First Impressions now using Dolby Vision LLDV:If you see the LK990 set to HDR Brightness of -2, color at ~7 and Color Enhancement at ~16, paired with an HDFury Vertex2 doing the LLDV/Dolby Vision tweak with a DV String set to Max Luminance (nits) of 10,000 nits, or anywhere between 2,000 - 10,000 really, then you certainly will be saying "Wow"!I can indeed get an absolutely incredible HDR/Dolby Vision image on screen with this combo.If anyone says brightness doesn't matter, then you render the great flat panels that can do that utterly moot. As anyone who actually knows what HDR is and does, the main and almost entirety of the movie is still within the 100 nit max range, or at least is supposed to be mastered that way, the same as its SDR counterpart. The extra brightness is reserved for specular highlights, etc. So by using a lower nit projector you're just forcing the display, source or tone mapping device/processor to compress that range available further down to whatever it has for its peak nits. No compression is always better than any compression. Tone Mapping is compression, but I am sure you already know that, so I guess that's for the newbies too.With this higher brightness LK990, I still have the same original 100 nits brightness for the main part of the HDR video, but also the extra brightness over the lower nit projector is now there for what is easily seen as amazing specular highlights and incredible "HDR Effects", like you see on some of the best HDR/DV flat panels, which haven't been compressed by tone mapping as much or at all.The laser levels in this projector are able to be lowered based on what is input, using the unit's Automatic Power Control and SmartEco features, similar to what is reported with the much more expensive JVC RS4500, and why many prefer its black levels and shadow details over the reportedly better RS3000 (@markmon for example). When you input and present a 10,000 nit Dolby Vision custom string using the HDFury Vertex2/DIVA/Maestro with a BT2020 color gamut along with setting the LK990 to an HDR Brightness setting of -2, color at around 7, Brilliant Color ON and Color Enhancement to about 16. This combo makes this happen along with great color tracking along the color gamut sweeps, unlike when you use default color at 50 and Color Enhancement OFF and Brilliant Color ON.Couple that with all the things that make the image much more detailed than even the mighty RS4500, as easily seen in the image attached, by @Woofer.And no, I am not going to go around and around with anyone again here. I have it here and have for a LONG time, doing extensive testing and configurations with many different sources and signals. Nobody else has. I know what it can and can't do, Nobody else does. And yes, I do agree the one area it can't compete totally is native on/off contrast and a little color saturation, and I have always said that. But I have also always said all the other attributes this presents are far superior, and I still completely stand by that. I can all but guarantee anyone else who sees what I see here with Dolby Vision would feel the exact same way too.The complete stubbornness to not want to really dig deep and understand this projector and what it can do never ceases to amaze me. I don't understand why. Who does it hurt? You all just keep living with old thoughts and paradigms that simply don't apply here.Please all be safe and healthy in these trying times!https://discuss.avscience.com/index.php?action=dlattach;topic=1924.0;attach=4867
Dave, you keep trying to make HDR all about brightness. When the very name is High Dynamic Range. That is not referring to brightness. If brightness is the end all be all for HDR, then the LCD flat panels would easily provide the best HDR image and yet it is clearly undisputed that an OLED TV provides a much better HDR image and the OLED TV's have less than half the brightness of some of the LCD TV's. What is it that an OLED has over an LCD flat panel, it is contrast.
Thanks for the info Dave. I think that all the owners of this unit that are not using DTM should definitely see a huge benefit from this mod. I have had an LK series unit for over 2 years and I still enjoy the unit greatly. I think that the room, content and personal preference all factor in when evaluating this unit or any other display. Prior to DTM, the LK's brightness was a huge advantage in dealing with a static map. With the introduction of DTM, the frame by frame mapping has evened the playing field a bit for projection units of standard brightness. Yes, the black floor of these units measure poorly, but the room environment and content will reveal how much of a problem that actual is IRL. I am using my unit probably 40% movies, 30% gaming and 30% sports.
Sorry to hear about your granddaughter. We had both our son and daughter with broken arms at the same time. I was not the one looking after either of them during their incidents.
Why is this answer ridiculous? back years ago, before Shifting DLP, when on off contrast for DLP was higher, the recommendation for anyone that was serious about image quality, they were advised to turn off DB. Javs was seriously interested in the LK990, because it was one of the few projectors that might have had enough light output to light up the size screen he was wanting and in his budget. Javs is pretty serious about image quality and could not live with the artifacts DB introduced, so he reviewed the projector without DB on. Each person has to decide if the benefits of BD outweighs the cost. ........
I have added a note, correcting my statement about Jav's turning off BC, rather than DB.
Added 7/18/2020It was pointed out to me that it was Brilliant Color in the LK990 that Javs turned off, not dynamic black. The 990 does not have dynamic black in the regular menu. Brilliant Color is another bit of processing that should be turned off, if you want an accurate picture.
I have found if you greatly reduce the main color control to less than 10 (I use about 7) and then go the Advanced Menu->CineMaster->Color Enhancer and set it to about 14-18 (I use 16) then the color tracks very nicely with Brilliant Color ON.
Since the only way you can possibly know that is by measuring it, hence you have the calibration file on your PC, can you please share the colour volume tracking results for any primary colour, say, green?If you are correct, your deltaE for Y should be reasonable.For reference, here is BC Off. This is Rec709 mind you... Note the two red circles, Y Target and Y measured. They are close enough to consider reasonable that there is no issue.Official BenQ LK990 Owners ThreadNot like this:BC ON, which from a standpoint of how the technology works based on what I read about it and even what you yourself have agreed, its my understanding its NOT possible to have the green with BC On reach the target Y which as you know is a target in relation to the measured white. If you have found a way around that, then that's great! But you should be easily able to show us so we can see the measurements.Official BenQ LK990 Owners ThreadAlso, note how much better tracking is overall with BC off, but also note that to the untrained (Not saying you at all) the tracking on BOTH appears to be somewhat reasonable (excluding Magenta) if one was looking only at the CIE chart, however the colour volume in comparison to the measured peak white takes a shit when BC is on. Hence, unusable to me.
......WRGB OLEDs come to mind due to their white pixels being mixed in with the RGB, where you have to lower Y luminance so much to get full 3D color luminance measured properly. Javs saw this and complained that he had to turn off Brilliant Color, which I disagree with when done properly in the end. These projectors were engineered from the ground up, taking this into account, as the Delta/Vivitek and Optoma engineers told me. This is the same with the WRGB OLEDs. The LK990 with it's RGBY (Y being derived from the phosphor wheel, excited by the blue lasers) imparts similar issues, as you, me, Kris and Javs have all noted when measuring the LKs.
This is exactly what I measured on the lk990. It's colour volume is crap. Dave Harper told me the 970 didn't suffer it, quite adamantly at that but never substantiated it with measurements. It just shows he was wrong also.On the lk990 you can turn off brilliant colour and measure correct values but the light output is 40% lower. Seems like you do indeed have brilliant colour hard coded on. Everyone (mainly Dave) was always telling me I was wrong. Thanks for showing I was not.Sent from my SM-G988B using Tapatalk
Yah, the bar is low given what the LK990 did, and to be frank, if its even similar, for the price its quite good.
