tc/n500
VIDEO CAMERAS:
Beginner ($500 at Amazon): Panasonic G7 help.tc/g7 & Panasonic 14-42mm help.tc/p42
Better ($1,400 at Amazon): Panasonic GH4 amzn.to/2p5dAmD & Panasonic 14-140 f/3.5-5.6 help.tc/p140
Best ($4,300 at Amazon): Panasonic GH5 help.tc/gh5 & Metabones Speed Booster XL help.tc/mbxl & Sigma 18-35 f/1.8 help.tc/s35 & Sigma 50-100 f/1.8 help.tc/s100
DRONES:
Beginner ($400 at Amazon): DJI Phantom 3 help.tc/p3
Travel ($1,000 at Amazon): DJI Mavic Pro help.tc/Mavic
Better Image Quality ($1,500 at Amazon): DJI Phantom 4 Pro help.tc/p4p
Can't change the name because photographers are very high ISO...we're super sensitive.
🤣🤣🤣🤣
sensitive to fakeness?😃
remember when ISO 400 was a fast film?
Fast and grainy as hell. I used to shoot Fujichrome 50 and it was sweet.
@@billkaroly Ilford 400 ASA
@@billkaroly Kodachrome 100?
And you could push Tri-X to 1600 by processing it in Accufine. Also, remember when it was not called ISO but was called ASA?
@@theuktoday4233 I loved Kodachrome 25. Then there was the speedy Kodachrome 64.
As a professional photographer who also teaches photography, I explain to my participants/students that the ISO setting in the camera is quite similar to the setting of a stereo system, where the volume on the amplifier is increased to make the music louder. At ISO it is basically the same. By increasing the ISO value I amplify the sensor's initial signal, means the exposure values which are set by shutter and F-Stop. And as with a stereo system, the signal also starts to produce noise when overloaded. ISO setting is per se not active to exposure... Hope this explaines it as well.
Good comparison. Like the quality of MP3 compared to WAV files for sound vs JPEG compared to RAW files for images.
In digital cameras the ISO is a mix of analog and digital amplification. Tony believes that it is just digital amplification, but in reality most cameras have analogic amps with coarse steps, and with digital amplification cameras process the data to match the intended ISO, aka gain.
@@dzsemx Yes, but the actual pre-exposure steps are now reduced to two or three at best.
no dijiste nada y no das una solución a lo que dice, explicate para que te puedan entender
You are partially wrong. You need to learn about quantum efficiency, shot noise, read noise, and how dual gain sensors work.
I vote for it to be called Gain as well. Thanks for clearing things up Tony. Ive always wondered why camera companies focused on how high they could make ISO. Personally I was was always hoping they would come out with lower ISO values so I could shoot f1.8 outside on sunny days. Turns out they have been!!!
Native ISO (not extended) is actually analog signal boosting done by the sensor, so not the same as dragging the exposure slider in software.
this.
At least this is how it is in the Nikon world with Hi1 and Lo0.3 settings and such. Not all manufacturers always do this and some do make it just sliding a bar in Lightroom. Also, some sensors such as the D5/D500 use dual gain strategies at the sensor. IIRC, Canon on some cameras doesn't delineate as cleanly, and as using digital manipulation costs CPU power and slows frame rate on their 1D cameras.
Stop saying factually accurate things on this channel, please.
Generally it's a mix of analog and digital amplification. Also the manufactureres process very differently the raw data, like LR does very differently compared to most cameras.
It would be great if camera manufacturers would totally reevaluate the metering. A setting for analog gain, automatic expose to the right, multi-shot digital ND, and a floating digital gain as the third leg of the exposure triangle. The problem is probably a matter of how to not confuse the user with an absurdly complex new exposure pentagon. Simplicity and continuity with other cameras both past and of other brands is why ISO persists.
"The great thing about standards is that there are SO many to choose from...."
Not when it comes to film speeds. Or digital cameras.
But there shouldn't be. In the material testing world ASTM is pretty consistent.
Actually there is only 2. SOS (used by Fuji, Oly, Panasonic) and REI (used by Canon, Nikon, Sony). Nothing is fake here.
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 patternmetering) is used.
The Standard Output Sensitivity (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 the exposure is controlled by an automatic exposure control system calibrated per ISO 2721and set to 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, and DC-008[66] updates the Exif specification to differentiate between these values. 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.
@@stansid3326 So if nothing is FAKE, if I were to say traveling at 2000 RPM in fourth gear I choose to claim that my car is traveling at 100KPH, regardless of what the radar guns state I am actually traveling at?
Good to know.
Standards? We don't need no stinking standards.
Just curious, since you seem to know what you’re talking about. Raw data isn’t sRGB ? So is it CIE or more like light intensity values per sensor pixel, without color data assigned yet (or just labeled for one of the RGB colors) ?
You forgot to take into account that different lenses have different T-stop values, affecting the final amount of light falling on the sensor. Also, ISO isn’t digital gain in all camera’s, it depends on whether the sensor is ISO invariant or not. And even then there’s a small difference in noise performance and artefacting. Also, I do believe dynamic range changes with ISO.
As an engineer I can't even count how many times I stepped into things with names that make no sense or are ambiguous only for "historical reasons"... That's annoying sometimes.
I have only one thing to point out (probably will sound like a Cpt. Obvious) the maximum ISO is not just invented. I'm doing my master thesis on image sensors from the electronic point of view and the number and nature of noise sources in the electronics are really nasty. Manufacturers make a lot of efforts to try to reduce those effects and increase the multiplication factor of digital amplification (which is not exactly digital because the amplification part in the camera is analog) and they have some specific measurement procedures to establish when noise is "acceptable" so when they state a new maximum ISO it means they managed to refine something in the electronics (when they don't lie).
The real funny thing is there are standars also for sensor characterization methods but every manufacturer just uses its own methods and from there it comes the difference in ISO effects on different brands.
There is a maximum, but you're right it's not an ISO, it's one photon. (Apparently humans can see a single photon, but the odds are we wouldn't notice it.) All of this comes down to counting photons.
If you make your thesis into a youtube, I'd like to see it
@@WillN2Go1 WARNING: NERDY AND BORING ANSWER:
It's hard to say to what a manifacturer applies the ISO since they don't follow standards.
If you mean that he maximum ISO is the one allowing to detect a single photon, well it's not so obvious. The ability to detect a single photon depends more on the manufacturing quality of pixel architecture (a photon produces an electron and this electron must survive within the pixel without recombine with defects in the chip). Now, there could be some techniques using voltage to assist electron survival that I don't know but generally ISO is an indication of the overall ability to amplify the electrical signals coming out from pixels without introducing too much noise in the process. The amplification can happen both outside and/or within the pixel, but as far as I know this should not influence the ability to produce these electrical signals which is more a physical-chemical property of the sensor.
Having said that, I'm still in the study process, so I might not know some special techniques in this regard.
@@jack002tuber It would be a very boring video! LOL
I would love to read your thesis when you are finished. I'm sure it will be fascinating!
In order for this to be an accurate video, you would have to really look at each camera manufacturer, and each model they produce and talk about them specifically. A general conversation about ISO in digital cameras is destined to be filled with half truths, as different cameras implement the gain changes in different ways. As many people have pointed out, there are ISO invariant cameras, and there are cameras which do apply an analog gain before the ADC process, there are cameras with dual base ISOs, etc. etc. This is a gross over-simplification and may leave many people with a false understanding of how their camera actually works.
Collective Thought Media yup. Example, the canon 5DR is NOT fully ISOless, nor are any CCD cameras. This brightness adjustment you demonstrated would not work. Those cameras use noise processing in a different sequence, . Nikon D850, Canon 5D MK IV , however are all ISOless.
