Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Best LCD grading monitor on the planet

Jeff, I'd like to know how you know it's an "8bit LG" panel... the HP User Guide states:

"As noted earlier, the LCD module in the LP2480zx monitor provides a 10 bits/color (30 bits/pixel) input, with true 10-bit drivers within the LCD itself."
 
http://www.hardforum.com/showthread.php?t=1366545

An extensive forum post on the Dreamcolor monitor. Apparently there are very visible dithering artifacts, so so much for the whole 10-bit thing. (Turning off the dreamcolor engine would get rid of the dithering, but then the colors wouldn't be accurate.)

2- Also, the Dreamcolor engine is not enabled for interlaced sources and "YUV" (Y'CbCr) sources.

From http://www.justechn.com/2008/06/10/hp-dreamcolor-lp2480zx-lcd-display
Sean,
I wrote HP about your problem and I just got an email back from one of their engineers. Here is what he had to say:

I am a engineer working on the LP2480zx monitor. Sorry to hear that you are having difficulty getting the results that you would like from the HP DreamColor monitor. I was forwarded your post below describing your SDI setup. First of all let me assure you that the issues you described do not indicate anything wrong with the monitor.

* The DreamColor Engine is only enabled if the input is:
o Progressive (not interlaced) AND
o RGB (not YUV)
* If the input signal is Interlaced OR YUV, then the DreamColor Engine is disabled and there is no color management (the colors are unmanaged, reverts to Full (Native) gamut and the color presets menu is grayed out).
* With an SDI input, you must use an adapter that converts to progressive RGB at the output.
* The only adapter we have found that does this is the Gefen EXTHDSDI2DVISP HDSDI to HDMI adapter.

I am not sure which product you are using to convert the SDI signal to HDMI, but if it offers the option, please set it to convert the Yuv signal to RGB, and make sure that the output is non-interlaced.

If you don’t have those options we recommend the Gefen SDI to HDMI converter. Make sure that it has the most recent firmware. Then follow the adjustments below to make sure that the black level is set correctly
Click to expand...
 
P.S. I spoke to Wes at ecinema today and their new DPX and Pro displays should be shipping sometime in january ($8k for the Pro, $12k for the Pro with dual link, and $40k for the DPX). They will also be offering a DVI version of their pro display to use as a GUI monitor along side a new 'finger printing' calibration system to sync them both up perfectly.
 
Glen,
Mr. ToastyX brings up some valid points, but some of his points are subjective (like "sRGB still doesn't look right to me."). On my Dreamcolor I see some dithering artifacts in the REC 601 colourspace but don't seem to see any in the REC 709 (maybe my eyes are just tired). Of course, there shouldn't be any dithering in any colourspace, so I'll be looking into that with HP and my reseller.
 
I've worked with the DreamColor display a couple times now and almost bought one. I would buy one if I could get it in a 30" 2560x1600 size...

But I have to say that the above article is pretty much spot-on. The Dreamcolor engine is not available in all modes and yes, there is visible dithering to approximate all those colors on, what is still in reality, an 8bit display panel. However, the dithering is handled well and the monitor performs admirably for its relatively low price. For a budget grading solution, this is a very good option, provided you have the means to properly calibrate it.

jdcineaste said:
Of course, there shouldn't be any dithering in any colourspace, so I'll be looking into that with HP and my reseller.
Click to expand...

Dithering or banding, take your pick.. The panel in use can't display all the colors produced by the monitor's internal 30bit processing. There's a whitepaper from HP floating around out there that sheds some light on the dithering algorithm used, but I'm not sure where it's at off the top of my head. I seem to recall that it may have been linked to in one of the threads on these forums about the dreamcolor.
 
subjective (like "sRGB still doesn't look right to me.")
Click to expand...
Have you compared the dreamcolor to a broadcast CRT? Do the colors look right?