QuoteQuoteQuoteQuoteQuoteQuoteIf the narrower band wavelength of the Leds is causing the chromatic aberation i guess the lens isnt the deciding factor, the Lk970 has 93-94% rec709, the M150 has 90% Bt2020 so that is a huge difference lol, and might be causing it.QuoteI have tried calibrating 3 different LK970 multiple times, but I have never got more than ~64% rec709 gamut coverage.QuoteIt's always super short in luminance, like 30-50% too low depending on the color and that leads to a coverage of only about 64%, despite the saturation and hue on the 2D CIE chart looking good.Have you actually measured the LK970 at >90% rec709?QuoteYes i measure it in calman, i can re-measure it now if ud like and share it hereQuoteThis is just now, With stock settings, normal laser.QuoteHmm. I mean yeah my 2D CIE looks just like that, but when it factors in the luminance, it falls way short.QuoteWhat are the luminance percentages of reference of the colors in your measurement?I was using DisplayCAL to measure my gamut.Do you get a similar gamut coverage after a 3DLUT?QuoteWhen i create 3dlut with displaycal or lightspace yes you are right it says 64-65% gamut. But how come is there that much difference between calman and lightsapce/displaycal.QuoteOn the other hand my Sim2 M150 measures even better in Lightspace/Displaycal compared to calman where it already measures great to begin with.I m getting 89.7% bt2020 stock settings , and 93-94% bt2020 in Lightspace/ Dispalycal so i guess every program measure it in its own way.QuoteThe reason for the low gamut % with 3DLUT in DisplayCAL is because the luminance values are super low and the 3DLUT is trying to correct them but it physically can't because the projector just can't do it.QuoteI have never used CalMAN so I am not sure really much about what it's doing or reporting.In your last picture, look at the Y vs. Target Y for your colors. The Y is way too low compared to the target. Remember, color gamut is a 3D volume, not just a 2D triangle. You can reach way out to the proper "saturation" level of a color which you see on the 2D CIE diagram (x and y), but still be way short on luminance (Y), and the gamut coverage therefore shouldn't be considered 100% or close to 100% if the luminance is way too low.[MEDIA=youtube]BQGbbobBsiw[/MEDIA]You can have a 2D CIE that looks like the corner points are perfect, but still be way short on luminance.[/size][/color]QuoteAh i see , i guess the luminance on the LK970 is really lacking.QuoteI took another measurement of the sim2 m150 but with proper settings because it was just default before and as u can see now the color luminance is almost spot on.This is with bt2020 and a 3dlut will make it exactly on point.The LK970 on the other hand cant get the luminance close, i just tried and i guess its a limitation.QuoteYeah your luminance values look fine.QuoteIt's just some projectors such as the LK970 that fall way short for some reason.I have seen the issue on a couple other DLPs as well. Never seen it on an LCD or LCoS though. But of course not all DLPs are like that at all, just some.And yeah you should definitely check out the theater fan/wind project, it's really awesome. Less gimmick than you might think (when done right). Could use some more users. I'm still improving and expanding things on the project. I hope it can really take off more once I make the fans easier to use and more generic to support all video sources.QuoteYeah, I will admit that I don't really know how the luminance is exactly supposed to play into the "gamut coverage" calculation and number. I am no expert.QuoteBut with the luminance values being so low, the actual vs. target colors (and deltaE 2000) are pretty far off and so I think perhaps the full gamut coverage is quite low, but again I am not sure exactly how to "properly" report that.QuoteI dont think 3chip dlp would have this problem and typically rgbrgb wheels do ok, its wheels other than rgbrgb or when the diodes arent to spec, like that panasonic pt-rz470(red and green leds with a laser/phosphor for green) has good red saturation but low luminance and less green saturation but luminance is too high.QuoteIt sounds awesome im absolutely going to look into it. It didnt come across as gimmickery at all... Tactile sensation and immersion is real. A gimmick is like... 1,000,000 dynamic contrast!QuoteI think it happens on RGBW wheels. Do we know if the LK970 is RGBW? It's because the white part messes with the proper color luminance I think.QuoteI meant like is the LK970, 94% rec709, or is it 64% rec709?QuoteIf you look at the chromaticity (x,y) and the 2D CIE diagram, it seems like it's 94% coverage.But if you look at the luminance values for the colors on the LK970, they are far short, upwards of 50% short in the worst cases and this leads to large deltaE 2000 errors and leads software like DisplayCAL to give you a number like 64% rec709 coverage when measured.I am not sure how this gamut coverage number is exactly supposed to be calculated and reported and how the luminance should be factoring in.Quotethere are tactile transducers/bass shakers which provide rumble and dbox which has lift roll yaw pitch motion. I have to temper my subs response so i isolate them off the floor with sorbothane then use bass shakers to replace the tactile rumble. Works really well for preventing low frewuencies from getting out of the room, persons with sub bass sensitivity, while still providing good immersive feedback. QuoteWould love to incorporate dbox and wind, but dbox wont work, so the idea that i can add wind is really cool...not sure, there are an array of color wheels and in some instances 2 color wheels. rgbcym, rgbcy(which is what led overlap does), rgbw, rgby, rgbrgbg orsometimes a dark green segment for p3 gamut.QuoteBlue core lasers are typically RGBY, so I think the 970 is similar.QuoteActually I think the LK970 is RBGY and the LK990 is RGBW so the LK990 might be even worse.QuoteI think it also has to do with BriliantColor on the LK990 because with that off color accuracy is good, but the brightness is very low, lower even than the LK970. With BriliantColor on the color is not able to be calibrated to any sort of low dE.QuoteIts one of the reasons why i get posterization when doing a 3dlut on the LK970 , its very obvious, the sim2 doesnt have any at all.QuoteI think its the main reason why the sim2 look bright, cause it has a lot of range for color luminance .QuoteYeah I do think it's why there is so much posterization. QuoteBecause it's trying to take a 64% rec709 to a 100% target.I did find that using a different rendering intent in ArgIICMS minimizes the posterization by a lot, and guess what that rendering intent does / is called...“Luminance preserving perceptual” (ArgyllCMS 1.8.3+) uses compression to make the source gamut fit within the destination gamut, but very heavily weights the preservation of the luminance value of the source, which will compromise the preservation of saturation. No contrast enhancement is used if the dynamic range is reduced. This intent may be of use where preserving the tonal distinctions in images is more important than maintaining overall colorfulness or contrast.I don't think it's a coincidence that the rendering intent that sacrifices saturation for luminance and that uses compression produces the best result on the LK970.QuoteThe thing is also when i do a p3 3dlut with my jvc rs440 whcih does 90%p3 i get posterization but to a lesser degree than the LK970, a lot less actually but its there and can be seen in some scenes.QuoteWith the Sim2 M150 i create a bt2020 3dlut and no posterization whatsoever even though the projectoe can do 90% bt2020 not 100% , i guess the RGB LED is a different animal.There is a workaround in LightSpace called Concatenation method, and DisplayCal also has good solution for it but at the expense of saturation like u said, i just dont know how come the Sim2 has absolutely zero posterization.QuoteI talked to someone who is much more knowledgeable on 3DLUTs about it once before and they said posterization is not really caused by targeting a larger than native gamut.QuoteFor example I tested targeting P3 on my rec709 PC monitor which only does about 78% P3 and I saw no posterization with the P3 3DLUT.It's definitely more complicated than that, but unfortunately is past my current understanding of color systems.QuoteThis is exactly what I measured on the lk990. It's colour volume is crap. Dave Harper told me the 970 didn't suffer it, quite adamantly at that but never substantiated it with measurements. It just shows he was wrong also.QuoteOn the lk990 you can turn off brilliant colour and measure correct values but the light output is 40% lower. Seems like you do indeed have brilliant colour hard coded on. Everyone (mainly Dave) was always telling me I was wrong. Thanks for showing I was not.Sent from my SM-G988B using TapatalkQuoteIf I had to describe it based on trying to calibrate 3 LK970 both manually and with 3DLUT, I would definitely describe the behavior resembles that of BriliantColor and what happens when you have a RGBY or RGBW color wheel.QuoteI also encountered the same behavior with this projector: https://www.projectorcentral.com/Optoma-ZH403-1080p-Laser-Projector-Road-Test.htm"]Optoma ZH403 1080p DLP Laser Projector Road Test[/URL]It had far-off luminance, but with BriliantColor off I could get low dE, but at less than 2000 lumens on a 4000 lumen spec machine.[/size][/color]QuoteOn the 970, I am not surprised that BC is hard-coded in. My thought was that BC was always there and adjusted to provide more brightness for the LK990. Once you have seen or measured enough units you can spot BC a mile away. I can see it in the rendering of whites. RGB Laser or high lumen LED is the path forward for DLP in my opinion.QuoteBrilliantColor, amazing name for a feature that f**ks up colors! :cool:QuoteI believe my Hitachi RGB LED DLP has something similar baked in to the different light source powerlevels. I am still waiting answer from Hitachi to tell me how those should work. In my testing the Low-setting is the only usable one with good colors with proper luminance, mid and high settings boost the whites, but color luminance goes dimmer than the Low-setting!
