Dave's settings cannot in any way shape or form come close to giving you the same performance as a Panasonic, Lumagen or MadVR. Let's look at this from the most basic imaging science approach:Dave is using nothing but the projector settings. For the BenQ 970 (I haven't seen the 990, but I imagine it is similar) you can set the gamma to a power law, in which he choses 2.8, but that is ALL you can do. There is no parametric setting to adjust gamma along the luminance levels. If he is trying to do a PQ calibration (which he stated before he is) then that gamma curve would look like an S curve, which you are not going to get. He is most likely trying to emulate this with the contrast/brightness controls, but that is going to kill the white and black point, which is what these controls are for. With any of the other solutions the S-Curve is not only taken care of in the source device, it is changed on a title by title or frame by frame basis. Then there is color. He wants to feed 2020 color in, but these are projectors that can't even do full 709. So anything beyond what the projector is capable of is clipped and layered on top of itself. You also have to do major adjustments of color/tint to realign the gamut to what 2020 is, which completely kills your linearity even inside the gamut you can do. This causes clear issues with greens, reds and yellows. By converting the matrix to 2020 you lose NOTHING in terms of what you would have had before, but you are accurately taking the 2020 container and mapping it to the display's gamut potential. This would nearly eliminate clipping (you're still not getting the full 709 gamut) and improve linearity inside your gamut. The thing that is so frustrating about Dave's argument is that it is like he doesn't even understand his own argument. He is trying to do the exact same things these devices will already do (create a tone map and remap color), but they will do it far better. And this is just for the HDR side of things. We are not even talking about the limitations of contrast and its effects. I'm not going to get in to all the other arguments about people using or not using the settings properly and all that other nonsense. Reading that thread through speaks volumes enough. If you couple the 970 with an outboard device you can get a decent HDR image out of it. It is extremely bright and very sharp, and would be a good fit for someone looking for those qualities in a solid state device for their room on a budget. The 9060 gives you better color coverage and nearly eliminates RBE completely, but it has issues with HDR tone mapping and creates a difficult proposition for using an outboard device to do the tone mapping for you. For SDR I like the 9060 better than the 970, but it isn't nearly as bright so that may be an issue depending on your room needs. Again, I haven't seen a 990.
Well, Dave, we can only use the settings you tell us to use. So if Kris set it wrong, so did you. The LK990 I played around with could not do more than REC709 and this was measured with and without your settings. So I don't know how you guys were achieving these results. Can you point me to posts that show gamut coverage results from 12GAGE, tnaik4 and others?
Dave, you realize you ALWAYS have some sort of excuse, right? What is the point in you giving us your settings if you need to do an additional calibration? Why not just tell us what we should be targeting for calibration instead? You could tell us that, then we could provide calibration results showing you that we achieved your targets and then we can tell you what we think of the image. Wouldn't that make more sense?It just seems you do it your way on purpose so you always have an excuse why we aren't agreeing/seeing the same thing you're seeing.Yes, please provide the MULTIPLE graphs showing >REC709 coverage that you claim corroborate your findings. The only thing I could find is this... which I wouldn't call a meaningful measurement in your favor:https://www.avsforum.com/forum/24-digital-hi-end-projectors-3-000-usd-msrp/2917348-benq-lk970-4k-dlp-laser-12-999k-msrp-44.html#post57772038You can click on the quoted post to see his original image. Which clearly shows he isn't hitting more than REC709...Yes there are points outside of the triangle he measured but there is a whole section of REC709 that can't be hit.
Dave, it doesn't matter whether or not a display accepts an HDR image because what you're doing can be done on any display that has a decent set of image controls. Trust me when I say, those of us who've argued with you on this topic know exactly what you're trying to accomplish. It has nothing to do with us misunderstanding you. Manipulating gamma like you're trying to do on these BenQ models isn't anything new. We did just that with the older JVC models. We've all been through this before.So now that we have that out of the way, can you please provide some actual data to back your claims up or are we just going to argue over the semantics of us understanding one another?
What is the point in you giving us your settings if you need to do an additional calibration? Why not just tell us what we should be targeting for calibration instead? You could tell us that, then we could provide calibration results showing you that we achieved your targets and then we can tell you what we think of the image. Wouldn't that make more sense?
The 970 nor 990 can't do DCI P3 within the BT2020 container. So you are having to remap to what it can do, which is a little less than Rec709. So you need a good method to remap BT2020 to Rec709. So it looks to me your choices are settings that you would have to constantly tweak per title or use an outboard source, like the Panasonic, MadVR or Lumagen.
Guys, this is not going to go anywhere. It is absolutely 100% clear that Dave is going to do everything he can to avoid having to do anything that would either allow you to accurately try and emulate what he has done or provide anything that backs up any of his claims other than him arguing. Any of the people here that would be interested in doing any of this are more than aware of how to calibrate a projector for HDR and would gladly provide clear evidence of not only the measurements and settings they used, but image showing the end result. Dave has provided nothing but pushback and excuses. Just walk away, I think he's shown more than ever it just isn't worth it.
I think Kris explained his theory pretty well already. He mentioned that you can get far better results if you just calibrate to D65/REC709 (or as close as these units will allow) with a normal 2.2/2.4 gamma and use a quality tone mapper to convert HDR to SDR. The Lumagen Radiance Pro, madVR, UB820 or UB9000 all seem like better candidates. As I said before, we got far better subjective results going down this route instead when I demo'ed the LK990. I do the same thing with my JVC projector too. I rarely let it do the tone mapping conversion as madVR (and the UB9000 I'm currently reviewing) does a noticeably better job. Kris even provided photos showing this.
Kris didn't provide photos comparing it against a JVC. He provided photos of the settings you suggested (both on the LK970 and UB820) against using straight up tone mapping with the unit set to SDR standards. The photos speak for themselves. I linked to them several posts ago.
275 Replies
90363 Views