Light meter calibration revisited

Over time I’ve written a lot of articles on the use of the light meter, and also some about the calibration of the meter. This has fueled some discussion and also made some people doubt their methods even more than before, so I thought that it was time for a revisit of this topic.

 

Why calibrate the meter
In fact you are not calibrating the meter, but the combination of the meter and the camera/lens combination.
Remember that every camera can be different, and that if ISO100 is correct it could be off for ISO200.
Knowing all this makes it easier to understand why the meter out of the box is not 100% accurate.

 

The quick and easy way
I’ve been experimenting a lot with the calibration process and found that if you calibrate to a QPcard101 and get the numbers to read app 128.128.128 for the gray patch you are pretty accurate. However there are some things you do have to realize.

There is no middle gray
Yes you read this correctly, there is no middle gray that works for everything.
When we edit we use a colorspace, for example sRGB, aRGB or ProPhotoRGB, but that’s not all, we also add color profiles and of course some people add some curves (contrast) to the images. This ALL has a huge impact on where the middle gray point is exactly.


Now it would be easy to just throw out some numbers and tell you that you should calibrate on that point, and nowhere else. And sorry it doesn’t work that way. The calibration itself is highly depending on your workflow. Meaning (as I always mentioned in the posts) you have to do the calibration in the workflow were you do the editing.

 

Now that we know that middle gray is not always the same how do we do a proper calibration ?

 

Sekonic
If you own the Sekonic 758 you are “lucky”
Sekonic offers software to do the process for you automatically. In the past the outcome was not something that was accurate in my workflow but since they released version 3 of their software I’ve been revisiting it and the outcome now is pretty much dead accurate. However I do have to not that I have to do the calibration on a windows machine because they still have not solved the kernel panics under MacOS (latest version).

 

What you do with the software is shoot 3 exposures with a 3 stop difference.
Edit this in YOUR workflow and store as TIFF 16 bits with the COLORSPACE you normally store your files in.
Now import these into the software and a few seconds/less than 3 minutes later you have a profile, you can do this for reflective and incident, ambient and flash and for several ISO values. In total you can store 3 cameras (calibrations).

Now in the past you had to buy the “expensive” Sekonic target, but since the new software they also support color checkers including the Color Checker Passport from X-rite which I use myself. And I can say that the calibration with the color checker passport went smooth and the result is accurate.

 

So I hear you ask “where does the 18% gray patch end up after the calibration in your workflow?”
Well in my workflow it ends up around 140.140.140 a little bit brighter than the 128.128.128.

 

Now you will start wondering, “was the 128.128.128 method wrong?”
Well no.
Overall one can say that the method works pretty well for getting a very accurate average (works on most cameras) calibration profile, in some workflows it’s dead accurate, in some workflows it’s off by 1/10 or 2/10 of a stop. But overall you will get great results. The reason I’ve been using this technique (and still teach people it) is that it’s a very easy and very cheap solution with instant results.

 

Now how to do it if you don’t own the Sekonic with software?

 

 

Alternative ways
Not everyone uses the L758 or a similar meter, so are you then “stuck to the 128.128.128 method”.
Well again, there is nothing wrong with that method, but if you want to be really really accurate there is another method, but it takes a lot more time.

The first thing you need is a color checker, doesn’t matter which brand as long as online you can find the values for the patches, X-rite has them listed for example, but you can also visit http://www.babelcolor.com/main_level/ColorChecker.htm for a nice collection of documents on the subject.

 

Now that we know all the values it’s not as easy as 1.2.3 for the simple fact that if you use a different curve there will be a difference in the values. If you have differences in the measured values and the ones on the list I would advise to make sure that you have white not higher than 250.250.250 whatever you do, often they are supposed to be in the 245-250 range. And then check your gray patches, and try to keep them as close as possible to the values subscribed. You can now adjust your curves to get the middle grays as close as possible to the listed values.