I think you are getting too deep in the weeds. The math is only dependent on sensor size (like the rain bucket analogy).
@@Bullwinkle056 some sensors increase the amount of analog gain being applied, where as other sensors do not modify the amount of electrical gain being applied to them but rather just increase the luminance on the recorded images. That's a pretty major difference and something that should be discussed.
I'm oldish ( only in numbers) .... yes old enough that as studio manager I would call the supply house ask how many boxes of an emulsion number they had.. if they had enough of what we wanted (say a case of 10 sheet boxes of 4 x 5 100 Fuji chrome.) I would have it sent over. It was then my job to test the film as to what its true ISO was. It was never ever ever what it said. It did stay true throughout an emulsion run. The point here is that ISO was never a set standard number of sensitivity even in film. Some times a 100 could be as low as 32! It also had a cast that needed to be tested for. When you spent 8 hours in the old days before you go to film on a single shot you cant slide a brightness button. You shot a set of 10 4 x 5 sheets. 2 on, 2 @ +1/3, 2 at +2/3, 2 at -1/3, 2 at -2/3. Run a single sheet of each exposure variation and hold the other sheets. After the lab ran the chrome (f you asked for a rush same day) You would get it back make very tiny adjustments on the second sheet as needed.... light meters where imperative. The point being the ISO was never intended to be anything more than a range of sensitivity...
I'm oldish too (56), and in the 80s I was working in CA for Mattel Toys in their photo dept. They'd buy 50 sheet boxes of Ektachrome 100 by the case and run emulsion tests. The saying then was Kodak wasn't as good as Fuji, but it was consistently not as good. Ha! Labs in Los Angeles used to have 2 hour E-6 turnarounds, which was pretty nice.
John MacLean I know right. Just run across the street for us. We spent hours making it perfect before film. Retouch?... absolutely dirty word and it was felt if you needed a retouch from a tiny piece of fun tack sticking out from a G.I. Joe shoe you just shouldn’t be calling yourself a professional.
@@silverleafcookies I also assisted car shooters from Boulevard Photographic that worked with 8x10 chrome film in the studio. The AD would grease pencil the acetate sleeve with corrections (usually reflections) that we would then reshoot until it was freaking perfect. The photographers said if you can shoot sheet metal, then you can shoot anything. So true. Back then the precursor to Photoshop called Scitex retouching and it was revolutionary for the things that couldn't be shot out. How far we've come!
@@JohnMacLeanPhotography so true! and all the round freakin chrome....dont wear bright pink to the shoot! and get out those giant panels of foam core. I love The designs of Harley Earl the father of the fishtail head stylist for over 28 years at GM in the 1950's. I'm writing a piece on Earl now and how he tapped into and created the culture of America of the time... he freaked out when he saw the P-38 airplane.. Most think is was rockets but it was an airplane that captured his imagination. (I work in equine portraits but my personal love is shooting Cars) .... check out the work of Sarel van Staden and Mayna Cotton. www.carfineart.com.....
@@darrellbissing well there are different opinions on todays craft. I think of it in this way. PS has taken the place of the set work. Yes, I can bounce a highlight into an image with a tiny mirror and take a couple tenth of exposure off a spot with a black piece of foam core, fun tack and armature wire. But it still all comes down to the ability to see the details and how it fits in the whole image. I will forever be grateful to Lou Goodman for teaching me this craft. The craft of seeing.
I’SO disappointed with this post
OK have a thumbs up from me
@@TonyAndChelsea not 100% true because the camera ISO boost's the analog signal from the sensor before it is converted to digital making less noise then low ISO digitally boosted because that also boosts the noise.
@@Narinjas Any proof it boosts analog signal from the sensor not digital?
6:17 - DSLR ISO is not exactly like "Lightroom Exposure Slider". There are additional mechanisms the camera can employ to reduce noise because it's interfaced directly with the sensor hardware; a slight tweak in voltage, I/O noise improvements, fourier algortithms on noise bands known to affect the sensor that can be broken down into regions... So, the end effect is similar (in that it's a digital gain) but the process could quite well be much more complex, so not exact in terms of just exposure n*X gain.
@@LieutenantMoustache all the ADC gain does is reduce quantization error; the rounding that occurs between the analogue signal and the digital result. It has no bearing on the inherent noise present on that analogue signal, so instead of watching a few videos and then think you know everything and go full Dunning-Kruger, understand that there are other processes that can be employed pre and post-ADC that can assist in the reduction of noice - such as I listed: Sensor voltage changes, I/O bus frequencies, oversampling, and fourier transformations (eg. which take an analogue frequency, identifies the component frequencies, and you can tweak that as you please). So yeah, I know a few scientificially sounding words... because I have direct exposure to ADC prototyping...
Educate yourself. Read some application notes from Microchip and TI on ADC, Topics include: using FFT in ADC, anti-aliasing, ADC data acqusition, noise suppression.
2:10 When comparing different cameras (and therefore I assume different lenses), could those lenses have a different T-stop that affects the result?
Yes, this is true, though I did a series of tests I don't show in this video using an adapter and the same lens.
I understand where you are coming from, and that is true. However, from what I've seen, the difference between f-stops and t-stops for, mechanically the same lens is about .2 stops (+/- .1 stops), so I use that as a general rule of thumb. Tony is indicating a full stop difference, so his point, IMO, is still valid for the differences in ISO.
@@TonyAndChelsea For clarity you should've just left this section out and shown the one with the same lens.
@@TonyAndChelsea Completely comparing it, even with adapters is quit impossible because it also the distance and adapters affect the light transmission, but otherwise like your point! ISO these days is totally arbitrary. I do a lot of workshops and always have to adjust per camera brand with the settings, Oly is mostly the lowest canon the brightest and Nikon somewhere in the middle. But that beside, it is just mainly a lot of marketing crap at this moment....
@@TonyAndChelsea I'm guessing using the adapter would mean the circle of projection would be be roughly the same diameter helping to eliminate variables?
Tony, to my knowledge none of the Sonys/Nikons/Canons are actually ISO-invariant (isoless). You even can easily test this in DPReview's comparison tool. For 1-2 stops it usually makes absolutely no difference to the naked eye, though going 4-5 stops quickly reveals sometimes huge differences. AFAIK the gain/ISO is applied actually when the picture is taken, not after that. The gain is applied electronically, not digitally. Also, at least the modern Sonys have a second base ISO. This really isn't THAT easy. You absolutely can say that it may not matter that much, it does to a degree though, even with the most modern cameras. Generalizing this for everybody is giving false advice.
I agree. I tested on my A9 a minute ago, took two shots at f/2.8 and 1/1250 sec. First at iso 20000 (properly exposed), second at iso 100 and really dark. No way to recover anything close to acceptable in LR.
@@conraddana It also makes a huge difference if you're shooting 12 or 14 bit RAW. I have a A7RII and it defaults to 12 bit instead of 14 when shooting in continuous or silent. No clue how your A9 behaves but either way in 12 bit this effect is way way worse!
«…none of the Sonys/Nikons/Canons are actually ISO-invariant….»
They ALL are. There is NOTHING that can possibly be done to make the semiconductor more/less sensitive to light after the fact. NOTHING!
«…the gain/ISO is applied actually when the picture is taken, not after that.»
It is impossible to apply gain before the picture is taken. The picture is taken, and, depending on the camera, the analog signal from the photosites are amplified, then sent to the ADC, and/or they are sent to the ADC, then digitally amplified, then converted to whatever raw format the camera uses, and/or converted to JPEG JFIF.