QuoteQuoteQuoteQuoteQuoteIf the narrower band wavelength of the Leds is causing the chromatic aberation i guess the lens isnt the deciding factor, the Lk970 has 93-94% rec709, the M150 has 90% Bt2020 so that is a huge difference lol, and might be causing it.QuoteI have tried calibrating 3 different LK970 multiple times, but I have never got more than ~64% rec709 gamut coverage.QuoteIt's always super short in luminance, like 30-50% too low depending on the color and that leads to a coverage of only about 64%, despite the saturation and hue on the 2D CIE chart looking good.Have you actually measured the LK970 at >90% rec709?QuoteYes i measure it in calman, i can re-measure it now if ud like and share it hereQuoteThis is just now, With stock settings, normal laser.QuoteHmm. I mean yeah my 2D CIE looks just like that, but when it factors in the luminance, it falls way short.QuoteWhat are the luminance percentages of reference of the colors in your measurement?I was using DisplayCAL to measure my gamut.Do you get a similar gamut coverage after a 3DLUT?QuoteWhen i create 3dlut with displaycal or lightspace yes you are right it says 64-65% gamut. But how come is there that much difference between calman and lightsapce/displaycal.QuoteOn the other hand my Sim2 M150 measures even better in Lightspace/Displaycal compared to calman where it already measures great to begin with.I m getting 89.7% bt2020 stock settings , and 93-94% bt2020 in Lightspace/ Dispalycal so i guess every program measure it in its own way.QuoteThe reason for the low gamut % with 3DLUT in DisplayCAL is because the luminance values are super low and the 3DLUT is trying to correct them but it physically can't because the projector just can't do it.QuoteI have never used CalMAN so I am not sure really much about what it's doing or reporting.In your last picture, look at the Y vs. Target Y for your colors. The Y is way too low compared to the target. Remember, color gamut is a 3D volume, not just a 2D triangle. You can reach way out to the proper "saturation" level of a color which you see on the 2D CIE diagram (x and y), but still be way short on luminance (Y), and the gamut coverage therefore shouldn't be considered 100% or close to 100% if the luminance is way too low.[MEDIA=youtube]BQGbbobBsiw[/MEDIA]You can have a 2D CIE that looks like the corner points are perfect, but still be way short on luminance.[/size][/color]QuoteAh i see , i guess the luminance on the LK970 is really lacking.QuoteI took another measurement of the sim2 m150 but with proper settings because it was just default before and as u can see now the color luminance is almost spot on.This is with bt2020 and a 3dlut will make it exactly on point.The LK970 on the other hand cant get the luminance close, i just tried and i guess its a limitation.QuoteYeah your luminance values look fine.QuoteIt's just some projectors such as the LK970 that fall way short for some reason.I have seen the issue on a couple other DLPs as well. Never seen it on an LCD or LCoS though. But of course not all DLPs are like that at all, just some.And yeah you should definitely check out the theater fan/wind project, it's really awesome. Less gimmick than you might think (when done right). Could use some more users. I'm still improving and expanding things on the project. I hope it can really take off more once I make the fans easier to use and more generic to support all video sources.QuoteYeah, I will admit that I don't really know how the luminance is exactly supposed to play into the "gamut coverage" calculation and number. I am no expert.QuoteBut with the luminance values being so low, the actual vs. target colors (and deltaE 2000) are pretty far off and so I think perhaps the full gamut coverage is quite low, but again I am not sure exactly how to "properly" report that.QuoteI dont think 3chip dlp would have this problem and typically rgbrgb wheels do ok, its wheels other than rgbrgb or when the diodes arent to spec, like that panasonic pt-rz470(red and green leds with a laser/phosphor for green) has good red saturation but low luminance and less green saturation but luminance is too high.QuoteIt sounds awesome im absolutely going to look into it. It didnt come across as gimmickery at all... Tactile sensation and immersion is real. A gimmick is like... 1,000,000 dynamic contrast!QuoteI think it happens on RGBW wheels. Do we know if the LK970 is RGBW? It's because the white part messes with the proper color luminance I think.QuoteI meant like is the LK970, 94% rec709, or is it 64% rec709?QuoteIf you look at the chromaticity (x,y) and the 2D CIE diagram, it seems like it's 94% coverage.But if you look at the luminance values for the colors on the LK970, they are far short, upwards of 50% short in the worst cases and this leads to large deltaE 2000 errors and leads software like DisplayCAL to give you a number like 64% rec709 coverage when measured.I am not sure how this gamut coverage number is exactly supposed to be calculated and reported and how the luminance should be factoring in.Quotethere are tactile transducers/bass shakers which provide rumble and dbox which has lift roll yaw pitch motion. I have to temper my subs response so i isolate them off the floor with sorbothane then use bass shakers to replace the tactile rumble. Works really well for preventing low frewuencies from getting out of the room, persons with sub bass sensitivity, while still providing good immersive feedback. QuoteWould love to incorporate dbox and wind, but dbox wont work, so the idea that i can add wind is really cool...not sure, there are an array of color wheels and in some instances 2 color wheels. rgbcym, rgbcy(which is what led overlap does), rgbw, rgby, rgbrgbg orsometimes a dark green segment for p3 gamut.QuoteBlue core lasers are typically RGBY, so I think the 970 is similar.QuoteActually I think the LK970 is RBGY and the LK990 is RGBW so the LK990 might be even worse.QuoteI think it also has to do with BriliantColor on the LK990 because with that off color accuracy is good, but the brightness is very low, lower even than the LK970. With BriliantColor on the color is not able to be calibrated to any sort of low dE.QuoteIts one of the reasons why i get posterization when doing a 3dlut on the LK970 , its very obvious, the sim2 doesnt have any at all.QuoteI think its the main reason why the sim2 look bright, cause it has a lot of range for color luminance .QuoteYeah I do think it's why there is so much posterization. QuoteBecause it's trying to take a 64% rec709 to a 100% target.I did find that using a different rendering intent in ArgIICMS minimizes the posterization by a lot, and guess what that rendering intent does / is called...“Luminance preserving perceptual” (ArgyllCMS 1.8.3+) uses compression to make the source gamut fit within the destination gamut, but very heavily weights the preservation of the luminance value of the source, which will compromise the preservation of saturation. No contrast enhancement is used if the dynamic range is reduced. This intent may be of use where preserving the tonal distinctions in images is more important than maintaining overall colorfulness or contrast.I don't think it's a coincidence that the rendering intent that sacrifices saturation for luminance and that uses compression produces the best result on the LK970.QuoteThe thing is also when i do a p3 3dlut with my jvc rs440 whcih does 90%p3 i get posterization but to a lesser degree than the LK970, a lot less actually but its there and can be seen in some scenes.QuoteWith the Sim2 M150 i create a bt2020 3dlut and no posterization whatsoever even though the projectoe can do 90% bt2020 not 100% , i guess the RGB LED is a different animal.There is a workaround in LightSpace called Concatenation method, and DisplayCal also has good solution for it but at the expense of saturation like u said, i just dont know how come the Sim2 has absolutely zero posterization.QuoteI talked to someone who is much more knowledgeable on 3DLUTs about it once before and they said posterization is not really caused by targeting a larger than native gamut.QuoteFor example I tested targeting P3 on my rec709 PC monitor which only does about 78% P3 and I saw no posterization with the P3 3DLUT.It's definitely more complicated than that, but unfortunately is past my current understanding of color systems.QuoteThis is exactly what I measured on the lk990. It's colour volume is crap. Dave Harper told me the 970 didn't suffer it, quite adamantly at that but never substantiated it with measurements. It just shows he was wrong also.QuoteOn the lk990 you can turn off brilliant colour and measure correct values but the light output is 40% lower. Seems like you do indeed have brilliant colour hard coded on. Everyone (mainly Dave) was always telling me I was wrong. Thanks for showing I was not.Sent from my SM-G988B using TapatalkQuoteIf I had to describe it based on trying to calibrate 3 LK970 both manually and with 3DLUT, I would definitely describe the behavior resembles that of BriliantColor and what happens when you have a RGBY or RGBW color wheel.QuoteI also encountered the same behavior with this projector: https://www.projectorcentral.com/Optoma-ZH403-1080p-Laser-Projector-Road-Test.htm"]Optoma ZH403 1080p DLP Laser Projector Road Test[/URL]It had far-off luminance, but with BriliantColor off I could get low dE, but at less than 2000 lumens on a 4000 lumen spec machine.[/size][/color]QuoteOn the 970, I am not surprised that BC is hard-coded in. My thought was that BC was always there and adjusted to provide more brightness for the LK990. Once you have seen or measured enough units you can spot BC a mile away. I can see it in the rendering of whites. RGB Laser or high lumen LED is the path forward for DLP in my opinion.QuoteBrilliantColor, amazing name for a feature that f**ks up colors! :cool:QuoteI believe my Hitachi RGB LED DLP has something similar baked in to the different light source powerlevels. I am still waiting answer from Hitachi to tell me how those should work. In my testing the Low-setting is the only usable one with good colors with proper luminance, mid and high settings boost the whites, but color luminance goes dimmer than the Low-setting!
