Exodies Posted December 12, 2018 Share #1 Posted December 12, 2018 Advertisement (gone after registration) Can someone please explain the gaps in this histogram. The lighting is LEDs (I think), certainly not full spectrum. Welcome, dear visitor! As registered member you'd see an image here… Simply register for free here – We are always happy to welcome new members! M typ 240, Apo-Summicron-M 1:2/50 ASPH. Quote Link to post Share on other sites Simply register for free here – We are always happy to welcome new members! M typ 240, Apo-Summicron-M 1:2/50 ASPH. ' data-webShareUrl='https://www.l-camera-forum.com/topic/292242-histogram-and-leds/?do=findComment&comment=3646026'>More sharing options...
Advertisement Posted December 12, 2018 Posted December 12, 2018 Hi Exodies, Take a look here Histogram and LEDs. I'm sure you'll find what you were looking for!
wda Posted December 12, 2018 Share #2 Posted December 12, 2018 (edited) The histogram shows a relatively low-key exposure with a slight bias towards the red channel and quite a bit of underexposure. If that is how you saw the scene, fine. As a detached viewer I would be tempted to adjust exposure and then use LR HSL sliders with eye dropper tool to make further slight adjustments to exposure and colour saturation. Easily done without upsetting the original file. The process is non-destructive so experiment until you are happy. Not too sure what you mean by 'gaps' unless you refer to the lack of pixels on the right side (discussed above) . Colour channels often show up this way but hint at what corrections are possible. Edited December 12, 2018 by wda Quote Link to post Share on other sites More sharing options...
Exodies Posted December 12, 2018 Author Share #3 Posted December 12, 2018 Thanks for the analysis but I was talking about the feathering along the top of the histogram. Quote Link to post Share on other sites More sharing options...
pgk Posted December 12, 2018 Share #4 Posted December 12, 2018 (edited) Whilst LEDs might be described as 'full spectrum' I'm not so convinced. It does look as though many are pretty good but some are still uneven in their output throughout the spectrum, which shouldn't really be that much of a surprise. I haven't looked but I'm sure some research would show manufacturer's claimed spectral outputs and some will no doubt be better at supplying a continuous output throughout all colours than others. Probably their are ISO 'standards' by now and 'photographic' output LEDs will have to fulfil these (I haven't looked). I have though been involved in looking into lighting for a specific system and erred on the side of caution and recommended LED tubes designed to replace fluorescents (in situ) which are for use in the tv/film industry (both are quite picky) and these seem to give a pretty good continuous output. Others, and especially those intended for use in domestic/office situations, are probably designed to be reasonable rather than perfect so may well exhibit gappy colour distribution. They are better than many fluorescents though! Edited December 12, 2018 by pgk typo/grammar Quote Link to post Share on other sites More sharing options...
frame-it Posted December 12, 2018 Share #5 Posted December 12, 2018 56 minutes ago, Exodies said: Can someone please explain the gaps in this histogram. The lighting is LEDs (I think), certainly not full spectrum. Welcome, dear visitor! As registered member you'd see an image here… Simply register for free here – We are always happy to welcome new members! M typ 240, Apo-Summicron-M 1:2/50 ASPH. is this the un-edited file? if not, please post a jpeg of the un-edited one https://forum.luminous-landscape.com/index.php?topic=41550.0 Quote Link to post Share on other sites More sharing options...
Exodies Posted December 15, 2018 Author Share #6 Posted December 15, 2018 Unedited histogram Welcome, dear visitor! As registered member you'd see an image here… Simply register for free here – We are always happy to welcome new members! Lightroom CC on iPad Auto tone applied Lightroom CC Classic Auto tone applied These are not from the same picture as the original post, but this picture was taken with the same light. Quote Link to post Share on other sites Simply register for free here – We are always happy to welcome new members! Lightroom CC on iPad Auto tone applied Lightroom CC Classic Auto tone applied These are not from the same picture as the original post, but this picture was taken with the same light. ' data-webShareUrl='https://www.l-camera-forum.com/topic/292242-histogram-and-leds/?do=findComment&comment=3647815'>More sharing options...