I actually shot those two with the mechanical shutter.
I'm curious, how exactly would you even test ISO invariance on dpreviews website?
Sorry, the proposal of not worrying ISO is wrong. It is OK advising beginners not worrying to much about ISO, use Auto ISO.
But thinking about ISO is highly relevant in digital photography. Basically the understanding of a signal path is not to complex:
At the frontend you have a nearly analog signal (not thinking about astro photography): the (high) number of photons hitting the sensor within the color channel and the exposure time.
The signal is changed to electricity by the sensor, thereby you can control the ampifying factor by the ISO setting. (the amount of photons you control by exposure time & aperture setting).
Being analog you have effects like noise.
Next step is the critical one: Analog-digital conversion: this means the analog signal is categorized in steps (i.e. 8 bit equals 256steps, 10 bits 1024 steps, 12bits 4096Steps)
What ever is showing up in signal computing you live with this input steps. every noise from the sensor is also pressed in this steps.
The only strategy for having the cleanest signal is hitting the white point already on sensor avoiding clipping (the analog signal is above range of A/D conversion) but giving as much steps as possible.
The result is the very simple photographic rule: get the exposure right in camera as bright as possible without clipping the highlights. you can do everything in post but you can not get any information of your subject in your file which went not through A/D conversion.
parece sacado de wikipedia
@@djblue00 jaja sí, pero tiene sentido. Northrup se equivocó.
@@fabrigarciacartoons no entiendo tu punto, explicate!
Thanks Tony! That was one part of digital photography that I didn't understand. As an engineer I can totally relate to your feelings about this. And yes, it's ridiculous that we can't set a lower ISO to avoid using ND filters.
wow, ive been doing photography for only couple years and never put much thought about it. But this really did open my eyes quite a bit. Thanks Tony, u are the best i dont care what Chelsea says about you!
I've found this with my a7s, in Slog 2 whenever you point at an 18 grey card it'll say it's exposed when it's actually 2 stops under. I wonder if Sony did this to make the dynamic range on the highlights seem more impressive.
Which is why you commonly hear people say "overexpose 2 stops with s log 2" they're not actually over exposing, they're just exposing correctly.
It would be interesting to take a spot reading of an 18 grey card and then set each camera to the value given by the meter and see how far off they are from the meter
No, it's purely a video thing. You don't get all of the sensor information in video recording so because of that you have to do some finagling slog you really do need to overexpose by 2 stops in order to get the image that you want.
That's just log doing what it's supposed to do. It's a logarithmic scale. By pretending it's linear and "exposing correctly", you are undoing the advantage of log recording.
Matt, Compression doesn't "change exposure" But it does bake in what ever exposure you have set. The difference in what you can do to a RAW file vs an MP4 or JPG is pretty huge. Most cinema cameras shoot in raw and you can change the ISO in post. Not so with a compressed format.
@@godboy159 "Most" cinema cameras shoot in RAW? I don't know... there are a ton that don't. For instance, isn't it really odd that Canon can't get it's sh*t together and just offer 4K RAW at a competitive price point? I mean, hell, the C100MK 2 has a 4K sensor... I have one-and love it. But it downsamples to 1080 (beautifully), and the best they unlock the firmware to do is uncompressed 8bit 4:2:2. I've said they could actually kill the competitive market if they offered paid upgrades to their firmware because they "hang their hat" on their color science.
Data rates and media used? Not sure why they don't. Don't they offer Raw with an external recorder? Don't know much about the Canon offerings.
I posted a longer explanation to F-Stoppers, but the reason why ISO is used instead of Gain is that its more accurate and user-friendly. Getting matching exposures with Gain requires you to know the base sensitivity of the two sensors and be able to match them (Assuming the companies offer gain adjustments that can be matched). ISO always gives you an exposure that can be easily edited in post to match another exposure and without losing detail.
ISO also made it easier for film photographers to switch to digital cameras.
I actually saw a small difference between the two images in Lightroom, the one that you boosted manually looked worse by a bit. A lot more color noise, so you maybe could change that in post as well? I'll still stick to using ISO through the camera, maybe boosting it a bit in post if I got the exposure wrong. There are some applications that can remove noise and what not from images, but I wouldn't use them for professional work personally.
This noise is expected: Quantization noise.
Modern cameras use 12, 14, or 16-bit numbers to store each pixel values in raw files. Their precision is limited. If you shot at -6 EV, then boost it by 6 stops, you get the same brightness, but 6 bits of data are lost forever.
Usually, quantization noise is much lower than sensor noise. However, keep in mind that *6 stops are a lot!* (The distance between middle gray and paper white is only 2.5 stops.)
As long as your exposure is *less than 4 stops off* from ideal, changing ISO or fixing in Lightroom makes *absolutely no difference,* because your final export will be an 8-bit JPEG file.
Oh, one more thing: You need to use Zebra Pattern to make sure your histogram doesn’t unintentionally clip. This is the only one more thing fixable through ISO only but not in Lightroom.
So now its a two angle triangle, thats disturbing
It's a line!
the two most important anyways was aperture (for the desired dof) and shutter speed (to eliminate or introduce motion blur). Iso was just whatever after that.
@@_morgoth_ iso is truly arbitrary, it's just numbers, even if sensors improve, we are always talking about stops, almost like exposure compensation, they both let us play more freely with shutter speed and aperture
I've argued this with lots of folks, and as far as a photographer is concerned strategically, ISO still matters. If you shoot at auto ISO in aperture priority, the camera might pick 1/4000 s and ISO 5000, or you could pick ISO 100 and 1/125s and get roughly the same picture. While you can adjust exposure in post, you can't remove noise from high ISO. It doesn't work both ways...
I find this obtuse.
1:55: Chelsea empties the dishwasher and starts putting away the silverware...
John Drummond haha tony goes somewhere and makes a video when it’s time to clean up..
Actually, they have a kid who probably does that as one of her chores. LOL
So I'm not the only one who heard that! So, was it Chelsea, or a subunit, or who? Inquiring minds want to know :)
no, she's more like: damn (throwing kitchenware roughly), he is making this tech videos again, just to fight with nerdy pixelpeapers for two weeks and i will not get any action again.
@@izoyt Said no wife ever..
hi tony, may i get more dynamic range if i underexposed ? in dxo we see more dynamic range with lower iso.
Will there be a part 2 video where you briefly discuss cameras with dual iso like the panasonic gh5s and black magic pocket cinema camera?
The main problem is that ISO (organization) does not provide one scale / measurable numbers for ISO (sensitivity), thus all manufacturers can choose which scale they want. The only standard today still valid is incorporated into flash meters which measure the light and calculate proper exposure (triangle), which is pretty handy, and can be used on film camera and on digital camera to get correct exposure. I can not imagine that digital camera does not have such measurable scale - how we can predict correct exposure with flash without having standardized sensitivity scale?
+ there is one simplification in the video - there are cameras, which incorporate a technology to physically change sensor behavior by changing ISO - called "dual gain" (all modern Sony mirror less cameras use ISO around 640 to change the sensor behavior and get less noise).
PavelR2 ISO 12232:2006 and 2019
Holy cow! This made me so happy I have a Fuji X-T3, as it appears to truly have the lowest base ISO. I've been dreaming of getting the d810 for that reason alone. You just made my year. I wish you continued success with your channel, Tony and Chelsea!
It just left over terminology from the days of film. My guitar amplifier volume goes up to 11.
How do gh5s have lower noise in the higher native iso?
What about dual-gain sensors? Like native 100 AND 400/640/800 etc ISO.