QuoteQuoteQuoteQuoteIf the narrower band wavelength of the Leds is causing the chromatic aberation i guess the lens isnt the deciding factor, the Lk970 has 93-94% rec709, the M150 has 90% Bt2020 so that is a huge difference lol, and might be causing it.QuoteI have tried calibrating 3 different LK970 multiple times, but I have never got more than ~64% rec709 gamut coverage.QuoteIt's always super short in luminance, like 30-50% too low depending on the color and that leads to a coverage of only about 64%, despite the saturation and hue on the 2D CIE chart looking good.Have you actually measured the LK970 at >90% rec709?QuoteYes i measure it in calman, i can re-measure it now if ud like and share it hereQuoteThis is just now, With stock settings, normal laser.QuoteHmm. I mean yeah my 2D CIE looks just like that, but when it factors in the luminance, it falls way short.QuoteWhat are the luminance percentages of reference of the colors in your measurement?I was using DisplayCAL to measure my gamut.Do you get a similar gamut coverage after a 3DLUT?QuoteWhen i create 3dlut with displaycal or lightspace yes you are right it says 64-65% gamut. But how come is there that much difference between calman and lightsapce/displaycal.QuoteOn the other hand my Sim2 M150 measures even better in Lightspace/Displaycal compared to calman where it already measures great to begin with.I m getting 89.7% bt2020 stock settings , and 93-94% bt2020 in Lightspace/ Dispalycal so i guess every program measure it in its own way.QuoteThe reason for the low gamut % with 3DLUT in DisplayCAL is because the luminance values are super low and the 3DLUT is trying to correct them but it physically can't because the projector just can't do it.QuoteI have never used CalMAN so I am not sure really much about what it's doing or reporting.In your last picture, look at the Y vs. Target Y for your colors. The Y is way too low compared to the target. Remember, color gamut is a 3D volume, not just a 2D triangle. You can reach way out to the proper "saturation" level of a color which you see on the 2D CIE diagram (x and y), but still be way short on luminance (Y), and the gamut coverage therefore shouldn't be considered 100% or close to 100% if the luminance is way too low.[MEDIA=youtube]BQGbbobBsiw[/MEDIA]You can have a 2D CIE that looks like the corner points are perfect, but still be way short on luminance.[/size][/color]QuoteAh i see , i guess the luminance on the LK970 is really lacking.QuoteI took another measurement of the sim2 m150 but with proper settings because it was just default before and as u can see now the color luminance is almost spot on.This is with bt2020 and a 3dlut will make it exactly on point.The LK970 on the other hand cant get the luminance close, i just tried and i guess its a limitation.QuoteYeah your luminance values look fine.QuoteIt's just some projectors such as the LK970 that fall way short for some reason.I have seen the issue on a couple other DLPs as well. Never seen it on an LCD or LCoS though. But of course not all DLPs are like that at all, just some.And yeah you should definitely check out the theater fan/wind project, it's really awesome. Less gimmick than you might think (when done right). Could use some more users. I'm still improving and expanding things on the project. I hope it can really take off more once I make the fans easier to use and more generic to support all video sources.QuoteYeah, I will admit that I don't really know how the luminance is exactly supposed to play into the "gamut coverage" calculation and number. I am no expert.QuoteBut with the luminance values being so low, the actual vs. target colors (and deltaE 2000) are pretty far off and so I think perhaps the full gamut coverage is quite low, but again I am not sure exactly how to "properly" report that.QuoteI dont think 3chip dlp would have this problem and typically rgbrgb wheels do ok, its wheels other than rgbrgb or when the diodes arent to spec, like that panasonic pt-rz470(red and green leds with a laser/phosphor for green) has good red saturation but low luminance and less green saturation but luminance is too high.QuoteIt sounds awesome im absolutely going to look into it. It didnt come across as gimmickery at all... Tactile sensation and immersion is real. A gimmick is like... 1,000,000 dynamic contrast!QuoteI think it happens on RGBW wheels. Do we know if the LK970 is RGBW? It's because the white part messes with the proper color luminance I think.QuoteI meant like is the LK970, 94% rec709, or is it 64% rec709?QuoteIf you look at the chromaticity (x,y) and the 2D CIE diagram, it seems like it's 94% coverage.But if you look at the luminance values for the colors on the LK970, they are far short, upwards of 50% short in the worst cases and this leads to large deltaE 2000 errors and leads software like DisplayCAL to give you a number like 64% rec709 coverage when measured.I am not sure how this gamut coverage number is exactly supposed to be calculated and reported and how the luminance should be factoring in.Quotethere are tactile transducers/bass shakers which provide rumble and dbox which has lift roll yaw pitch motion. I have to temper my subs response so i isolate them off the floor with sorbothane then use bass shakers to replace the tactile rumble. Works really well for preventing low frewuencies from getting out of the room, persons with sub bass sensitivity, while still providing good immersive feedback. QuoteWould love to incorporate dbox and wind, but dbox wont work, so the idea that i can add wind is really cool...not sure, there are an array of color wheels and in some instances 2 color wheels. rgbcym, rgbcy(which is what led overlap does), rgbw, rgby, rgbrgbg orsometimes a dark green segment for p3 gamut.QuoteBlue core lasers are typically RGBY, so I think the 970 is similar.QuoteActually I think the LK970 is RBGY and the LK990 is RGBW so the LK990 might be even worse.QuoteI think it also has to do with BriliantColor on the LK990 because with that off color accuracy is good, but the brightness is very low, lower even than the LK970. With BriliantColor on the color is not able to be calibrated to any sort of low dE.QuoteIts one of the reasons why i get posterization when doing a 3dlut on the LK970 , its very obvious, the sim2 doesnt have any at all.QuoteI think its the main reason why the sim2 look bright, cause it has a lot of range for color luminance .QuoteYeah I do think it's why there is so much posterization. QuoteBecause it's trying to take a 64% rec709 to a 100% target.I did find that using a different rendering intent in ArgIICMS minimizes the posterization by a lot, and guess what that rendering intent does / is called...“Luminance preserving perceptual” (ArgyllCMS 1.8.3+) uses compression to make the source gamut fit within the destination gamut, but very heavily weights the preservation of the luminance value of the source, which will compromise the preservation of saturation. No contrast enhancement is used if the dynamic range is reduced. This intent may be of use where preserving the tonal distinctions in images is more important than maintaining overall colorfulness or contrast.I don't think it's a coincidence that the rendering intent that sacrifices saturation for luminance and that uses compression produces the best result on the LK970.QuoteThe thing is also when i do a p3 3dlut with my jvc rs440 whcih does 90%p3 i get posterization but to a lesser degree than the LK970, a lot less actually but its there and can be seen in some scenes.QuoteWith the Sim2 M150 i create a bt2020 3dlut and no posterization whatsoever even though the projectoe can do 90% bt2020 not 100% , i guess the RGB LED is a different animal.There is a workaround in LightSpace called Concatenation method, and DisplayCal also has good solution for it but at the expense of saturation like u said, i just dont know how come the Sim2 has absolutely zero posterization.QuoteI talked to someone who is much more knowledgeable on 3DLUTs about it once before and they said posterization is not really caused by targeting a larger than native gamut.QuoteFor example I tested targeting P3 on my rec709 PC monitor which only does about 78% P3 and I saw no posterization with the P3 3DLUT.It's definitely more complicated than that, but unfortunately is past my current understanding of color systems.QuoteThis is exactly what I measured on the lk990. It's colour volume is crap. Dave Harper told me the 970 didn't suffer it, quite adamantly at that but never substantiated it with measurements. It just shows he was wrong also.QuoteOn the lk990 you can turn off brilliant colour and measure correct values but the light output is 40% lower. Seems like you do indeed have brilliant colour hard coded on. Everyone (mainly Dave) was always telling me I was wrong. Thanks for showing I was not.Sent from my SM-G988B using TapatalkQuoteIf I had to describe it based on trying to calibrate 3 LK970 both manually and with 3DLUT, I would definitely describe the behavior resembles that of BriliantColor and what happens when you have a RGBY or RGBW color wheel.QuoteI also encountered the same behavior with this projector: https://www.projectorcentral.com/Optoma-ZH403-1080p-Laser-Projector-Road-Test.htm"]Optoma ZH403 1080p DLP Laser Projector Road Test[/URL]It had far-off luminance, but with BriliantColor off I could get low dE, but at less than 2000 lumens on a 4000 lumen spec machine.[/size][/color]QuoteOn the 970, I am not surprised that BC is hard-coded in. My thought was that BC was always there and adjusted to provide more brightness for the LK990. Once you have seen or measured enough units you can spot BC a mile away. I can see it in the rendering of whites. RGB Laser or high lumen LED is the path forward for DLP in my opinion.QuoteBrilliantColor, amazing name for a feature that f**ks up colors! :cool:QuoteI believe my Hitachi RGB LED DLP has something similar baked in to the different light source powerlevels. I am still waiting answer from Hitachi to tell me how those should work. In my testing the Low-setting is the only usable one with good colors with proper luminance, mid and high settings boost the whites, but color luminance goes dimmer than the Low-setting!