Exodies Posted December 15, 2018 Author Share #7 Posted December 15, 2018 (edited) Advertisement (gone after registration) This is the unedited picture from the three histograms above. The original histogram is from Lightroom CC Auto tone on iPad. Welcome, dear visitor! As registered member you'd see an image here… Simply register for free here – We are always happy to welcome new members! Edited December 15, 2018 by Exodies Quote Link to post Share on other sites Simply register for free here – We are always happy to welcome new members! ' data-webShareUrl='https://www.l-camera-forum.com/topic/292242-histogram-and-leds/?do=findComment&comment=3647817'>More sharing options...
zeitz Posted December 20, 2018 Share #8 Posted December 20, 2018 Since no one else, such as Jeff or Jaap, is making a statement based on actual knowledge, I will throw out a guess for members to criticize. First I recall that spikes in the profile can result from over processing of a low bit count (8-bit) and high compression (jpeg) image. I know there are many advocates who say that .jpg is all you need; others disagree and point to histograms with spikes that result. You don't say whether you are working with .dng or .jpg files. Second I am guessing that the autotone algorithms in Lightroom are different for Lightroom on a full-spec computer vs Lightroom on a reduced-spec computer. We don't know what Adobe did to resize a very large MacOS / Windows 10 program for an iOS device with limited processing power. I know Photoshop CC is coming soon to iPads. We will see. Saying that I don't know what I'm talking about is a fair comment because I almost always work with raw files in Photoshop CC and then only on a MacPro. Attached is a histogram from a .jpg file I just opened. It was taken with a normal electronic flash. The only adjustment I made in Photoshop CC is contrast and brightness in a layer. The as-shot histogram is smooth. Welcome, dear visitor! As registered member you'd see an image here… Simply register for free here – We are always happy to welcome new members! 1 Quote Link to post Share on other sites Simply register for free here – We are always happy to welcome new members! ' data-webShareUrl='https://www.l-camera-forum.com/topic/292242-histogram-and-leds/?do=findComment&comment=3650260'>More sharing options...
pgk Posted December 20, 2018 Share #9 Posted December 20, 2018 (edited) Any adjustment which 'stretches' data deficient images (such as those taken under non-contimuous spectra) risks producing gaps, which is exactly what the adjustments are doing. Greater than 8-bit files don't make up for lacking data even if they may be a little more tolerant. Edited December 20, 2018 by pgk 1 Quote Link to post Share on other sites More sharing options...
Exodies Posted December 20, 2018 Author Share #10 Posted December 20, 2018 I’m only editing with DNG. Quote Link to post Share on other sites More sharing options...
pgk Posted December 20, 2018 Share #11 Posted December 20, 2018 But editing a DNG and saving the edits as an .xmp file or whatever, will still adjust the histogram and reveal areas lacking in data, and that seems to be what's happening. Quote Link to post Share on other sites More sharing options...
zeitz Posted December 20, 2018 Share #12 Posted December 20, 2018 Paul, I'm beginning to think there is an Adobeism going on here. There is that little caution emblem with an exclamation point in it, at the top right of the histogram. Hovering over it, the symbol is described as "click here for histogram with uncached data". When I click on it for most files, the spikes completely go away. In some cases a few spikes remain. We may never know what Adobe is doing to facilitate processing and display. Quote Link to post Share on other sites More sharing options...
pgk Posted December 20, 2018 Share #13 Posted December 20, 2018 (edited) The histogram doesn't seem to update automatically for some, no doubt, very good reason. A problem with these histograms is that they are relatively small and can't do much more than indicate what's going on. It would be interesting to find out whether the people responsible for the histogram display make it accentuate any potential problems (spikes) because its difficult for it to show any absolute differences except perhaps in heavily processed 8-bit files. All that said they remain a useful tool but one which has to be used in conjunction with the image on screen. In this thread the images posted do look somewhat colour deficient and the histogram suggests this too. A link posted on a Large Format forum is worth having a look at: https://www.waveformlighting.com/high-cri-led?gclid=EAIaIQobChMI7P63q_ed3wIVS77ACh3cBQRqEAAYASAAEgKr8PD_BwE Edited December 20, 2018 by pgk Quote Link to post Share on other sites More sharing options...