 

If you have huge differences you will have to decide what to do.
In most cases there won’t be a lot of difference, but if you use extreme curves it will. The question you have to ask yourself then is “do I want to blow out my highlights to get the middle tones correct, or keep the whites and underexpose” or in other words it will probably make your realize that the curves you are using as standard are not “accurate”

 

Added
As mentioned in a reply below.
I’m still in the testing phase of this procedure.
What I did find out which is rather remarkable is that when I calibrate in Capture one on 128.128.128 with the Leaf profiles in place, the numbers in Photoshop for the RAW files (with the Colorchecker profiles) end up at 142.142.142

 

This is of course easy to explain due to the fact that a profile contains curves and will change the outcome of the middle gray point, it also points out that it’s incredibly important to check the outcome of your calibration and do it in the software you shoot with (your workflow).

 

I will get back to this topic as soon as I’ve done enough testing and have a conclusion.
At the moment my conclusion is simple, the 128.128.128 method has worked very well for me over the years, I kept detail in the whites and the shadow areas were also spot on. The difference is not huge with the Sekonic method so we will see what happens. What I do experience is that especially the newer cameras like the D800 and 5DMKIII were underexposing with the 128.128.128 method on whites although the skin tones were perfect so the question also comes to mind, where is the calibration based on with the sekonic calibration, is it based on the dynamic range of the camera meaning it will make sure that white is on the top end of the scale and does not blow out, or is is based on where 18% gray should be….

 

If it’s on the last I think that is a great way to do it to maximize dynamic range, seeing that modern cameras are using a different dynamic range method than previous models, during a workshop a few weeks ago we experimented with a 5DMKII, 5DMKIII, D800 and D70 to see when a white background would burn out, skin brightness was correct on all cameras, however both the 5DMKIII and D800 needed about 1/2 to 2/3 stops more power on the background before the background would blow out 100%. At the moment I’m in contact with Sekonic and will ask them this of course.

 

 

33 replies
  1. Stefan
    Stefan says:

    Hi Frank, interesting topic and thx you found some time to test the Color Checker Passport using relaese 3 from the Sekonic software.

    So if I understood your explanation and procedure well, after you calibrated with the Color Checker Passport with the new Sekonic software and your figures are as you said around 140 140 140 you can start shooting properly. So you do not have to correct them also according to your well known teached method with for example with a QP101 card to 128 128 128 ?

    Gr. Stefan

    • Frank Doorhof
      Frank Doorhof says:

      Depends, I still have to test this.
      Also remember that I always stated that the 128.128.128 method is the best starting point for accurate exposures. I still stand behind that.

      It will however depend on the way you approach the matter. I still have to wrap my head around the fact that it’s embedded in my brain that if something is 18% gray it’s supposed to be in the middle of the histogram.

      I’ve been using that method for years and it has gotten me perfect exposures with the Leaf, BUT slight underexposure with the Canon 5DMKII and 5DMKIII, but again perfect with 5DMKI, 20D and 10D

      So I don’t think there is one truth at the moment. I am however (as always) bringing out what I experience and what I think about it.

      Let’s say it like this.
      With the version 2 of the software it was so buggy and the results were so “weird” that I stopped using it and started looking for another solution and started using the 128.128.128 for 18% gray method that gave me the perfect exposures.

      During the workshops the last few months it appeared that especially the new generation of cameras are showing something I never saw before, a Canon 5DMKII with my meter gave perfect exposures, while the D800 was underexposing like half a stop on ISO100, the Leaf was perfect on 50-100, a D200 was slightly too hot on ISO100.

      In other words, the fact that you have to calibrate per camera is without a doubt true, so calibration in needed.

      However which starting point is something that seems to be not so easy, because on both the D800 and 5DMKIII white metered reflective blows out almost a full stop later than with the 5DMKI. Which makes you wonder if by now it’s not time for a different calibration method, at least something that takes into account both the top and middle part.

      When Sekonic released version 3 of the software I had to make some times to test it (which I did yesterday) and total truth to be said I still have to start shooting with it.

      So time will tell which is perfect.
      The difference between 128.128.128 and 140.140.140 is by the way not that radical but it’s interesting.

      I know for example that Photoshop (which is my workflow) always adds a little but of a curve, in comparison with Capture one.

      The funny thing (well call it funny) is that when I calibrate in Capture one on 128.128.128 the same file into photoshop reads (and I really kid you not) 142.142.142

      The reason behind this (I think) is in the profiling, Capture one uses profiles for the Leaf back, when I export the RAWs into Photoshop, I use my own profiles. Profiles also use curves of course and as soon as we start shifting just a few percent a jump from 128.128.128 to 140.140.140 is there without a doubt.