He has no idea what he's talking about. He learned just about the basics of sensors and how analog to digital conversion works and now he thinks that he's an engineer. The only thing he identified correctly is that iso varies between cameras. Nobel prize incoming
Wait... I'm wondering why is then ''DANGER'' written next to my max ISO setting, if this is all fake and just a postprocessing thing?
Gain from sensors (and noise reduction) is applied on both sides of the A/D converter, so you can't do it all in post, as analog gain is done on the ORIGINAL analog signal, not the digitized version which will always involve some loss of integrity. This is why many prefer vinyl records over digital recordings.
And yet digital audio recordings can store more detail/higher fidelity than vinyl ever could.
Wow Tony! What an awesome topic. You're like leading the deconstruction of a paradigm. Thank you!
Ahhhh, Tony, please be careful, only some cameras are ISO Invariant, not all! And also some which have dual ISO also are only ISO Invariant with only one of the circuits while the other is ISO Variant! People already don't understand this fact and you are perpetuating the misunderstanding with how you treated it in this video. Most cameras still have an analog gain at the sensor level, before being processed through the analog to digital converter and processed with digital gain.
Can you name some cameras with analog gain?
@@oscaralvarez714 literally all of em right now
thanks for pointing that out, you're absolutely right!
@@oscaralvarez714 how do you think sensor works? according to tony, by lightroom.
Well I'm getting mixed information. The video states that ISO is simulated now in most modern cameras, then here in the comments I hear that it's not. I know the signal from the sensor has to be boosted somehow but digital gain is a bit different than analog gain. *Edit: And I'm talking about digital vs. analog post processing gain. It would make more sense to make it digital as you can just change the digital gain value in post. With analog gain, the image would change a bit too permanently. That's what RAW is for right? For digital post-processing? I didn't sleep too well last night.
WOOHOO!
FINALLY! SOMEONE TELLS IT LIKE IT IS.
Thank you Tony!
As a audio engineer, I VOTE FOR: "GAIN". It's all about SIGNAL!!
BAM!
Great comparison!!
+1 for GAIN
calling it ISO makes no sence
Coming from a similar background - when I first learned about ISO, I actually remember saying "Oh, so it's like Gain"
It's just a shame you can't Pan the gain left and right - up and down.... it's kinda logical in a way.
Really it is signal to noise ratio, ISO is the gain applied to both signal and noise and as ISO increases then both increase to the extent where noise eventually swamps the signal. Calibration is a possible answer?
And then they lie about the gain values.
Great video as always, but question: when you compare 2 different cameras you’re using 2 different lenses too, is that correct? Wouldn’t the light trabsmission effectiveness from lens to lens give you a different exposure under the same settings sometimes too? I’ve seen it before in videos comparing lenses from different manufacturers in the same category.
the focal ratio is the determining factor on how many photons hit the sensor. You could have a 50mm 1.8 (which is tiny, ubiquitous and fits in your pocket) and you could have a 400mm 1.8 (which is bloody huge, like Eye of Sauron huge, and costs an absolute bleedin' fortune so if you spot one take a photo - you might never see another), the same amount of light hits the sensor. The difference between them is the field of view.
Hi Tony. The iso changes the analog gain of the sensor a\d converters. Like in audio, if you cranck too much, the white noise begins to rise to unacceptable levels. You can also mix with digital gain to give more smooth steps between but basically, is analog. This is the reason why some hi end sensors use cooling to lower the thermal noise floor of the camera and why bigger sensors achieve better signal to noise ratio. At least, the sensors has this options. Of the manufacturers perder so it digitally, I'm wondering why. Congratulate by the channel.😁
7:20 but cameras like the A7III have a dual base iso circuits, so you better shoot in either iso 100 or iso 640
What do you mean?
@@polaroidphoenix175 Google: "Sony a7 III dynamic range and high ISO improve over its predecessor"
@@aliendroneservices6621 eh I just googled dual base iso circuit
@@polaroidphoenix175 Nothing to do with "circuits"
@@polaroidphoenix175 "Note the jump in dynamic range at ISO 640 for both cameras. That's essentially the camera's second 'base' ISO, where the second stage of the dual-gain architecture kicks in. At ISOs 640 and above, most recent Sony sensors use a higher gain mode that essentially amplifies the signal at the pixel-level to get it above the (already pretty low) noise floor."
If the highest ISO setting is just made up by multiplying the signal intensity, what is expanded ISO?
Tony simply got it wrong. Expanded ISO is in the digital domain. Also, not only is there analogue gain applied, but there are so-called "dual-gain" sensors that accumulate the electron charge into a small capacitor for higher ISOs thus providing an effective gain without the noise that's added by a adding more gain via the active amplifier.
Really great video! I discovered by accident that a lower iso can be boosted in post for a correctly exposed photo, but I always felt badly when I had to do it (since it implied I screwed up my settings) & I never understood why. Thx!
Hey...
Is it the same reason why i didn't got any change in brightness of the two pics i took from my oppo phone by manually changing the ISO first time at ISO300 and second time at ISO1000 , when the shutter speed and aperture were same it those two situations??
Isn't the Fuji and Olympus difference partially caused by their using Standard Output Sensitivity rather than Exposure Index ISO? I say partially because also Fuji has a 0x9650 tag to tell the converter how much compensation to do that not all converters follow. I'm not sure if Olympus has that as well. Standard Output Sensitivity makes the most sense as a standard to me, but for the fact that ISO already meant certain exposure settings and SOS is not consistent with the film ISO-related exposure settings. In fact, the existence of the REI method of ISO means ISO must not find it at all offensive that camera manufacturers play games with ISO since REI method allows for that.
Tony has been lurking on DPreviews, Science and technology forum. good luck with that intellectual bearpit
logtothebase2, yeah, I’d call it a cesspool, personally. Dpreview in general is one of the unhappiest places on the internet.
Question: how do you “increase” ISO in post? Or do you mean just open the exposure more? Because thats fairly limited. And if you use autoISO for raw, would you not introduce more noise? Or would it be less noise in a darker image?
So i took a heap of shots at my son's birthday and they came up very grainy afyer looking fine on the camera screen and i was really bummed.. haven't got around to doing anything more than looking at the jpegs but if i go play with the raws is there a good chance i can recover them?
Canon 6dmkii
Excellent! And this is why I've subscribed. Your cutting through the photography world BS one video at a time! Thanks!
Tony - shooting RAW as you pointed out, lets you basically set the ISO after the fact, while with jpgs, you have to get pretty close. When a DSLR shoots video, isn't the video that comes out of most (if not all) DSLRs processed in camera and roughly equivalent to jpgs in that the settings are "burned in" when processed by the camera. There is no "RAW" version when shooting video, so you have to be more careful with exposure when shooting video, correct?
Chris, this is an internet comment section. Where are all the obligatory curse words and defamatory accusations against Tony as a professional and as a human being?? You need to step up your game, man.
Chris - you're absolutely right. I do mention at the end of the video that getting the "ISO" right is important for JPG and video.
Some very very high end video (cinema) cameras do shoot in raw.
@@villageblunder4787 not only high end. Depending what you mean by high end. Black Magic Pocket Cinema Camera shoots raw and it's a relatively cheap, "low end" camera.
And the DPReview article that goes into more ISO Invariant info - www.dpreview.com/articles/7450523388/sony-alpha-7r-ii-real-world-iso-invariance-study
Tony does it matter to know the sensitivity of a sensor? Where would you get the value? number? Thanks.
So, if I am using a Sekonic light meter, which standard is it measuring to? Is there a way to calibrate that to my Fujifilm camera?