I think that a lot of the time the theory is wrong or doesn't capture all the complexities of human vision... I wouldn't immediately dismiss subjective comments because they are subjective (especially since human perception is a subjective thing).
 
So, they're doing what TV Logic is doing. High quality dithering of 10 bit inputs to achieve an accurate picture on an 8bit display. Looking forward to the day when we can see highly precise modern display technology that doesn't have to rely on trickery. Lots of variables though that make this easier said than done.
 
No Glen, I haven't compared the DreamColor to a CRT broadcast, but then again, I wasn't the one doing the review/testing. That comment ("sRGB still doesn't look right to me") wasn't the only thing that I found to be subjective either. By the way, I do have a DreamColor monitor.

Dan Bennett, from HP, has replied to one of my questions on one of the forums over at creative cow and has provided much interesting information concerning this display (see the color forum).

Note: I tried to enter the link, but it kept getting changed to something completely different, for some strange reason.
 
jdcineaste said:
Note: I tried to enter the link, but it kept getting changed to something completely different, for some strange reason.
Click to expand...

Links to Creative Cow is not allowed. Next time you will be banned! ha ha. just kidding

I was kind of hoping Dan Bennett would answer your question about whether the panel is 8 or 10 bits. I still haven't seen any proof, and I'm not going to take my monitor apart to find out. :detective2:
 
But he does clearly state the following:
"* A true 30-bit panel (despite rumors to the contrary!) which allows the HP DreamColor Engine to map input colors to a true palette of 1.07 billion physically displayable colors. This reduces color error when performing the color and tone mapping."

And this is after stating that the DreamColor engine is 36 bits.

I just hope he'll explain how we can get the best rendering without dithering.
 
Note: I tried to enter the link, but it kept getting changed to something completely different, for some strange reason.
Click to expand...
It's a long story (possibly because some people were unhappy about cow's past censorship/anti-competition policies). The short answer is to use tinyurl.com on your links.
 
I purchased the monitor on the basis of what I read in HP marketing materials
Click to expand...
Nooooooooooooooo.... lol. I think there's a danger in purchasing anything expensive based on marketing materials... products tend to differ significantly from what is advertised.

2- What is the point of having a 10-bit panel if the end result is that the image has visible dithering artifacts? (Granted, I have not seen this for myself.) In theory, the image ought to be good... in practice, it doesn't look like this is the case.

To quote the HP employee:
2 - 10-bit pixels (when you get that 10-bit graphics card!) are passed straight through to the panel; at least in "Full" mode. Of course, in a color managed mode such as Rec 709, there'll be some reduction of the monitor's native color space to map onto the Rec 709 color space, which means that not all 1.07 billion colors are available. It's hard to say what the real number is, but if you imagine, say, 800-900 million it wouldn't be too far off. Still much better than an 8-bit panel. :-)
Click to expand...
A 10-bit Rec. 709 signal actually doesn't have 800-900 million colors to begin with. There is significant waste because the Y'CbCr color space defines many out-of-gamut and illegal colors, including values that call for negative light (which doesn't exist / is an impossible color).

Or looking at it from the monitor's point of view, I believe that significantly less than 800 million colors that the monitor can reproduce would be used for a Rec. 709 image. (Here it's about the relative areas of the color gamuts, rec. 709 versus the monitor's maximum gamut. Too lazy to calculate it, but the wider the gamut of the monitor is... the more thinly spread out its bits are and therefore less of those bits are relevant to displaying a Rec. 709 image.)
In practice it probably doesn't matter, but I would point out that this is likely a case of technical BS.

And this is after stating that the DreamColor engine is 36 bits.
Click to expand...
You can think of it as bottlenecks... usually, the bottleneck in terms of bit depth is in the panel of the monitor. The various parts of the monitor have different bit depths.

--e.g. the input can be 8-bit, 10-bit, etc. RGB versus Y'CbCr makes a difference, as you can get rounding error converting between the two. Same deal with full-scale RGB versus studio video-scaled RGB. Apparently the Dreamcolor can't take YUV signals, so you'd need to convert that to RGB (likely losing some bits in the process).