      The reason I ended up with the 128.128.128 method also has some other base, according to some calculations the middle gray should render as 110.110.110 and 119.119.119. I of course also tried that a few times and ended up with images that were clearly way too dark. With the 128.128.128 method I always kept detail in the whites and good detail in the blacks.

      Time will tell (probably within 1-2 weeks) if the Sekonic method is the one I will be using or still the 128.128.128 as soon as I’ve done some shoots I will know, the difference is not that much but especially in the whites we will see if the dynamic range of the camera compensates for it, or if it blows out.

      Every day we learn, every day there is a new development, if we stand still we don’t progress 😀

    • Stefan
      Stefan says:

      thx and it is, or at least for me, a very interesting topic. At the same time, the more I know and understand about it, the more I discover the difficulty from this subject. Therefor I really like your approach with the grey card and the 128’s in your own workflow. And if That works it works as simple as it is. I am going to do some testshots coming weekend also to discover and try to understand possible differences. And as soon I discovered ( again) my workable flow I will spend time to shoot rather than being to much involved in technics. But…..it remains a very intersting topic 🙂

  2. John_Skinner
    John_Skinner says:

    Kudos Frank for even bringing to our attention that Sekonic had now included the XRite as an alternative to the extortion of $199.00 +/- for a target to use our $600.00 meters properly. The addition of the firmware update is a nice perk too. I’m happy Sekonic has not forgotten about it’s users. I’ve not had the time to invest in this newer method of creating profiles with the passport inside ver.3. Nut I am looking forward to the results. So between your take on things, the F1 help file, maybe I’ll have better luck with this. My past attempts to archive 128 – 129 – 128 across the board with either body, OR lens. failed miserably and I just gave up and just used the meter ‘out of the box’. Again, thanks for all you do for all of us.

    • Frank Doorhof
      Frank Doorhof says:

      In what case failed ?
      I’ve used the method for so long with so many different cameras and it has always been very accurate, although with the newer cameras (as mentioned in the article) I’m seeing the dynamic range is higher so calibrating to white peak does make sense, although the skin tones will shift, but the difference is not that huge, so I’ll see what happens, if I have to touch the exposure slider too much I’m back 12 points within a heartbeat 😀

  3. SitchaChante
    SitchaChante says:

    Hi Frank

    After carefully following your procedure for calibrating the meter to my camera, I also have the discrepancy you describe regarding the mismatch between Capture 1 and ACR. My numbers are about the same as yours…that is…128 128 128 in C1 reads about 143 143 143 in ACR. This is actually kind of a bummer for me.

    I use Capture 1 mostly for tethering and do my work in Lightroom and PS5. I am thinking it would be better to nail the numbers in ACR and let the levels fall where they may in C1…even though they seem a bit underexposed.

    I was wondering how you reconcile the difference between the 2 in your workflow. You mentioned in a previous post that you had an ‘offset’ in C1 to even them out. How are you doing that?

    I’m actually warming up to Capture 1, especially the way it renders skintones…but the difference I am seeing in the grey values makes me think there is something amiss..like maybe I did something wrong.

    I hope you keep us posted on the new calibration software you are experimenting with on the Sekonic L758. I am curious to learn your final thoughts.

  4. Elias Mereb
    Elias Mereb says:

    Hi Frank, thanks for all the tips! Did you heard back from sekonic? I own a D800 and just got a L-758, I also have a ColorChecker and began playing with the sekonic tool, but it asks for both incident and reflected light values.
    So here are my questions:
    1. When calibrating with the ColorChecker, in which color should I meter the reflected light?
    2. Is there any difference between the quick and advance mode? any difference in the end result?
    3. Does it matters if the ColorChecker is in perfect focus? the help file says to disable VR and AF
    Any other tips for calibrating/profiling for my D800?
    Thank you so much for taking the time!
    Best Regards

    • Frank Doorhof
      Frank Doorhof says:

      I’m not behind my studio computer at the moment but the reflective is discussed in the software or manual from the top of my head it’s the third.