There is one thing that should also be looked at, and that would be not all lenses are made the same, in terms of light at the same F-Stop.
I have only had the chance at one point to test the Kit-Lens & A Sigma 16mm both at F4 and the Sigma produced a bit more Light, but that said
It does make sense since the kit lens has such a small lens. But still begs the question, how stagnated are lenses in their numbering?
At museum of flights, in low\dim light and no flash conditions, I took many pictures(RAW files) of same plane with diff ISO(100, 1000, 2000).Other setting remained same. All pictures looked dark in camera. Later on I processed all pictures in LightRoom.
After raising exposure in LR, I could not recover picture taken at ISO 100. It was all dark.
I could recover picture taken at ISO 1000 but it was very noisy.
What surprised me most is that I could recover picture taken at ISO 2000 and it was no noise at all. It looked like ISO 100.
So, your theory suggested in this video is something I don't agree with. ISO is not about post processing but is about signal processing.
Also, I found that common understanding of lower ISO means lower noise is also wrong.
Regarding the "lower ISO means lower noise is also wrong", you're right, in a way... Check out Matt Grum's answer here, along with the discussion in the comments that follows. Alas, I never did follow up with further testing... perhaps I will someday?!?
photo.stackexchange.com/questions/6615/what-is-iso-on-a-digital-camera/
@@DavidLindes thanks for sharing this link. This explains what I experienced.
Could the differences in photos be down to the difference between F stop and T stop on the lens? Really interesting, though, is been wondering about this since the dual native iso thing on the gh5s.
We did comparisons (not shown in the video) using the same lens adapted to different systems. The third-party numbers equilize t-stops.
based on this i can overexpose with 25600 and then use lightroom to get it back to perfectly exposed iso 400 without noise/blown highlights? or have the same dynamic range as iso 100 with higher iso if i just use lightroom?
As everything in life, it's not that simple. Oversimplifying this issue does a disservice to a wide part of your audience that doesn't understand the differences that exist even between camera lines within the same manufacturer in this regard. Even in the film days not all film makers labeled their emulsions strictly following the ASA and later ISO standard, not even Kodak. The vast majority of cameras amplify their sensor's analog signal output before converting them to 1s and 0s, effectively changing the S/N ratio curves of the sensor depending on the selected amount of gain (or ISO). While there are some ISO invariant cameras, they are more the exception than the norm, and knowing how your particular camera works in this obscure detail will let you know how to wring the most juice (in this case dynamic range) out of your camera given any particular lighting situation. DPreview.com and Bill Claff's photonstophotos.net are the best online resources that tackle the particular behavior of each camera model regarding gain and ISO invariance, I highly recommend that you learn how your own camera behaves so that you can adjust your camera settings for the best outcome.
exactly. not all sensors are designed that way. A misleading video
amen.
@Ziggi Mon okay, but Sony isn't the only sensor manufacturer so generalization isn't appropriate.
A shot taken at ISO 100(base ISO/native ISO) then exposure increased by 4 stops in post processing and a shot taken at ISO 1600 aren't always the same. It's simply because increasing ISO in camera is (sometimes) analog computation but increasing exposure in post processing is (always) digital computation.
Though, I agree with ISO not being standardized.
And you get quantization errors when amplifying the digital signal in post and that type of error (banding) happens less when preamplying the analog signal before ADC by changing the ISO value (though amplifying too much causes visible noise in the picture).
@@AllanSavolainen > preamplying the analog signal before ADC by changing the ISO value
Yes, the video maker has no idea that most cameras do this.
So the +1EV wheel on the camera use the same principle of increasing ISO, just have higher gain? Do I get it right?
Hi tony, where can i get the data from the info posted at 4:30
Wait for the ISO V ISO pronunciation arguments...:-)
nah, it took a long time but I think we're all finally good with this pronunciation now.
Anyone who pronounces ISO as ISO is clearly wrong. We know it's ISO not ISO.
Hope that helps 😉
@@rod4eva LOL!
Thanks for the laugh!
I’m a big fan of your videos, but this one seems way off base. Here is the abstract for the ISO 12232 standard:
This document specifies the method for assigning and reporting ISO speed ratings, ISO speed latitude ratings, standard output sensitivity values, and recommended exposure index values, for digital still cameras. It is applicable to both monochrome and colour digital still cameras.
As you can see, there does appear to be an appropriate standard. The fact that some camera vendors do a bad job of calibration and reporting does not make the standard fake. Your complaint about the standard only muddies the waters. I would like to suggest that you correct your comments. The fact that I can take most modern digital cameras and a film camera and get equivalent exposure is something you might want to notice. The fact that my hand held light meter from many years ago matches my newest digital cameras light meter is more practical evidence of the value of these standards which have evolved over the years. Note that the use of iso as a post processing control name is a bit odd (it really is just gain), but it is meaningful for digital photographers.
Well, for being ISO = International Standard Organization, defining a non-standard, that's a hell of a contradiction to start with.
@@ducav2: how is this "defining a non-standard"?
It's a standard. And it's been updated as recently as last month:
www.iso.org/standard/73758.html
Adherence to standards isn't necessarily enforced or done well, but that doesn't mean it's not a standard... ?
@@DavidLindes "Non standard", I must admit, is not correct. But de facto, it can hardly be called a standard if it gives such freedom to camera manufacturer, to the point of leading to exposures differing between each other of more than one stop (given all same settings). They might as well have written "This standard specifies that the gain at base iso 100 is up to the manufacturer". That's not a standard as a matter of fact, it's just labeled as such. A standard which doesn't provide any tangible reference is just pointless.
ducav2: mmmm... sort of. I haven't actually read the standard, but I'm guessing that it has a much more restrictive specification included, and manufacturers are simply not following the standard (*at all*), but still use the name ISO. Perhaps the ISO body ought to have sought a trademark, as lots of other non-governmental standards (eg USB) have done: you only get to use the logo if you actually follow the standard (and probably pay money). ISO doesn't have that, which probably means (though I admit I'm well into speculation land here, though we could research this point) that it's easier for a small shop to actually get and adhere to the standard (or some arbitrary standard from the organization - which makes standards for *all sorts* of things, many of them having nothing to do with photography, so this matters), but also easier for any size shop to seemingly apply it ("it's called ISO, it must conform to the standard, right?") without actually doing anything of the sort, or even trying.
Point being: the appropriate criticism here falls one of two places:
1. The standards body not doing more to enforce that their name is only used in products that comply to their standards; or
2. Manufacturers for not adhering to the published standard.
What's not (I suspect) a valid criticism is to say that the standard lacks precision of meaning. (Though I suppose that's possible... if it's the case, then yes, that's utterly ridiculous.)
ISO is a standard, that defines a multiple other standards that manufacturers can use to calibrate and define the ISO.
That is like saying "World is ##"¤% place" only because one looks at the USA foreign politics, but that is not a standard for foreign politics, but different countries has different foreign politics...
And that is the case here, people think that because there are words "ISO" that it means "International Organization for Standards" (notice, it is IOS) that every single one is implementing one specific thing... no! There are multiple ways to implement the sensor sensitivity for photons (a digital sensor is sensitive for photons, not for radio frequencies, not for magnetism, not for anything else than photons, why even the IR spectrum needs to be filtered out as they are sensitive for those photons).
The ISO is now very sensitive topic because some people just want to claim that all the engineers, physicists, mathematicians etc are lying to the camera users by just holding on of some ISO for "backward compatibility with film".
Yeah.... The world most secret society just got revealed (again) by these people calling that ISO is a fake....