adan Posted December 21, 2018 Share #14 Posted December 21, 2018 (edited) 1) I don't think this has anything to do with LED lighting particularly. LED output is not especially "gappy" or "spiky" (certainly far, far less so than fluorescent or other gas-discharge light, and mostly in the IR range - 700+ nanometer wavelength). http://blog.durablescope.com/post/BuiltASpectrometer/ ...and that assumes that homebuilt spectrometer isn't buggy in the IR. 2) a histogram is not a spectrogram, any more than a car fuel gauge or tachometer is a speedometer (even though the "displays" may be identical needles-and-scales). A spectrogram graphs wavelengths of light (as a special case of a histogram), a photo histogram graphs brightnesses of a single pure primary "wavelngth" (red, green, or blue) 3) a picture histogram is going to be heavily influenced by the subject colors, not just the "white light" color. The shirt and skin and hair colors are going to impose their own values on top of whatever the light is producing, as far as the data in the pictures is concerned. If anything, that will soften or blur any spikes. 4) Histograms always collapse the actual data to 8-bit for display purposes - because as pgk says, it is hard enough to see and distinguish 256 individual columns of pixels, let alone 65536. I have to see this as just a quirk of the down-converting from 12/14-bit image data to 8-bit display data*, combined with quantization in the display processing ("draw a green pixel here, and another green pixel here ") that builds the picture of the histogram on the viewing device. An "Adobeism" as zeitz calls it. * or simply that the software uses the embedded 8-bit preview/thumbnail .jpg within the .DNG to draw its "this many pixels have a red value of 182 and this many have a value of 183" graph. What happens if you put the card back in the camera (or if the original picture on the card has been deleted, put the .DNG back on the card in the correct "Leica 10x" folder). How does the camera's display software render this picture's histogram? Edited December 21, 2018 by adan Quote Link to post Share on other sites More sharing options...
pgk Posted December 21, 2018 Share #15 Posted December 21, 2018 3 hours ago, adan said: I don't think this has anything to do with LED lighting particularly. Take a look at the photos. Quote Link to post Share on other sites More sharing options...
Exodies Posted December 21, 2018 Author Share #16 Posted December 21, 2018 (edited) Before getting round to adan's suggestions: Histogram, generated in Lightroom CC on iPad and displayed in Lightroom Classic CC Welcome, dear visitor! As registered member you'd see an image here… Simply register for free here – We are always happy to welcome new members! Histogram above after being recalculated in Lightroom Classic CC (Reset button) It looks like iPad LR only calculates every other value (or combines the counts for odd and even intensity values) and the histogram is stretched for display in Classic. 1 hour ago, pgk said: Take a look at the photos. I think adan was referring to the feathering of the histogram, not the limited colour range here. Edited December 21, 2018 by Exodies Quote Link to post Share on other sites Simply register for free here – We are always happy to welcome new members! Histogram above after being recalculated in Lightroom Classic CC (Reset button) It looks like iPad LR only calculates every other value (or combines the counts for odd and even intensity values) and the histogram is stretched for display in Classic. I think adan was referring to the feathering of the histogram, not the limited colour range here. ' data-webShareUrl='https://www.l-camera-forum.com/topic/292242-histogram-and-leds/?do=findComment&comment=3650641'>More sharing options...
pgk Posted December 21, 2018 Share #17 Posted December 21, 2018 51 minutes ago, Exodies said: I think adan was referring to the feathering of the histogram, not the limited colour range here. Indeed, but most photography is finally about the image not the technical aspects of its images which simply reflect what you can see and interpret it into a numerical or graphical explanatio Quote Link to post Share on other sites More sharing options...
adan Posted December 21, 2018 Share #18 Posted December 21, 2018 Are we talking about the same thing? The first sentence of this post is "Can someone please explain the gaps in this histogram?" The only "gaps" I see are these: Welcome, dear visitor! As registered member you'd see an image here… Simply register for free here – We are always happy to welcome new members! ...which as Exodies' followup checks demonstrate, are just an artifact of the various LR versions means of creating and displaying the data. I don't see any other gaps. I do see a normal histogram for uncorrected indoor lighting, which is usually somewhat red (regardless of whether it is LED or incandescent bulb) compared to sunlight. LEDs come in all colors, but those used for indoor lighting are intentionally biased reddish to replicate what we are used to after 120 years of incandescent lighting. Such LEDs (as with regular bulbs) introduce a need for white-balancing (to taste). They don't produce gaps. Quote Link to post Share on other sites Simply register for free here – We are always happy to welcome new members! ...which as Exodies' followup checks demonstrate, are just an artifact of the various LR versions means of creating and displaying the data. I don't see any other gaps. I do see a normal histogram for uncorrected indoor lighting, which is usually somewhat red (regardless of whether it is LED or incandescent bulb) compared to sunlight. LEDs come in all colors, but those used for indoor lighting are intentionally biased reddish to replicate what we are used to after 120 years of incandescent lighting. Such LEDs (as with regular bulbs) introduce a need for white-balancing (to taste). They don't produce gaps. ' data-webShareUrl='https://www.l-camera-forum.com/topic/292242-histogram-and-leds/?do=findComment&comment=3651008'>More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.