      Advance mode gives you more stops

      I think it should be in good focus but I don’t believe it has to be pinsharp 😀

      Sekonic is working on it.

  5. Fidel
    Fidel says:

    Frank, this is a dumb question, but last night I was trying to calibrate my sekonic 758.
    I started using the big grey card without any other color. then I noticed it got really dark I had to use -2.4 calibration.
    Then I use the 4 grey on the other side bottom row. then I have it at -0.7.
    which color should I use?

    • Frank Doorhof
      Frank Doorhof says:

      Hi,
      It has to be an 18% gray card.
      I don’t know what you did use but in theory when you get a real 18% gray card it should work.

      Also remember that with the new Sekonic software you can use the color checker passport to do the calibration.

  6. Joseph Holland
    Joseph Holland says:

    Great articles, you have convinced me ten times over to finally take the plunge and buy a 758. I have a question regrading its ability to store 3 profiles. Would you recommend profiles for 3 different lens at one set iso, or one lens and 3 different iso’s?

  7. Michael Katz
    Michael Katz says:

    Hi Frank,

    A lot of interesting info, thanx!

    I intend to replace my old Minolta IV lightmeter with a Sekonic, but can’t decide between two models: the 758DR with RT-32CTL and the new 478DR. The advantage of the 478DR seems to be the ability to adjust the Canon Speedlites mounted on PocketWizard FlexTT5 from the light meter directly, while on the 758DR you have to do it via the PocketWizard AC3 – quite awkward, need 3 hands for that… Also, the 478DR stores 10 presets as opposed to 3 presets on the 758DR.

    I would like very much to hear your opinion and recommendation on this.

  8. Dinora
    Dinora says:

    So basically I can save for any three of the four reflective, incident, flash and ambient (which is, I presume, actualy available light and can include the sun for outdoors or lightbulbs for indoors), but not all four?

    By “Leaf profiles” you mean profiles stored in the light meter when calibrated using the Leaf?

    How can I check the numbers in PhotoShop and Capture One?

  9. Dinora
    Dinora says:

    I also presume that “3 exposures with a 3 stop difference means” -1, 0, +1, not -3, 0, 3 and even that isn’t what the quoted line means. I am nitpicking, but this is important. What colourspace have you chosen?

    • Frank Doorhof
      Frank Doorhof says:

      As in the manual from Sekonic and in the software is mentioned it’s 3 stops so -3,0,+3
      Colorspace you choose the colorspace you normally work in, so for me the conversion was done with Photoshop in ProPhotoRGB

    • John Robinson
      John Robinson says:

      Hi Frank, If you are doing the conversion of the test images using the ProPhoto colour space, shouldn’t you set the mid tone in the sekonic DTS software to 99 (ProPhoto mid tone). In the sekonic DTS manual is states the 118 value is for sRGB colour space which would be correct. I may be wrong here but there’s two ways to do this…

      1) Convert test images using ProPhoto colour space but set default values in the Sekonic DTS software to 99 for mid tone.
      2) Convert test images using sRGB colour space and leave default values at 118 in the Sekonic DTS software.

      I have searched for clarification online but have yet to find any, I may call Sekonic for answers unless you know better?

    • Frank Doorhof
      Frank Doorhof says:

      I normally calibrate on a 18% gray card. And get it to read middle gray. When using the 128.128.128 in the raw convertor everything comes out fine. And accurate.

      Remember that it gets it colorspace when you export.

      This is a bit of a problem I have with the dts software so I prefer to keep it in the convertor i use and the tone curve I export in.

  10. Russ
    Russ says:

    Hi, Like you I use the Colourchecker to calibrate and correct colour but find that the grey patch is very small to get a reading from so does it need to be X-Rite or Sekonic’s idea of 18% grey or can you use another reputable maker of an 18% grey card to spot from?
    I also understand that the grey patch on the new Colourchecker that comes with the 478 is actually closer to 20% grey and not 18%???
    Thank you
    Russ

    • Frank Doorhof
      Frank Doorhof says:

      Hi,
      I think you misunderstood me.
      The colorchecker should not be used in the calibration process as an 18% graycard, when you use it, use the sekonic software with it, not alone.