So lets see, one to trust a famous professional CZcams reviewers and celebrities, or to trust the engineers, physicists, mathematicians doing the theoretical science and applied science and producing all kind technology to peoples lives?
Hmm.... I think I go with the "God made it!" answer... Or not...
I plea everyone who thinks or claims that ISO is a lie/fake etc, to get their mouth where they think it is and travel to visit some of the experts, like www.enas.fraunhofer.de/en.html that they are calling as liars and fakers....
Or get the engineers and scientists responsible to design and produce those sensors in the digital cameras for an hour interview and tell them on their faces that they are liars and fakers...
Hi Tony. Do you shoot film?
Would be very interesting to hear your take on film photography and special effects photography (infrared photography etc). Cheers
Hi Tony , do you think digital cameras hardware is being overridden by software to some extent?, with new tech cameras coming with e-shutter, do you think shutter speed also will become like ISO today in future?
"this ISO doesn't seem right" when I was fiddling with the Sunny 16 rule on one of my digital cameras about four years ago.
I suspect the Sunny 16 rule doesn't work with ANY recent digital. camera
I haven’t done a scientific test, but I use sunny f/16 all the time with my X100F, and it seems fine.
@@joeltunnah huh. Not for me. Seems a tad too dark. It also doesn't match my light meter...but as I've discovered from some other comments, there are multiple ISO standards for digital photography. Fuji and Oly use one, the other brands use another.
Only works for jpgs in the camera
Great video Tony. I've been saying this for a long while. Even recently, I pointed out how I have to "calibrate" my light meter to every digital camera I use. They're all over the map. Well spoken and explained to the masses. Thanks for producing this & I hope everyone watches it! - Dale
I’m using my D850 to “scan” Kodak my old collection of negative film of 400 iso and I can see the actual grain in the film so clearly I can use it to focus on. Makes me realise that we are so fortunate to now have digital sensors that produce such high quality images.
I often use my granpa 1938 Voigtlander with ISO 125 film and sometimes I need to check the exposure so I take out my Fuji XT-20 and set to ISO 100 or 200.
Does it makes sense?
Why do you believe that Olympus and Fuji are out in their calibration rather than your reference camera, whatever that is?
One can use a calibrated light meter. Or just follow the Sunny 16 Rule with a Fuji and you will see it's a tad bit darker.
To compensate for their crop sensors' pooror noise performance than fill frame.
So reading some of the comments, there are two ISO standards for measuring exposure - Fuji and Oly use SOS standard, and the others REI standard. So...maybe they aren't being as nefarious as originally thought.
@@AIP404 suspicious? Yes. But not sure enough to make claims.
Quick searching suggests SOS is less subjective in measuring, so there is at least a benefit to using that standard. But sure, Fuji may reap some benefits of chosing the different standard, but ISO made it. And because ISO made it, Fuji and Oly don't seem so nefarious any more.
A follow up that would be interesting (but I have no intention on researching) is how that second standard came to be and why it was made. Was there a camera company who made it and ISO adopted? Or was it just a less subjective way to measure gain? That's why I remain suspicious, but I lack info to make the claim.
Here comes Tony with another video to make self described camera nerds incandescent with rage.
Anyway, I think I read some place that base ISO has more dynamic range than higher ISO settings. Does that only apply to JPG?
Smaakjeks K No also raw
As Tony said, ISO is a post-processing directive. If you take a properly exposed raw photo regardless of which ISO it had been taken with, you are utilizing the full dynamic range of the sensor. The sensor will effectively record the same information regardless of ISO you used. If you take a photo in JPG, or any other format than raw, the ISO you picked (and a reduced dynamic range) is basically baked in and any post-processing results in a loss of detail. The "base" or "native" ISO just means that if you properly expose at that ISO, you will get the most dynamic range. You will always loose dynamic range when you shoot other than raw. Only raw gives you a full dynamic range.
@@ocukor1
Alright. What about under-exposed images? They will have less dynamic range, right?
@@smaakjeks No. Dynamic range refers to a sensors ability to gather luminance information. Basically more dynamic range means the sensor will be able to record scenes with a larger luminance difference between the darkest and lightest part of a scene without losing any information. If you underexposed then you might be crushing the shadows in a dark scene, but that's not strictly speaking an issue of dynamic range but user error. If however you're underexposing without crushing the shadows, then you can still pull everything back in post. Doing so would result in unnecessary shadow noise though, since you could have exposed further to the right. In that case your sensors dynamic range didn't disappear, you just didn't utilize it.
@@youknowwho9247
But I never asked if the sensor itself lost dynamic range. I wrote "image".
I think part of the variability comes from using f-stops not t-stops for lens to set equivalent settings for images
Now that was a really informative video, especially the demonstration in LR. Can you provide a list of "isoless" cameras as a follow-up? I think having one setting less to care about during the creative process while maintaining full control is a very relevant feature.
So changing ISO doesn't change the sensitivity of the sensor. The load will transfered and calculated to digital numbers. When you increase the ISO you will increase the amplification of the stored load. In Photoshop with raw you do the exact same, only on the digital version, so the result is nearly the same. That's because some cameras name it db. So your result with the black picture makes absolute sense to me.
So let's take a closer look. We have a red pixel with value 64 and beside a red pixel with 65. That's a different of 1, but you would not see the difference with naked eye. So if we change ISO from 100 to 800 we increase with factor 8. The difference will now be 8, based on the two pixels. You begin to see the difference and will recognize noise. The interesting question now is: Are there ISOs that reduce amplification or reduce the value of the load. Why I think, that cameras not only increase the load, but also reduce the load? If a camera has ISO 25600, it means factor 256. On a possible value of red from 0 to 255 you would increase the difference of 1 to 256. That would result absolute crap. Okay, I know RAW has more bit, but even there 256 is much. So I think it is possibly that ISO 800 is the setting where the load is take 1:1. ISO 25600 than has factor 32, that could work on RAW 14bit. ISO 100 would reduce the load by also factor 8.
It could be interesting on which ISO the load would be taken 1:1, because it is the purest signal with the best result.
Before WW II we had around six standards; Steiner, DIN, Weston, ASA to name a few. ASA has changed at least once in the fifties. The only one never changed is DIN. It is a similar discussion in the analogue time about film. Btw, why do you hold a Gossen lightmeter, but do not mention it?
He did mention light meters. Probably all of us dinosaurs recognise a Gossen Lunasix 3 on sight.
@@stephenarling1667 hahaha I'm not That old :D
Don´t forget GHOST from Russia, HAHAH !!
When the photo magazines back in the days of Kodachrome mentioned Russia and film, "GOST" was the term. en.wikipedia.org/wiki/GOST
Then what determines noise. Appetite and Shutter speed only? And what determines how far we can push the “iso” or brightness before it becomes noises and unusable?
AJ Prospectives ISO only effects the Noise in the Image. The more digitally you brighten up your image the noisier it gets. And how high you can push a ISO while it is still good looking depends on the sensor and camera brand.
Hi. In other words you say that you keep your ISO in auto while clicking in manual mode. Right?
All of us old film shooters know that 😁
I read the title as "ISS is totally FAKE" and thought the Flat Earther's had got to you!
:-D
Tony I have a question, also anyone for the community can help. I just got the EOS R a few months ago and I tried using C-Log for the first time. In the C-Log mode when I try and set the ISO below 400, it adds an L to the number. I have read that means it's an artificial low ISO. From what you're saying in this video, do those L ISO numbers not really mean much, or is it different C-Log?
Did you correct for t-stop? Or use lenses where the t-stop was the same was f-stop?
waiting for Ken's rant
max factor ken agrees, he covered this a long time ago.