QuoteQuoteQuoteIf the narrower band wavelength of the Leds is causing the chromatic aberation i guess the lens isnt the deciding factor, the Lk970 has 93-94% rec709, the M150 has 90% Bt2020 so that is a huge difference lol, and might be causing it.QuoteI have tried calibrating 3 different LK970 multiple times, but I have never got more than ~64% rec709 gamut coverage.QuoteIt's always super short in luminance, like 30-50% too low depending on the color and that leads to a coverage of only about 64%, despite the saturation and hue on the 2D CIE chart looking good.Have you actually measured the LK970 at >90% rec709?QuoteYes i measure it in calman, i can re-measure it now if ud like and share it hereQuoteThis is just now, With stock settings, normal laser.QuoteHmm. I mean yeah my 2D CIE looks just like that, but when it factors in the luminance, it falls way short.QuoteWhat are the luminance percentages of reference of the colors in your measurement?I was using DisplayCAL to measure my gamut.Do you get a similar gamut coverage after a 3DLUT?QuoteWhen i create 3dlut with displaycal or lightspace yes you are right it says 64-65% gamut. But how come is there that much difference between calman and lightsapce/displaycal.QuoteOn the other hand my Sim2 M150 measures even better in Lightspace/Displaycal compared to calman where it already measures great to begin with.I m getting 89.7% bt2020 stock settings , and 93-94% bt2020 in Lightspace/ Dispalycal so i guess every program measure it in its own way.QuoteThe reason for the low gamut % with 3DLUT in DisplayCAL is because the luminance values are super low and the 3DLUT is trying to correct them but it physically can't because the projector just can't do it.QuoteI have never used CalMAN so I am not sure really much about what it's doing or reporting.In your last picture, look at the Y vs. Target Y for your colors. The Y is way too low compared to the target. Remember, color gamut is a 3D volume, not just a 2D triangle. You can reach way out to the proper "saturation" level of a color which you see on the 2D CIE diagram (x and y), but still be way short on luminance (Y), and the gamut coverage therefore shouldn't be considered 100% or close to 100% if the luminance is way too low.[MEDIA=youtube]BQGbbobBsiw[/MEDIA]You can have a 2D CIE that looks like the corner points are perfect, but still be way short on luminance.[/size][/color]QuoteAh i see , i guess the luminance on the LK970 is really lacking.QuoteI took another measurement of the sim2 m150 but with proper settings because it was just default before and as u can see now the color luminance is almost spot on.This is with bt2020 and a 3dlut will make it exactly on point.The LK970 on the other hand cant get the luminance close, i just tried and i guess its a limitation.QuoteYeah your luminance values look fine.QuoteIt's just some projectors such as the LK970 that fall way short for some reason.I have seen the issue on a couple other DLPs as well. Never seen it on an LCD or LCoS though. But of course not all DLPs are like that at all, just some.And yeah you should definitely check out the theater fan/wind project, it's really awesome. Less gimmick than you might think (when done right). Could use some more users. I'm still improving and expanding things on the project. I hope it can really take off more once I make the fans easier to use and more generic to support all video sources.QuoteYeah, I will admit that I don't really know how the luminance is exactly supposed to play into the "gamut coverage" calculation and number. I am no expert.QuoteBut with the luminance values being so low, the actual vs. target colors (and deltaE 2000) are pretty far off and so I think perhaps the full gamut coverage is quite low, but again I am not sure exactly how to "properly" report that.QuoteI dont think 3chip dlp would have this problem and typically rgbrgb wheels do ok, its wheels other than rgbrgb or when the diodes arent to spec, like that panasonic pt-rz470(red and green leds with a laser/phosphor for green) has good red saturation but low luminance and less green saturation but luminance is too high.QuoteIt sounds awesome im absolutely going to look into it. It didnt come across as gimmickery at all... Tactile sensation and immersion is real. A gimmick is like... 1,000,000 dynamic contrast!QuoteI think it happens on RGBW wheels. Do we know if the LK970 is RGBW? It's because the white part messes with the proper color luminance I think.QuoteI meant like is the LK970, 94% rec709, or is it 64% rec709?QuoteIf you look at the chromaticity (x,y) and the 2D CIE diagram, it seems like it's 94% coverage.But if you look at the luminance values for the colors on the LK970, they are far short, upwards of 50% short in the worst cases and this leads to large deltaE 2000 errors and leads software like DisplayCAL to give you a number like 64% rec709 coverage when measured.I am not sure how this gamut coverage number is exactly supposed to be calculated and reported and how the luminance should be factoring in.Quotethere are tactile transducers/bass shakers which provide rumble and dbox which has lift roll yaw pitch motion. I have to temper my subs response so i isolate them off the floor with sorbothane then use bass shakers to replace the tactile rumble. Works really well for preventing low frewuencies from getting out of the room, persons with sub bass sensitivity, while still providing good immersive feedback. QuoteWould love to incorporate dbox and wind, but dbox wont work, so the idea that i can add wind is really cool...not sure, there are an array of color wheels and in some instances 2 color wheels. rgbcym, rgbcy(which is what led overlap does), rgbw, rgby, rgbrgbg orsometimes a dark green segment for p3 gamut.QuoteBlue core lasers are typically RGBY, so I think the 970 is similar.QuoteActually I think the LK970 is RBGY and the LK990 is RGBW so the LK990 might be even worse.QuoteI think it also has to do with BriliantColor on the LK990 because with that off color accuracy is good, but the brightness is very low, lower even than the LK970. With BriliantColor on the color is not able to be calibrated to any sort of low dE.QuoteIts one of the reasons why i get posterization when doing a 3dlut on the LK970 , its very obvious, the sim2 doesnt have any at all.QuoteI think its the main reason why the sim2 look bright, cause it has a lot of range for color luminance .QuoteYeah I do think it's why there is so much posterization. QuoteBecause it's trying to take a 64% rec709 to a 100% target.I did find that using a different rendering intent in ArgIICMS minimizes the posterization by a lot, and guess what that rendering intent does / is called...“Luminance preserving perceptual” (ArgyllCMS 1.8.3+) uses compression to make the source gamut fit within the destination gamut, but very heavily weights the preservation of the luminance value of the source, which will compromise the preservation of saturation. No contrast enhancement is used if the dynamic range is reduced. This intent may be of use where preserving the tonal distinctions in images is more important than maintaining overall colorfulness or contrast.I don't think it's a coincidence that the rendering intent that sacrifices saturation for luminance and that uses compression produces the best result on the LK970.QuoteThe thing is also when i do a p3 3dlut with my jvc rs440 whcih does 90%p3 i get posterization but to a lesser degree than the LK970, a lot less actually but its there and can be seen in some scenes.QuoteWith the Sim2 M150 i create a bt2020 3dlut and no posterization whatsoever even though the projectoe can do 90% bt2020 not 100% , i guess the RGB LED is a different animal.There is a workaround in LightSpace called Concatenation method, and DisplayCal also has good solution for it but at the expense of saturation like u said, i just dont know how come the Sim2 has absolutely zero posterization.QuoteI talked to someone who is much more knowledgeable on 3DLUTs about it once before and they said posterization is not really caused by targeting a larger than native gamut.QuoteFor example I tested targeting P3 on my rec709 PC monitor which only does about 78% P3 and I saw no posterization with the P3 3DLUT.It's definitely more complicated than that, but unfortunately is past my current understanding of color systems.QuoteThis is exactly what I measured on the lk990. It's colour volume is crap. Dave Harper told me the 970 didn't suffer it, quite adamantly at that but never substantiated it with measurements. It just shows he was wrong also.QuoteOn the lk990 you can turn off brilliant colour and measure correct values but the light output is 40% lower. Seems like you do indeed have brilliant colour hard coded on. Everyone (mainly Dave) was always telling me I was wrong. Thanks for showing I was not.Sent from my SM-G988B using TapatalkQuoteIf I had to describe it based on trying to calibrate 3 LK970 both manually and with 3DLUT, I would definitely describe the behavior resembles that of BriliantColor and what happens when you have a RGBY or RGBW color wheel.QuoteI also encountered the same behavior with this projector: https://www.projectorcentral.com/Optoma-ZH403-1080p-Laser-Projector-Road-Test.htm"]Optoma ZH403 1080p DLP Laser Projector Road Test[/URL]It had far-off luminance, but with BriliantColor off I could get low dE, but at less than 2000 lumens on a 4000 lumen spec machine.[/size][/color]QuoteOn the 970, I am not surprised that BC is hard-coded in. My thought was that BC was always there and adjusted to provide more brightness for the LK990. Once you have seen or measured enough units you can spot BC a mile away. I can see it in the rendering of whites. RGB Laser or high lumen LED is the path forward for DLP in my opinion.QuoteBrilliantColor, amazing name for a feature that f**ks up colors! :cool:QuoteI believe my Hitachi RGB LED DLP has something similar baked in to the different light source powerlevels. I am still waiting answer from Hitachi to tell me how those should work. In my testing the Low-setting is the only usable one with good colors with proper luminance, mid and high settings boost the whites, but color luminance goes dimmer than the Low-setting!