    • Russ
      Russ says:

      Hi, Thank you for your reply.
      Yes I no how to calibrate the meter (L-758) with the CCP and Sekonic software but the grey patch(4th in from the left on the bottom row) is very small to get a reflective exposure from even with the spot meter so what I am asking can I use an 18% grey card from another manufacturer to take that reflective reading then the rest do with the CCP. I no you can buy the Sekonic grey card on it’s own but they are also quite expensive and I already own the Lastolite EzyBalance 12 and 18 % grey cards.
      Hope you understand the question.
      Again Thank you
      Russ

  11. Russell Spears
    Russell Spears says:

    I have a bit of a confusion…

    I am part of the few that need to understand everything about photography-with this predilection I was hesitant but went ahead and I got a Sekonic L-358 light meter. This whole 12% to 18% Grey issue (at least for the particular model I have in my hand) had to be settled in my head. So I set up my own test. But my results are way different and I can’t figure out where it all went wrong. My goal is to be able to precisely set every key value in a shot to a corresponding luminance value the final Adobe RGB file.

    Here are my working assumptions:

    1. A reflective reading and an Incident reading tell us two different things about middle grey: Incident meter readings tell me what is Middle Grey under this lighting conditions. Reflective meter readings tell me what this object meters if it was middle grey.
    2. What ever tone the object of light we meter, the meter will give us a middle Grey Value in terms of ISO, F/Stop and Speed.
    3. A any consistent tone, 18% or 12% Grey Card, Skin, Sky, the bright side of a model’s ass-even a 5100k 2.2 Gama bright monitor on an iMac, will all be converted to a meter’s middle grey.

    Here is how I tested my meter and camera:

    Rather than introduce more variables like flash or worse ambient light, I thought why not use the brightest and most consistent light I have. I used my iMac screen (yea laugh) but at full brightness I had a larger dynamic range than flat paper. I took meter readings to see how even the light was across my screen and it varied a bit towards the edges (0.1 to 0.2 tenths of a stop variance)-but it did not vary over time so I turned off all the lighting and even removed a mirror behind me to avoid any reflections from my screen back onto itself.

    Now-with the lights off, the screen was metered with the three attachments I have: both my 5 Degree View Finder and Lumigrid attachments gave me at ISO100, f8 at 1/30 sec and my Rebel t3i gave me this same value too after I calibrated the meter +0.2 tenths for the camera’s apparent difference. However the Lumisphere was off by 3.7 Stops (I am guessing that the meter’s “incident” reading adjusted the same light source for a middle grey). Thinking everything in the Sekonic universe is making some sense, I went ahead and did the test shots.

    I did a series of shots in full stop increments from 1 sec to 1/6000 sec at f/8.2. and this was my results when brought into Camera Raw with all adjustments set to 0 and I cropped the image down to the very middle 2 inches-which I took all the measurement from, then I brought the images into Photoshop and checked the mean value in the info palette for each. Here was the results:

    1 sec= 255
    .2 sec=255
    .4 sec=225
    .8 sec=158
    .15 sec=112
    .30 sec=76
    .60 sec=52
    .125 sec=35
    .250 sec=25
    .500 sec=19
    .1000 sec=12
    .2000 sec=7
    .4000 sec=4
    .8000 sec=0
    .16000 sec=0

    So from this my Sekonic Meter’s Middle Grey Seems to be set at 76 or 80 on the Grey Scale not 128 or 119. Now from this post, I can see why there is variation and dispute. But do you or anyone see the glaring mistake that pointed to 76 as middle grey?

    • Frank Doorhof
      Frank Doorhof says:

      If you are testing this please use strobes, or constant light and a proper 18% graycard for calibration.
      After you have calibrated the meter to read 18% gray from the card it should function properly.
      If you’re metering the screen from a mac you are metering a reflective surface, that can actually throw in some variables if you’re metering under angles etc. Normally 18% gray should be around the middle of the grayscale.