I was about to say the same ..... he was saying the same thing for don’t know how many years but no one believes on him. But now this channel says the same and while world goes crazy !!!!
Jan Novák .......master 🤣🤣🤣🤣🤣 !!!!
I am not sure about what you said about ISO being just "gain." I would expect that the ISO setting would adjust the upper voltage in the A/D converter. From that sense, it is a lot more than just a software post-process.
So are you saying, use auto ISO? For video also?
he said, is fine any iso if you shoot raw because you can achieve the same result in post. But no if you shoot jpg or video, in that case you need to nail the exposure.
It does not matter for the exposure if you use the camera to measure it. The problem is that you think your camera is better than it actually is. you think you shoot with ISO 800 and in reality you shoot with ISO 400. Then you are happy how little noise you get with ISO 800.
I saw your another video where you explained T-stops. In the part where you compare 3 camera ISO brightness, did you chceck T-stop? Because problem might be in lenses.
Just learn your particular camera and how I.S.O. affects it....problem solved..I think not really worrying about your ISO is a mistake, if you have it way too high your not going to get rid of all the noise, too low and I dont see you being able to recover shawdow details
One easy experiment anyone can do to get more detailed knowledge of their camera’s ISO behavior is to spend an hour or two doing photographic experiments with their own camera. Set up (on a tripod) a RAW shot on an appropriate test subject (I used some magazine covers on a brown table, with a color card for reference). Set it up so the exposure is “right” at a nice high ISO. Then take the exact same shot with exactly the same shutter speed and aperture, dropping the ISO as you go. If you are a detail nut, do it 1/3-1/2 “stop” of ISO at a time. If you can do with less detail, just cut ISO in half each time (1 “stop”). Go down about 5 stops. So, for example, I did ISO 12800 down to ISO800 in 1/3 stop steps. Then dive into and boost the exposure by the exact compensation (.33, .5, .66, 1, 2, 2.33, whatever) needed for each shot you took,. Then look at the differences between shots. My camera is COMPLETELY ISO invariant above ISO 800, and my experiment confirmed this up to ISO 12800. Raising my ISO800 exposure by 5 stops resulted in EXACTLY the same image -with full shadow recovery and dynamic range relative to the ISO 12800 image.
So while your statement about “get to know your own camera” is spot-on, the assertion that shadow recovery is likely to be a problem is less so. If your camera is truly ISO-invariant then you will NOT have ANY problem with shadow recovery - other than the extra effort needed to do it in post. At least not with RAW images.
Shooting ISO-Invariant might even change how you shoot. Given how ISO settings usually only amplify signal rather than attenuating it, on an ISO-invariant sensor it turns out that highlight recovery is a far greater risk than shadow recovery. So if two settings are both within the same ISO-invariant range and you plan on post-processing, then it may be safer to turn the ISO down a stop or two from “proper exposure” in order to be able to maximize the amount of highlight detail you preserve. You will not be crushing your blacks since ISO settings are expressions only of amplification (digital or analog) and not attenuation. (ok - maybe the “fake” ISOs under the base ISO can crush blacks - and that sounds like another good experiment).
@@AcGtrNut good idea on the test, I usually use the minimum iso I can get away with and get the exposure right or within half a stop or so because when it comes to post I dont like to have to do much to my files because it makes me feel like its no longer my photo but a doctored image..His idea of just shooting and fixing it all in post makes no sense to me, I would rather get the exposure right or really close in camera...I shoot birds in flight 99% of the time and all I did was test how high I could go with iso and get an acceptably clean shot but what you did is a good idea..you were talking about being safe and turning your iso down a stop or two which works well so you wont blow the whites ( highlights ) another good way is to just dial in a half stop of - exposure comp..it works great or it does on mine..negative tells my camera that it needs less light and just sets the meter in the bottom of the view finder accordingly, I shoot Nikon and im unsure if exposure comp is the same from brand to brand.. - exposure comp may work opposite on other brands, I really have no idea
Wait... I though shooting at the lowest iso made it eaiser to, and cleanly recover shadow detail
@@evertking1 yes, but not like he was saying. He was saying just don't worry about iso and fix it in post, well if you are 5 to 10 stops underexposed its going to be hard to recover anything
@@bassangler73 Cameras can recover around 13 stops or more of dynamic range.
I liked it better when it was ASA.
After a few years of trying to nail my iso, I just switched to aperture priority and automatic iso. Seems to work for me in most situations.
Hey is that a gossen lunasix light meter!? Could you please do a break down on how to use it I have one and get confused alot with the two modes and how to use them
Tony is wrong about max ISO. Bigger pixel pitch sensors, like A7sii, is the reason why it can go to a higher max ISO. Bigger pixel allows the analog SNR to be pushed further before falling apart. Yes, you can push raw from A7ii to same max ISO as A7sii but A7sii will give a much more usable result because it has higer analog SNR thanks to its sensor design.
You are saying the same thing Tony is saying.If you can push the A7ii to the same max as A7sii, then he is right. Whether is usable or unusable, doesn't matter. Because that interpretation is left to the end user and no the camera maker. For example, the a7iii max 204800, but I don't consider that usable. I max out at 12,800. Everyone has their max range based on how comfortable they are and what work they do. Someone might be ok with 204800, and still add more gain in post. If they are okay with that, that's up to them.
@@TOKSVISIONZ Please don't twist what I said.
I said A7sII will give a much more usable result at high ISO range(>256000) and able to go higher native ISO than A7II. If you don't believe me, please go to dpreview image quality comparison tool to check for yourself.
Tony said "Max ISO is fake." I give an example that shows it is NOT fake. That's all.
DON't mixing "usable depends on user" and thing like that. Thanks.
@@cks7249
yes Tony could explain every sentencen he said with 100 addition sentencen
the vid will never end...
theoretical there is no Max ISO
You can multiply the measurements with ridiculous huge numbers.
EVERY gauge can be used the way. But by this way measurement errors will be multiplied too!
You can measure the thickness of your thumb and than measure the lumber thicknis in "thumbs". Multiply by the thumb thickness, you get close to the nominal lumber size.
But if you measure your whole house with this technique, there will be huge errors.
So in PRACTICE you should use a better gauge for these work.
and that is what tony claim:
theoretical there is no ISO Limit
but because not all sensors share the same measurements errors, of course there is a different practical max ISO
@@tommyjacobi2054 No. You're twisting what Tony said.
Please re-watch the video from 5:36, Tony said "ISO is just a digital gain, same as just dragging exposure in light room", the underlying message is that Tony thinks Max ISO is fake, you can shoot everything in base ISO and it is just a adjustable gain in post-process.
This message will mislead people in believing there is no reason to get GH5S and A7sII because they can archive the same result with any camera, which is wrong.
From 7:33, Tony's claim is misleading. Max ISO is not makeup lie.
Some camera listed the Max ISO in marketing materials, which may be meaningless, but it is meaningful on some camera like A7sII.
From 8:20, Tony's claim that everyone can archive high ISO with old camera, it is also misleading with the same reason stated above.
@@cks7249
Tony said beginners should watch another vid.
i dont share your fear, there could be thousands of people which now dont buy the needed A7sII...
Of course this isnt a buying guide or advice !
Yes there may be two, three or even five people world wide, dumb, brainless people which good ammount of money, which now buy the wrong cam.
Independently from the cam those people will use, they NEVER make good photos.
So who cares?
Thats what i said: Tony could spent hours to explain every sentence in ten ways - but why should he!?
For people which think for themselves its absolutly clear what he mean:
There ist NOT build in feature that acitivates ISO 200.000 on a sensor.
