or Connect
AVS › AVS Forum › Display Devices › Display Calibration › ee/ColorBox, LightSpace, and 3D LUT Calibration
New Posts  All Forums:Forum Nav:

ee/ColorBox, LightSpace, and 3D LUT Calibration - Page 8

post #211 of 253
Quote:
Originally Posted by Iron Mike View Post

Let me know how to do a 17^3 with Argyll and export the LUT in the correct ee format, and I'll do some comparison... been using Argyll and DispCal for some of our Notebooks and results were better than basICColor and a few other solutions, but - to be fair - all these other solutions (incl. basICColor) do not support meter profiling (Argyll does), so they were somewhat limited in accuracy...

- M

super, the more the merrier. Follow the tutorial here and post questions and results in that thread. To generate a 17^3 grid + 50 neutral axis patches use the target generator command:

targen -v -d3 -f 0 -m 17 -g 50

instead of the default, although I recommend you use the default for initial testing before moving on to the longer measurement set.
post #212 of 253
Quote:
Originally Posted by zoyd View Post

super, the more the merrier. Follow the tutorial here and post questions and results in that thread. To generate a 17^3 grid + 50 neutral axis patches use the target generator command:

targen -v -d3 -f 0 -m 17 -g 50

instead of the default, although I recommend you use the default for initial testing before moving on to the longer measurement set.

yeah, I only used Argyll so far through dispCal... gotta look at some of these parameters and what they do... is it possible to use the dispCal GUI ?

also, if using A from the command line, I can still use a spectro meter offset, correct ?

what would be the pattern generator ? HDMI Out from the PC or does A support third party pattern generators ?

- M
post #213 of 253
You could use dispCal to calibrate the video card, create the targets and profile. The last step to create the eebox LUT has to be command line at the moment. Yes, colorimeter correction files are supported in the command line as well as spectral sample file corrections. HDMI out via the video card is the example provided, external pattern generators are not supported but we are thinking about time synchronizing to a DVD player like Ted does.
post #214 of 253
Thread Starter 
My Samsung F8000 LCD acts as if it was made for LUTs. It almost reaches full Blue and Green saturation and that helps make a LUT that nearly totally fills out the Rec.709 color space. For LightSpace setup other than the basics, I used the gain control to balance 100% White. The LUT did the rest.

post #215 of 253
Im a little lost and new to these. Just reading the LI site, I see this for the eecolor:
Quote:
On it's own the eeColor box comes with eeColor's TruVue eeColor application software, including eeColor's own pre-set image enhancement LUTs, which include LUTs for different room lighting and colour intent adaptation.

LUT 1: General Color with Standard Skin Tones and Dark Lighting
LUT 2: General Color with Standard Skin Tones and Dim Lighting
LUT 3: Cinema Color with Warm Skin Tones and Dark Lighting
LUT 4: Cinema Color with Warm Skin Tones and Dim Lighting
LUT 5: Vivid Color with Standard Skin Tones and Standard Lighting
LUT 6: Vivid Color with Warm Skin Tones and Standard Lighting

If I connect this between my source and VT50, I can select one of these presets? I get that if I buy LS software I can do my own LUT's, but thats not really a cost Im looking for. My question would be is the box worth it alone, for me with a VT50? Are these just like preset picture modes on a display, or is it only a color affect?
post #216 of 253
If you have CalMAN 5 Enthusiast then you can use eecolor to do a 3375 LUT cube.
post #217 of 253
Quote:
Originally Posted by gadgtfreek View Post

Im a little lost and new to these. Just reading the LI site, I see this for the eecolor:
If I connect this between my source and VT50, I can select one of these presets? I get that if I buy LS software I can do my own LUT's, but thats not really a cost Im looking for. My question would be is the box worth it alone, for me with a VT50? Are these just like preset picture modes on a display, or is it only a color affect?

obviously they do adjust the signal, but they will not magically give u a great picture... they're more for illustration purposes of the different LUT's you could setup for your display...

u can only switch between LUT's if a source is connected, u cannot switch through the LUT's if u just turn on the box w/o anything connected...

naturally, it only makes sense to purchase a LUT holder box if you gonna upload LUT's to the box to adjust / correct your image... rolleyes.gif

u can create LUT's with LS, CM or Argyll...

tip: get LS and don't look back cool.gif

- M
post #218 of 253
Something you guys may want to try using LightSpace.

Use your best profile (4913) make a LUT without Video sizing or any filters, set for 709 leave default Gamma of 2.2. Make sure you didn't change any pre calibration settings in your display, set your meter the same distance off screen using same pattern screen size. Also set your eecolor to LUT 1 (Null)
Use the above LUT you just made and set it for active LUT, now run a profile (4913).
Here is a link for some of the charts after my first try doing the above.
http://www.avsforum.com/t/1466472/2013-panasonic-settings-issues-thread/1110#post_23439581
btw, I did not use any filters or touch up when making the second LUT except for Video size (16-235)

ss
post #219 of 253
Quote:
Originally Posted by sillysally View Post

Something you guys may want to try using LightSpace.

Use your best profile (4913) make a LUT without Video sizing or any filters, set for 709 leave default Gamma of 2.2. Make sure you didn't change any pre calibration settings in your display, set your meter the same distance off screen using same pattern screen size. Also set your eecolor to LUT 1 (Null)
Use the above LUT you just made and set it for active LUT, now run a profile (4913).
Here is a link for some of the charts after my first try doing the above.
http://www.avsforum.com/t/1466472/2013-panasonic-settings-issues-thread/1110#post_23439581
btw, I did not use any filters or touch up when making the second LUT except for Video size (16-235)

ss

the problem with this test / verification is that it will not address the (possible) distortion that the LUT box introduces...

what this test does is demonstrate the mathematical LUT perfection that LS creates, but since you will be using that LUT in your eeColor which has a slight luma distortion, you need to test with the LUT active in the box...