--The monitor has some internal processing. 32-bit floating point precision per channel would not be unreasonable (so that would be 3 channels X 32-bit). If you want to do things correctly, you need to do linear light processing... this takes a lot of bit depth. You may be able to approximate the calculations well enough with a 3-D LUT with 12-bit depth... that I don't know.

Or if the monitor naturally had a Rec. 709 gamut, then you wouldn't need to do linear light processing.

--The panel itself has a certain bit depth. Because the transfer function of the display is far off the from the ideal transfer function (that of a CRT, in the case of broadcast) then the bit depth is spread thinly for some parts of the tonal range.
This is a bigger problem for plasmas and DLPs, since their transfer function is linear and requires major work. e.g. A 10-bit plasma panel is not good enough for professional work... shadows will have noisy dither artifacts.

--The display might also lose some available bits after calibration, e.g. if setting the white point is done digitally. (This can also be done by varying RGB LED backlighting... but then I think you have to stabilize the LEDs.)

Or, you take the simple approach... and see if the end results are good. Can you see banding or dither-related artifacts?
 
D-ILA projector for grading

D-ILA projector for grading

I've just relocated my studio and I'm looking into a ceiling mounted projector for grading. I'm currently using the JVC 24" DT (which I plan to keep as an extra reference). I'm hoping to spend $US10-15k. So far the front-runner is the new JVC RS20 (also sold in other countries as HD750). They're marketing it as a 'reference' projector. It's supposedly THX certified (from what I understand you get a preset color profile that's been calibrated to THX standards). It's 1080p, 50,000:1 contrast ratio & 900 lumens. Should be perfect in terms of brightness for the size of room I'm setting up (20' x 20').

Keen for any thoughts on the viability of this for grading. I'm running Apple Color with an Eclipse CX console through a Kona 3. The proj takes HDMI 1.3 so I was either going to use a Blackmagic converter to go from HD-SDI to HDMI (which I've used with some success on a Bravia), or possible JVC's external video processor which is upgradeable to HD-SDI.

The other interesting option is a motorised anamorphic lens to go in front. JVC sell it as an RSAL1 but it's really a re-badged Panamorph UH480. These are mostly designed to enlarge 2.35:1 content that's letterboxed within 16:9 - to fill a native 2.35 screen. Could be useful when editing RED 2:1 footage which through the Kona3 gets letterboxed inside 1080p24. The RS20 has a trigger output which can automate the anamorphic lens moving into place when a certain mode is selected. Could possibly come into play with RED footage that has been shot ANA - not sure yet whether it can stretch that far.

As far as a screen goes it seems that simple matte white is the most color accurate option - with some kind of black velour trim to improve visibility. High gain reflective screens apparently affect color reproduction and viewing angle.

Any suggestions would be appreciated. Obviously if I could pony up for a 2k or even the new JVC 4k (or even better - Meridian's modded JVC 4k - the 810) I would. But this is sort of an interim solution while I finish the edit and grade on the feature I'm currently working on.

Thanks in advance,
Dan
 
Evin Grant said:
We have a clibrated JVC DLA RS-1 in our Baselight suite...
Click to expand...

I'm curious:

What exactly is the projector calibrated to? This particular projector is a high end consumer unit, so it does not have internal lookup tables and the ability to do color space conversions - in other words, it is what it is. So if one is looking for accuracy to a particular specification for professional use, the only way to achieve that is through visual emulation provided by creating a 3D LUT after performing a profile on the projector once you've done a basic setup based on test charts. Is that basically what you mean by "calibrated"?

There are a lot of high end consumer projectors that "look nice." But for professional use, it's about accuracy, adherence to a standard. Looking good is not necessarily what one is after.
 
You must log in or register to reply here.
Back
Top