ISO works on every sensor the same way: amplify the measurement
that is simplified and not an exhaustive explanation
But even in school your first learn only natural numbers - simplified, to understand the system. But not the exhaustive knowlege to be an engineer.
you take his statement much too serious.
Tony never said: hello audience, now you can switch of your brains
he never said or even imply that somebody should choose his next cam by ignoring the ISO quality
he just simplified the process how digital iso works
sure for an engineer this could sound OVERsimplified - if he left his brain home and didnt get the purpose of this simplification
"My whole life was a lie" 😶
How did you compare APS-C to fullframe in your chart at 4:30 ? Because of their different sensorsize they are "calibrated" differently aren't they?
So, when I'm adjusting the exposure compensation... is that just fine tuning the ISO/gain on the sensor?
On isoless camare (most are now days) YES.
Nobody should care about iso numbers in the digital word. This is just a number. People should care about result for the same scene whatever number is used.
And NO : iso is not a digital gain it is purely analog, it is made on a voltage signal before reading out the value, before it is digitalized. That why iso in general, makes a better job than increasing brightness in post processing, because you do not amplify the read out noise, contrary to post processing digital amplification. However , the sensors read out noise being lower and lower, the two kinds of amplification get more and more similar. All camera have an iso for which the iso is so called invariant, meaning I put référés read out noise get smaller.
Cheers.
In many cameras it is a mix of analog and digital gain.
For example in many Canon cameras there is highlight tone priority mode. Usually when this is activated, you lose ISO100. What happens here, the sensor exposes for ISO 100 and the digital postprocessing brightens the image with 1 stop. With ISO 50/64 happens the opposite: the camera exposes as usually for ISO100 but uses all of the headroom from the highlights. Basically if you have the JPEG overexposed, you RAW will be overexposed too. At normal ISO 100 usually you have 1-2 stops reserve in the highlights So basically dragging the ISO100 exposure slider to -1 in post processing makes a very similar result to ISO50
@@dzsemx yes, you can find exceptions. Canon used also digital amplification for intermediate iso step and analog for full step 100,200,300, ...
What you describe is also call DR mode, the sensor is at iso 100 underexpose by 1 stop to conserve DR and than either a flag is written in the raw files to tells your software to increase by one stop the brightness or a curve is apply to the jpeg. In normal operation and in most cases gain is analogical.
i'm MrBlue thanks for this constructive input. It real helps.
@@photaudiotech5550 we need to learn about you, try to say something, that real helps
Make Photography Great Again !!!
So if I have Sony a6300 and I’ll take my photos in iso 3200 and above, I would be able to reduce the noise by bringing the exposure down in Lightroom?
Is it really just post processing and they actually don't change the voltage of the sensor or something to change physical sensitivity?
A+ clickbait title
It's not clickbait, though. He actually talked about what's in the title.
Maybe a better title would have been "Modern ISO is totally FAKE. Seriously"
They have been creating a lot of click bait lately but this is legit!
@@fixitrod4969 not 100% true because the camera ISO boost's the analog signal from the sensor before it is converted to digital making less noise then low ISO digitally boosted because that also boosts the noise.
@@Narinjas Every big claim like this isn't going to be 100 percent. The statement the sky is blue isn't 100 percent. I'd love to learn more if you have documentation on the 80D sensor and Digic 6 processing. I see a lot of people disputing this but can't seem too find documentation to support this for every/most sensor as people disputing Tony have claimed as well. So, I decided to test it on my camera. I took some pictures from iso 200 to 12800 in 1 stop increments and adjusted the exposure the amount of stops needed to match 6400 with an 80D and I can't tell which was taken at 200 or 12800. I don't know what the science is but the outcome I saw is 100 percent what he said. The grain at 10x might be a slight bit different but I'm guessing that's just the difference in how light room processes the gain/exposure change in iso vs how the software in the camera does it. If there is more science to it on that camera I'd love to see canons documentation on how that sensor processes this conversion that I can't seem to visually see. I really would like to read more about it because I just don't see it being different visually with my 80D, which for me is what matters.
Here are the images I refrenced.www.dropbox.com/sh/lop1u0rjrz4iruc/AACXKn4T9zUaflaX2F0Awlega?dl=0 These are zoomed/cropped quite a bit, I don't remember how much. I don't know if the meta data is still with them either. These are just hand held and realitivly low light and no flash. F 2.8 and I think 1/60 second but would have to check. Exposure to match the amount of light was approx iso/gain of 6400
@5:47 - "in fact the camera captures the exact same image in the exact same way, and then after the picture is taken it uses your ISO setting to determine how it's going to post-process"..doesn't the amount of electrical current flowing through the sensor increase so as to make the photo-diodes more sensible to light at least in ISO-variant sensors ? Not considering myself an expert, but there were quite of few things that seem oversimplified, compared to my understanding and knowledge. Why would you call this a video for nerds with all the over-simplification ?
@6:05 "the ISO that you choose in your camera is just a post-processing instruction EXACTLY like dragging the exposure slider in lightroom" - now that - I think - must be false. What cameras are we talking about? There's only a few very recent cameras, like the modern mirrorless cameras and the D500/D5 as far as I know that are ISO-invariant. Perhaps I'm wrong and anybody may correct me, so that I can improve my knowledge.
This definitely varies by camera model. I'm sure his statement is true of some cameras, but I just did a test on mine and I got very different results capturing at 100 and increasing exposure in lightroom vs capturing at 6400, with all other settings the same (in particular aperture/shutter).
I don't know how exactly ISO is implemented in my camera, but it wouldn't surprise me if there is analog amplification at the sensor going on, in addition to whatever postprocessing the camera does. I captured raw so if it were purely postprocessing I wouldn't think it would matter.
an excellent detailed article on ISO-Invariant cameras and the pros and cons - improvephotography.com/34818/iso-invariance/
Thanks for this video! Ever since I started reading about ISO-less cameras on DP-review I changed my approach to concert photography, using fixed settings for shutter speed and aperture, and underexpose a few stops to avoid highlight clipping, knowing it wouldn't be a problem to crank up the overall exposure. However, I do use a lightmeter when using strobes, so there is still an importance to have a standardized value for ISO, no matter what you call it. Maybe 'gain' is a more appropiate name, but it should be standard like ISO once was.
I stoped useing a meter for a many years. I started useing my meter more recently but have found that with my camera it seems off. This may be why thanks Tony. So how do I find the bace ISO of my camera so I can use a meter to do things and have the right seting? I shoot JPG most of the time.
Wait until,the person with a big hair,sees this video... 😗
That person is compensating for something with their big hair. And it's not exposure what I'm talking about, if you catch my drift
@@sergeretaj907 Im gonna guess "little fro" has the same hair style as up on top.
@@timjones4594 I think that person's hair might remind some people of Colin Kaepernick.
In other words, shoot RAW and have the lowest ISO you can preview with, and the rest doesn't matter! (except for video I guess...)
pixlplague, or shoot jpeg in an auto mode. The camera’s meter is still going to be correct, no matter how they fudge “ISO”.
You can cheat on many cameras:
- create a 1-2 stop brighter image profile and load to the camera
- use highlight priority mode on Canon cameras. Basically shoots 1 stop lower ISO and boosts 1 stop digitally. I saw on other brands similar functions.
Could you check pls, what happens to the sensor when you're changing ISO? it's physical/electrical change...
Thanks for the video! @4:26 you show a nice graph of real ISO values - any chance you can provide a video showing the tests? Does this mean that Fuji X-Trans has the best sensitivity? Or do I understand it completely wrong.