      I know there are a lot of discussions about this, but over the years I’ve found out that when I calibrate my meter to read both incindent and reflective to be 128/128/128 in my workflow (in this case Lightroom) my exposures are spot on. I do have to add that I don’t add any curves etc. to my RAWs before I retouch. For example in Capture One the outcome is different, and my files are actually all underexposed in the blacks but correct in the Whites, meaning my middle point will also shift. If I calibrate in Capture One I will get a different calibration and varying results during the shoot, UNLESS I use a flat profile.

      I don’t really understand what you did with the monitor but normally you should test with an 18% graycard under flat lighting and get that reading.

      So please take that into account.

  12. Sunshinez
    Sunshinez says:

    Hi Frank,

    Fantastic article and really helpful. Just wondering if you can help. I have just bought the L-758 and the X-Rite Color Checker Passport. Do I also need the QPcard 101, in order to do the 18% grey reflective exposure calibration?

    I have seen a video and read the DTS Software guide, where you are told you can spot meter the small grey patch (3rd from the right) on the CCP, but I was wondering what would I do about a reflective reading?

    Just wondering what you thought about this.

    Thanks,

    D Cooper.

    • Frank Doorhof
      Frank Doorhof says:

      The QPcard is 18%, the Colorchecker has no 18% gray patch.
      However the software from Sekonic does work with a color checker so you can use that.
      For reflective it also states which patch to use indeed, it’s a very simple procedure.

  13. Rama Kumar
    Rama Kumar says:

    Hello Frank, thank you very much for this very informative post. I recently acquired a L-758DR. I have been using CCP for quite sometime. I shall now use CCp to calibrate my 758DR. It’d be really good if you can put together either a video or an book not only talks about Calibration of the Light meter but also about the right usage of the meter in different shooting scenarios. Thanks once again for taking time to share your knowledge.

    • Frank Doorhof
      Frank Doorhof says:

      Can be in a few sentences :

      1. Calibrating.
      Get an 18% gray card and calibrate in your software/colorspace to 128.128.128 (or close)
      Or use the software with the CCP.

      2. Pointing it.
      Towards the light hitting the area you want to be lit correct, that’s it.
      In the studio mostly towards the light source.

      On the right side you find a menu on light meters, there are articles there that will help.

  14. bukboy
    bukboy says:

    Hi Frank

    About your suggestion to calibrate a light meter to 128.
    Ive just completed my own calibration and am contemplating why my respective 0 ev level is rendered at 118 pixel luminosity in the configuration DTS profile for my 5dMkii.

    DR+ 245 +3.1
    CP+ 230 +2.9
    MT 118
    CP- 35 -2.0
    DR- 20 -2.9

    I suspect that what this means is that my camera actually really really renders 18% gray at 118 (and not 128 as I have been misinformed to believe by the majority out there) because the sensor dynamic range curve is not strictly linear, and hence not symmetrical around the 0ev point.

    After having verified that indeed my highlights get blown out exactly 2.9Ev and above the mid tone exposure (that I shoot with) I am progressively inching towards the conclusion that calibrating a meter for 128 is a bad idea at last for my combo.

    And the reason for this is – my camera packs the top 2.9ev (it can record) into top 112 pixel levels above 118 (38.6 pixels per ev on average), and packs the bottom 2 ev (it can record) into the bottom 83 pixel levels below 118 (41.5 pixels per ev on average).
    If I choose to offset the middle point, this will not change the fact that my sensor has a curve that is a little left of the theoretically sold 5 ev zone models. So if I do I will find that the meter will be telling me to under expose by .25 ev and hence gain somewhere around .25 ev headroom in the highlights and lose around .25ev in my shadows

    Well either that or sekonic has my mind in a twist. Or my calibration is wrong, or the software is buggy. etc etc etc.

    But I would imagine that this is not accidental.I would think that canon would want more room to overexpose at lesser detail, while having greater detail in the underexposed sections. In which case again, the 128 calibration is probably a bad idea.

    Do you have any thoughts on this?

    • Frank Doorhof
      Frank Doorhof says:

      There is a difference.
      When I calibrate the 128 on an 18% graycard (not colorchecker) I do this in the raw convertor. Never leaving the convertor.

      In all cases my exposures are spot on. When I use 118 they are not. There is indeed a lot of discussion about this. I can only say that this works for me.

      Also it makes sense because an 18% gray card should be…. well 18% gray when you shoot it.

Comments are closed.