QuoteQuoteIf the narrower band wavelength of the Leds is causing the chromatic aberation i guess the lens isnt the deciding factor, the Lk970 has 93-94% rec709, the M150 has 90% Bt2020 so that is a huge difference lol, and might be causing it.QuoteI have tried calibrating 3 different LK970 multiple times, but I have never got more than ~64% rec709 gamut coverage.QuoteIt's always super short in luminance, like 30-50% too low depending on the color and that leads to a coverage of only about 64%, despite the saturation and hue on the 2D CIE chart looking good.Have you actually measured the LK970 at >90% rec709?QuoteYes i measure it in calman, i can re-measure it now if ud like and share it hereQuoteThis is just now, With stock settings, normal laser.QuoteHmm. I mean yeah my 2D CIE looks just like that, but when it factors in the luminance, it falls way short.QuoteWhat are the luminance percentages of reference of the colors in your measurement?I was using DisplayCAL to measure my gamut.Do you get a similar gamut coverage after a 3DLUT?QuoteWhen i create 3dlut with displaycal or lightspace yes you are right it says 64-65% gamut. But how come is there that much difference between calman and lightsapce/displaycal.QuoteOn the other hand my Sim2 M150 measures even better in Lightspace/Displaycal compared to calman where it already measures great to begin with.I m getting 89.7% bt2020 stock settings , and 93-94% bt2020 in Lightspace/ Dispalycal so i guess every program measure it in its own way.QuoteThe reason for the low gamut % with 3DLUT in DisplayCAL is because the luminance values are super low and the 3DLUT is trying to correct them but it physically can't because the projector just can't do it.QuoteI have never used CalMAN so I am not sure really much about what it's doing or reporting.In your last picture, look at the Y vs. Target Y for your colors. The Y is way too low compared to the target. Remember, color gamut is a 3D volume, not just a 2D triangle. You can reach way out to the proper "saturation" level of a color which you see on the 2D CIE diagram (x and y), but still be way short on luminance (Y), and the gamut coverage therefore shouldn't be considered 100% or close to 100% if the luminance is way too low.[MEDIA=youtube]BQGbbobBsiw[/MEDIA]You can have a 2D CIE that looks like the corner points are perfect, but still be way short on luminance.[/size][/color]QuoteAh i see , i guess the luminance on the LK970 is really lacking.QuoteI took another measurement of the sim2 m150 but with proper settings because it was just default before and as u can see now the color luminance is almost spot on.This is with bt2020 and a 3dlut will make it exactly on point.The LK970 on the other hand cant get the luminance close, i just tried and i guess its a limitation.QuoteYeah your luminance values look fine.QuoteIt's just some projectors such as the LK970 that fall way short for some reason.I have seen the issue on a couple other DLPs as well. Never seen it on an LCD or LCoS though. But of course not all DLPs are like that at all, just some.And yeah you should definitely check out the theater fan/wind project, it's really awesome. Less gimmick than you might think (when done right). Could use some more users. I'm still improving and expanding things on the project. I hope it can really take off more once I make the fans easier to use and more generic to support all video sources.QuoteYeah, I will admit that I don't really know how the luminance is exactly supposed to play into the "gamut coverage" calculation and number. I am no expert.QuoteBut with the luminance values being so low, the actual vs. target colors (and deltaE 2000) are pretty far off and so I think perhaps the full gamut coverage is quite low, but again I am not sure exactly how to "properly" report that.QuoteI dont think 3chip dlp would have this problem and typically rgbrgb wheels do ok, its wheels other than rgbrgb or when the diodes arent to spec, like that panasonic pt-rz470(red and green leds with a laser/phosphor for green) has good red saturation but low luminance and less green saturation but luminance is too high.QuoteIt sounds awesome im absolutely going to look into it. It didnt come across as gimmickery at all... Tactile sensation and immersion is real. A gimmick is like... 1,000,000 dynamic contrast!QuoteI think it happens on RGBW wheels. Do we know if the LK970 is RGBW? It's because the white part messes with the proper color luminance I think.QuoteI meant like is the LK970, 94% rec709, or is it 64% rec709?QuoteIf you look at the chromaticity (x,y) and the 2D CIE diagram, it seems like it's 94% coverage.But if you look at the luminance values for the colors on the LK970, they are far short, upwards of 50% short in the worst cases and this leads to large deltaE 2000 errors and leads software like DisplayCAL to give you a number like 64% rec709 coverage when measured.I am not sure how this gamut coverage number is exactly supposed to be calculated and reported and how the luminance should be factoring in.Quotethere are tactile transducers/bass shakers which provide rumble and dbox which has lift roll yaw pitch motion. I have to temper my subs response so i isolate them off the floor with sorbothane then use bass shakers to replace the tactile rumble. Works really well for preventing low frewuencies from getting out of the room, persons with sub bass sensitivity, while still providing good immersive feedback. QuoteWould love to incorporate dbox and wind, but dbox wont work, so the idea that i can add wind is really cool...not sure, there are an array of color wheels and in some instances 2 color wheels. rgbcym, rgbcy(which is what led overlap does), rgbw, rgby, rgbrgbg orsometimes a dark green segment for p3 gamut.QuoteBlue core lasers are typically RGBY, so I think the 970 is similar.QuoteActually I think the LK970 is RBGY and the LK990 is RGBW so the LK990 might be even worse.QuoteI think it also has to do with BriliantColor on the LK990 because with that off color accuracy is good, but the brightness is very low, lower even than the LK970. With BriliantColor on the color is not able to be calibrated to any sort of low dE.QuoteIts one of the reasons why i get posterization when doing a 3dlut on the LK970 , its very obvious, the sim2 doesnt have any at all.QuoteI think its the main reason why the sim2 look bright, cause it has a lot of range for color luminance .QuoteYeah I do think it's why there is so much posterization. QuoteBecause it's trying to take a 64% rec709 to a 100% target.I did find that using a different rendering intent in ArgIICMS minimizes the posterization by a lot, and guess what that rendering intent does / is called...“Luminance preserving perceptual” (ArgyllCMS 1.8.3+) uses compression to make the source gamut fit within the destination gamut, but very heavily weights the preservation of the luminance value of the source, which will compromise the preservation of saturation. No contrast enhancement is used if the dynamic range is reduced. This intent may be of use where preserving the tonal distinctions in images is more important than maintaining overall colorfulness or contrast.I don't think it's a coincidence that the rendering intent that sacrifices saturation for luminance and that uses compression produces the best result on the LK970.QuoteThe thing is also when i do a p3 3dlut with my jvc rs440 whcih does 90%p3 i get posterization but to a lesser degree than the LK970, a lot less actually but its there and can be seen in some scenes.QuoteWith the Sim2 M150 i create a bt2020 3dlut and no posterization whatsoever even though the projectoe can do 90% bt2020 not 100% , i guess the RGB LED is a different animal.There is a workaround in LightSpace called Concatenation method, and DisplayCal also has good solution for it but at the expense of saturation like u said, i just dont know how come the Sim2 has absolutely zero posterization.QuoteI talked to someone who is much more knowledgeable on 3DLUTs about it once before and they said posterization is not really caused by targeting a larger than native gamut.QuoteFor example I tested targeting P3 on my rec709 PC monitor which only does about 78% P3 and I saw no posterization with the P3 3DLUT.It's definitely more complicated than that, but unfortunately is past my current understanding of color systems.QuoteThis is exactly what I measured on the lk990. It's colour volume is crap. Dave Harper told me the 970 didn't suffer it, quite adamantly at that but never substantiated it with measurements. It just shows he was wrong also.QuoteOn the lk990 you can turn off brilliant colour and measure correct values but the light output is 40% lower. Seems like you do indeed have brilliant colour hard coded on. Everyone (mainly Dave) was always telling me I was wrong. Thanks for showing I was not.Sent from my SM-G988B using TapatalkQuoteIf I had to describe it based on trying to calibrate 3 LK970 both manually and with 3DLUT, I would definitely describe the behavior resembles that of BriliantColor and what happens when you have a RGBY or RGBW color wheel.QuoteI also encountered the same behavior with this projector: https://www.projectorcentral.com/Optoma-ZH403-1080p-Laser-Projector-Road-Test.htm"]Optoma ZH403 1080p DLP Laser Projector Road Test[/URL]It had far-off luminance, but with BriliantColor off I could get low dE, but at less than 2000 lumens on a 4000 lumen spec machine.[/size][/color]QuoteOn the 970, I am not surprised that BC is hard-coded in. My thought was that BC was always there and adjusted to provide more brightness for the LK990. Once you have seen or measured enough units you can spot BC a mile away. I can see it in the rendering of whites. RGB Laser or high lumen LED is the path forward for DLP in my opinion.QuoteBrilliantColor, amazing name for a feature that f**ks up colors! :cool:QuoteI believe my Hitachi RGB LED DLP has something similar baked in to the different light source powerlevels. I am still waiting answer from Hitachi to tell me how those should work. In my testing the Low-setting is the only usable one with good colors with proper luminance, mid and high settings boost the whites, but color luminance goes dimmer than the Low-setting!
QuoteIf the narrower band wavelength of the Leds is causing the chromatic aberation i guess the lens isnt the deciding factor, the Lk970 has 93-94% rec709, the M150 has 90% Bt2020 so that is a huge difference lol, and might be causing it.QuoteI have tried calibrating 3 different LK970 multiple times, but I have never got more than ~64% rec709 gamut coverage.QuoteIt's always super short in luminance, like 30-50% too low depending on the color and that leads to a coverage of only about 64%, despite the saturation and hue on the 2D CIE chart looking good.Have you actually measured the LK970 at >90% rec709?QuoteYes i measure it in calman, i can re-measure it now if ud like and share it hereQuoteThis is just now, With stock settings, normal laser.QuoteHmm. I mean yeah my 2D CIE looks just like that, but when it factors in the luminance, it falls way short.QuoteWhat are the luminance percentages of reference of the colors in your measurement?I was using DisplayCAL to measure my gamut.Do you get a similar gamut coverage after a 3DLUT?QuoteWhen i create 3dlut with displaycal or lightspace yes you are right it says 64-65% gamut. But how come is there that much difference between calman and lightsapce/displaycal.QuoteOn the other hand my Sim2 M150 measures even better in Lightspace/Displaycal compared to calman where it already measures great to begin with.I m getting 89.7% bt2020 stock settings , and 93-94% bt2020 in Lightspace/ Dispalycal so i guess every program measure it in its own way.QuoteThe reason for the low gamut % with 3DLUT in DisplayCAL is because the luminance values are super low and the 3DLUT is trying to correct them but it physically can't because the projector just can't do it.QuoteI have never used CalMAN so I am not sure really much about what it's doing or reporting.In your last picture, look at the Y vs. Target Y for your colors. The Y is way too low compared to the target. Remember, color gamut is a 3D volume, not just a 2D triangle. You can reach way out to the proper "saturation" level of a color which you see on the 2D CIE diagram (x and y), but still be way short on luminance (Y), and the gamut coverage therefore shouldn't be considered 100% or close to 100% if the luminance is way too low.[MEDIA=youtube]BQGbbobBsiw[/MEDIA]You can have a 2D CIE that looks like the corner points are perfect, but still be way short on luminance.[/size][/color]QuoteAh i see , i guess the luminance on the LK970 is really lacking.QuoteI took another measurement of the sim2 m150 but with proper settings because it was just default before and as u can see now the color luminance is almost spot on.This is with bt2020 and a 3dlut will make it exactly on point.The LK970 on the other hand cant get the luminance close, i just tried and i guess its a limitation.QuoteYeah your luminance values look fine.QuoteIt's just some projectors such as the LK970 that fall way short for some reason.I have seen the issue on a couple other DLPs as well. Never seen it on an LCD or LCoS though. But of course not all DLPs are like that at all, just some.And yeah you should definitely check out the theater fan/wind project, it's really awesome. Less gimmick than you might think (when done right). Could use some more users. I'm still improving and expanding things on the project. I hope it can really take off more once I make the fans easier to use and more generic to support all video sources.QuoteYeah, I will admit that I don't really know how the luminance is exactly supposed to play into the "gamut coverage" calculation and number. I am no expert.QuoteBut with the luminance values being so low, the actual vs. target colors (and deltaE 2000) are pretty far off and so I think perhaps the full gamut coverage is quite low, but again I am not sure exactly how to "properly" report that.QuoteI dont think 3chip dlp would have this problem and typically rgbrgb wheels do ok, its wheels other than rgbrgb or when the diodes arent to spec, like that panasonic pt-rz470(red and green leds with a laser/phosphor for green) has good red saturation but low luminance and less green saturation but luminance is too high.QuoteIt sounds awesome im absolutely going to look into it. It didnt come across as gimmickery at all... Tactile sensation and immersion is real. A gimmick is like... 1,000,000 dynamic contrast!QuoteI think it happens on RGBW wheels. Do we know if the LK970 is RGBW? It's because the white part messes with the proper color luminance I think.QuoteI meant like is the LK970, 94% rec709, or is it 64% rec709?QuoteIf you look at the chromaticity (x,y) and the 2D CIE diagram, it seems like it's 94% coverage.But if you look at the luminance values for the colors on the LK970, they are far short, upwards of 50% short in the worst cases and this leads to large deltaE 2000 errors and leads software like DisplayCAL to give you a number like 64% rec709 coverage when measured.I am not sure how this gamut coverage number is exactly supposed to be calculated and reported and how the luminance should be factoring in.Quotethere are tactile transducers/bass shakers which provide rumble and dbox which has lift roll yaw pitch motion. I have to temper my subs response so i isolate them off the floor with sorbothane then use bass shakers to replace the tactile rumble. Works really well for preventing low frewuencies from getting out of the room, persons with sub bass sensitivity, while still providing good immersive feedback. QuoteWould love to incorporate dbox and wind, but dbox wont work, so the idea that i can add wind is really cool...not sure, there are an array of color wheels and in some instances 2 color wheels. rgbcym, rgbcy(which is what led overlap does), rgbw, rgby, rgbrgbg orsometimes a dark green segment for p3 gamut.QuoteBlue core lasers are typically RGBY, so I think the 970 is similar.QuoteActually I think the LK970 is RBGY and the LK990 is RGBW so the LK990 might be even worse.QuoteI think it also has to do with BriliantColor on the LK990 because with that off color accuracy is good, but the brightness is very low, lower even than the LK970. With BriliantColor on the color is not able to be calibrated to any sort of low dE.QuoteIts one of the reasons why i get posterization when doing a 3dlut on the LK970 , its very obvious, the sim2 doesnt have any at all.QuoteI think its the main reason why the sim2 look bright, cause it has a lot of range for color luminance .QuoteYeah I do think it's why there is so much posterization. QuoteBecause it's trying to take a 64% rec709 to a 100% target.I did find that using a different rendering intent in ArgIICMS minimizes the posterization by a lot, and guess what that rendering intent does / is called...“Luminance preserving perceptual” (ArgyllCMS 1.8.3+) uses compression to make the source gamut fit within the destination gamut, but very heavily weights the preservation of the luminance value of the source, which will compromise the preservation of saturation. No contrast enhancement is used if the dynamic range is reduced. This intent may be of use where preserving the tonal distinctions in images is more important than maintaining overall colorfulness or contrast.I don't think it's a coincidence that the rendering intent that sacrifices saturation for luminance and that uses compression produces the best result on the LK970.QuoteThe thing is also when i do a p3 3dlut with my jvc rs440 whcih does 90%p3 i get posterization but to a lesser degree than the LK970, a lot less actually but its there and can be seen in some scenes.QuoteWith the Sim2 M150 i create a bt2020 3dlut and no posterization whatsoever even though the projectoe can do 90% bt2020 not 100% , i guess the RGB LED is a different animal.There is a workaround in LightSpace called Concatenation method, and DisplayCal also has good solution for it but at the expense of saturation like u said, i just dont know how come the Sim2 has absolutely zero posterization.QuoteI talked to someone who is much more knowledgeable on 3DLUTs about it once before and they said posterization is not really caused by targeting a larger than native gamut.QuoteFor example I tested targeting P3 on my rec709 PC monitor which only does about 78% P3 and I saw no posterization with the P3 3DLUT.It's definitely more complicated than that, but unfortunately is past my current understanding of color systems.QuoteThis is exactly what I measured on the lk990. It's colour volume is crap. Dave Harper told me the 970 didn't suffer it, quite adamantly at that but never substantiated it with measurements. It just shows he was wrong also.QuoteOn the lk990 you can turn off brilliant colour and measure correct values but the light output is 40% lower. Seems like you do indeed have brilliant colour hard coded on. Everyone (mainly Dave) was always telling me I was wrong. Thanks for showing I was not.Sent from my SM-G988B using TapatalkQuoteIf I had to describe it based on trying to calibrate 3 LK970 both manually and with 3DLUT, I would definitely describe the behavior resembles that of BriliantColor and what happens when you have a RGBY or RGBW color wheel.QuoteI also encountered the same behavior with this projector: https://www.projectorcentral.com/Optoma-ZH403-1080p-Laser-Projector-Road-Test.htm"]Optoma ZH403 1080p DLP Laser Projector Road Test[/URL]It had far-off luminance, but with BriliantColor off I could get low dE, but at less than 2000 lumens on a 4000 lumen spec machine.[/size][/color]QuoteOn the 970, I am not surprised that BC is hard-coded in. My thought was that BC was always there and adjusted to provide more brightness for the LK990. Once you have seen or measured enough units you can spot BC a mile away. I can see it in the rendering of whites. RGB Laser or high lumen LED is the path forward for DLP in my opinion.QuoteBrilliantColor, amazing name for a feature that f**ks up colors! :cool:QuoteI believe my Hitachi RGB LED DLP has something similar baked in to the different light source powerlevels. I am still waiting answer from Hitachi to tell me how those should work. In my testing the Low-setting is the only usable one with good colors with proper luminance, mid and high settings boost the whites, but color luminance goes dimmer than the Low-setting!
If the narrower band wavelength of the Leds is causing the chromatic aberation i guess the lens isnt the deciding factor, the Lk970 has 93-94% rec709, the M150 has 90% Bt2020 so that is a huge difference lol, and might be causing it.
I have tried calibrating 3 different LK970 multiple times, but I have never got more than ~64% rec709 gamut coverage.
It's always super short in luminance, like 30-50% too low depending on the color and that leads to a coverage of only about 64%, despite the saturation and hue on the 2D CIE chart looking good.Have you actually measured the LK970 at >90% rec709?
Yes i measure it in calman, i can re-measure it now if ud like and share it here
This is just now, With stock settings, normal laser.
Hmm. I mean yeah my 2D CIE looks just like that, but when it factors in the luminance, it falls way short.
What are the luminance percentages of reference of the colors in your measurement?I was using DisplayCAL to measure my gamut.Do you get a similar gamut coverage after a 3DLUT?
When i create 3dlut with displaycal or lightspace yes you are right it says 64-65% gamut. But how come is there that much difference between calman and lightsapce/displaycal.
On the other hand my Sim2 M150 measures even better in Lightspace/Displaycal compared to calman where it already measures great to begin with.I m getting 89.7% bt2020 stock settings , and 93-94% bt2020 in Lightspace/ Dispalycal so i guess every program measure it in its own way.
The reason for the low gamut % with 3DLUT in DisplayCAL is because the luminance values are super low and the 3DLUT is trying to correct them but it physically can't because the projector just can't do it.
I have never used CalMAN so I am not sure really much about what it's doing or reporting.In your last picture, look at the Y vs. Target Y for your colors. The Y is way too low compared to the target. Remember, color gamut is a 3D volume, not just a 2D triangle. You can reach way out to the proper "saturation" level of a color which you see on the 2D CIE diagram (x and y), but still be way short on luminance (Y), and the gamut coverage therefore shouldn't be considered 100% or close to 100% if the luminance is way too low.[MEDIA=youtube]BQGbbobBsiw[/MEDIA]You can have a 2D CIE that looks like the corner points are perfect, but still be way short on luminance.[/size][/color]
Ah i see , i guess the luminance on the LK970 is really lacking.
I took another measurement of the sim2 m150 but with proper settings because it was just default before and as u can see now the color luminance is almost spot on.This is with bt2020 and a 3dlut will make it exactly on point.The LK970 on the other hand cant get the luminance close, i just tried and i guess its a limitation.
Yeah your luminance values look fine.
It's just some projectors such as the LK970 that fall way short for some reason.I have seen the issue on a couple other DLPs as well. Never seen it on an LCD or LCoS though. But of course not all DLPs are like that at all, just some.And yeah you should definitely check out the theater fan/wind project, it's really awesome. Less gimmick than you might think (when done right). Could use some more users. I'm still improving and expanding things on the project. I hope it can really take off more once I make the fans easier to use and more generic to support all video sources.
Yeah, I will admit that I don't really know how the luminance is exactly supposed to play into the "gamut coverage" calculation and number. I am no expert.
But with the luminance values being so low, the actual vs. target colors (and deltaE 2000) are pretty far off and so I think perhaps the full gamut coverage is quite low, but again I am not sure exactly how to "properly" report that.
I dont think 3chip dlp would have this problem and typically rgbrgb wheels do ok, its wheels other than rgbrgb or when the diodes arent to spec, like that panasonic pt-rz470(red and green leds with a laser/phosphor for green) has good red saturation but low luminance and less green saturation but luminance is too high.
It sounds awesome im absolutely going to look into it. It didnt come across as gimmickery at all... Tactile sensation and immersion is real. A gimmick is like... 1,000,000 dynamic contrast!
I think it happens on RGBW wheels. Do we know if the LK970 is RGBW? It's because the white part messes with the proper color luminance I think.
I meant like is the LK970, 94% rec709, or is it 64% rec709?
If you look at the chromaticity (x,y) and the 2D CIE diagram, it seems like it's 94% coverage.But if you look at the luminance values for the colors on the LK970, they are far short, upwards of 50% short in the worst cases and this leads to large deltaE 2000 errors and leads software like DisplayCAL to give you a number like 64% rec709 coverage when measured.I am not sure how this gamut coverage number is exactly supposed to be calculated and reported and how the luminance should be factoring in.
there are tactile transducers/bass shakers which provide rumble and dbox which has lift roll yaw pitch motion. I have to temper my subs response so i isolate them off the floor with sorbothane then use bass shakers to replace the tactile rumble. Works really well for preventing low frewuencies from getting out of the room, persons with sub bass sensitivity, while still providing good immersive feedback.
Would love to incorporate dbox and wind, but dbox wont work, so the idea that i can add wind is really cool...not sure, there are an array of color wheels and in some instances 2 color wheels. rgbcym, rgbcy(which is what led overlap does), rgbw, rgby, rgbrgbg orsometimes a dark green segment for p3 gamut.
Blue core lasers are typically RGBY, so I think the 970 is similar.
Actually I think the LK970 is RBGY and the LK990 is RGBW so the LK990 might be even worse.
I think it also has to do with BriliantColor on the LK990 because with that off color accuracy is good, but the brightness is very low, lower even than the LK970. With BriliantColor on the color is not able to be calibrated to any sort of low dE.
Its one of the reasons why i get posterization when doing a 3dlut on the LK970 , its very obvious, the sim2 doesnt have any at all.
I think its the main reason why the sim2 look bright, cause it has a lot of range for color luminance .
Yeah I do think it's why there is so much posterization.
Because it's trying to take a 64% rec709 to a 100% target.I did find that using a different rendering intent in ArgIICMS minimizes the posterization by a lot, and guess what that rendering intent does / is called...“Luminance preserving perceptual” (ArgyllCMS 1.8.3+) uses compression to make the source gamut fit within the destination gamut, but very heavily weights the preservation of the luminance value of the source, which will compromise the preservation of saturation. No contrast enhancement is used if the dynamic range is reduced. This intent may be of use where preserving the tonal distinctions in images is more important than maintaining overall colorfulness or contrast.I don't think it's a coincidence that the rendering intent that sacrifices saturation for luminance and that uses compression produces the best result on the LK970.
The thing is also when i do a p3 3dlut with my jvc rs440 whcih does 90%p3 i get posterization but to a lesser degree than the LK970, a lot less actually but its there and can be seen in some scenes.
With the Sim2 M150 i create a bt2020 3dlut and no posterization whatsoever even though the projectoe can do 90% bt2020 not 100% , i guess the RGB LED is a different animal.There is a workaround in LightSpace called Concatenation method, and DisplayCal also has good solution for it but at the expense of saturation like u said, i just dont know how come the Sim2 has absolutely zero posterization.
I talked to someone who is much more knowledgeable on 3DLUTs about it once before and they said posterization is not really caused by targeting a larger than native gamut.
For example I tested targeting P3 on my rec709 PC monitor which only does about 78% P3 and I saw no posterization with the P3 3DLUT.It's definitely more complicated than that, but unfortunately is past my current understanding of color systems.
This is exactly what I measured on the lk990. It's colour volume is crap. Dave Harper told me the 970 didn't suffer it, quite adamantly at that but never substantiated it with measurements. It just shows he was wrong also.
On the lk990 you can turn off brilliant colour and measure correct values but the light output is 40% lower. Seems like you do indeed have brilliant colour hard coded on. Everyone (mainly Dave) was always telling me I was wrong. Thanks for showing I was not.Sent from my SM-G988B using Tapatalk
If I had to describe it based on trying to calibrate 3 LK970 both manually and with 3DLUT, I would definitely describe the behavior resembles that of BriliantColor and what happens when you have a RGBY or RGBW color wheel.
I also encountered the same behavior with this projector: https://www.projectorcentral.com/Optoma-ZH403-1080p-Laser-Projector-Road-Test.htm"]Optoma ZH403 1080p DLP Laser Projector Road Test[/URL]It had far-off luminance, but with BriliantColor off I could get low dE, but at less than 2000 lumens on a 4000 lumen spec machine.[/size][/color]
On the 970, I am not surprised that BC is hard-coded in. My thought was that BC was always there and adjusted to provide more brightness for the LK990. Once you have seen or measured enough units you can spot BC a mile away. I can see it in the rendering of whites. RGB Laser or high lumen LED is the path forward for DLP in my opinion.
BrilliantColor, amazing name for a feature that f**ks up colors! :cool:
I believe my Hitachi RGB LED DLP has something similar baked in to the different light source powerlevels. I am still waiting answer from Hitachi to tell me how those should work. In my testing the Low-setting is the only usable one with good colors with proper luminance, mid and high settings boost the whites, but color luminance goes dimmer than the Low-setting!
15 Replies
3950 Views