- M
post #220 of 253
SillySally,

just read your post again, maybe I misunderstood (or did I ?):

you are creating a second LUT doing a full 17p profile with a previous 65^3 LUT active in LS ? confused.gif

- M
post #221 of 253
Quote:
Originally Posted by Iron Mike View Post

SillySally,

just read your post again, maybe I misunderstood (or did I ?):

you are creating a second LUT doing a full 17p profile with a previous 65^3 LUT active in LS ? confused.gif

- M

Yes that is what I am doing, except I am using the first profile (65^3) to make a new LUT without using any filters or video scaling it. This LUT I am using as the active LUT, then I run a full measure (16^3) profile.
I am using the same pre settings in my VT60 as shown in my link, also the same window pattern and same off screen distance for my Klein.
With eecolor I set it to the normal LUT 1 (null) used for creating a profile.
Once the profile is done that I used the active LS profile with I simply create the new LUT (709, 2.2) video scale and export, no need to use any filters or touch up any settings in the VT60.

What has happened with the second profile is that it has improved upon Gamma, RGB balance, basically everything. The only downside is the first profile had a 98% Gamut the second had a 96% Gamut.

However here is a screen shot from the first profiles Cube so you can compare to the second Cube from the second profile.Its hard to see but look for the boundary's of the Cube box and where the actual cube is.



As far as the "slight luma distortion" that may or may not be visible, the only way I know to overcome any downstream error or distortion is the pre settings in the display itself.

ss
Edited by sillysally - 8/8/13 at 1:05am
post #222 of 253
okay, I think I'm following 80% but I don't understand how this would improve results... ;-)

I know the active LUT option, I've tested it against in-box LUT results (Lumagen | eeColor) for LUT and box performance verification... the active LUT option is very useful to verify LUT's created by LS, especially when you use LS as the pattern generator via HDMI Out... I've done this extensively with the Lumagen and it demonstrated the perfection of the LS LUT's and revealed the distortion the Lumagen adds to the signal...

what I don't quite get is how an active LUT can improve a 17p that is being used to create a new LUT ?!?

in my understanding, an active LUT would completely distort the profile readings that are then being used to create the second LUT... basically u're distorting the readings of the 17p with the active first LUT... and when you're creating the second LUT, LS will create it based on data that was affected by the first LUT being active but once you make the second LUT active within eeColor, that first LUT is nowhere hence the distortion is gone and the second (now active) LUT is adjusting the signal based on a "different setup" (signal + first LUT)...

example:

I would use my current best 65^3 LUT (which makes the TV already outstanding) as the active LUT - now I do a full 17p - LS will see an almost perfect signal (due to the active LUT) and when I create a second LUT from that fresh 17p it will barely have corrections, as the signal was almost perfect... now I load that second LUT into eeColor... since the first LUT is nowhere active or in the signal path anymore, I get small corrections from LUT 2 for a bare bones signal that needs much more corrections... all the corrections that are within the first LUT...

even if the active LUT does improve the 17p profile (and I would like to know how), the fact that the active, first LUT was done based on a 17p from a different point in time, hence the TV has drifted, meter have drifted etc would already diminish results... or not ?! ;-)

or...did you mean that you are combining LUT 1 and LUT 2 ?

- M
post #223 of 253
Maybe this screen shot will clear this up, on what I am doing and my meaning of a active LUT.



I don't think meter drift will play much of a part in this when using the K10-A, properly warmed up and placed off screen the same distance and area in the window.
Yes there is always a possibility of the display drifting when a full profile is run. However I did leave the display on (screen sweeper) for about 14 hours plus viewing two movies before doing what I did .

As for how it works, I think the paragraph or so that Steve has written about this will better explain it than me.
There is also one other possibility in my mix, and that is I use CM5 pro to set my settings in my VT60. Naturally when I ran the second profile I didn't need to use CM to setup the settings in the VT60 because they were already there.

Doing what I did can be very tricky if not setup just right and the pre settings don't affect the Gamut, Gamma, or get to detailed for RGB balance. Note I didn't use any settings in Gamma detail. Basically using the RGB settings to setup temp.
Also profiling your active meter with a reference meter seems not to work. I tried profiling using my I1pro 2 meter as reference meter.

ss
Edited by sillysally - 8/8/13 at 5:45am
post #224 of 253
Sounds like ..... ITERATION!!!
post #225 of 253
@ ss:

yes, that's the active LUT menu, I used it quite a lot on the past but I still don't understand how this method improves results... smile.gif

IMO, anything in the signal path (possibly) distorts results...

are you using HDMI Out via LS or are you using the Lumagen as a pattern generator ?

- M
post #226 of 253
Quote:
Originally Posted by Iron Mike View Post

@ ss:

yes, that's the active LUT menu, I used it quite a lot on the past but I still don't understand how this method improves results... smile.gif

IMO, anything in the signal path (possibly) distorts results...

are you using HDMI Out via LS or are you using the Lumagen as a pattern generator ?

- M

Yes I can understand your reservations, in the past using CM or LS I didn't get these kinds of positive results redoing a calibration/profile. imo it was a waist of time.

I am not saying I can pinpoint why the big change to the positive, all I am saying is with this last try everything points to the positive.
My guess is the biggest positive is the LUT is more linear in it's response throughout the cube for my VT60.
I don't know why but the Gamut when set to Native in the VT60 seems to be more compressed in the Cube than when the VT50 is set to Wide, when using a normal one pass profile. See the two charts I posted (1st pass/2nd pass) and compare to a Cube chart from your VT50.

When profiling the path is, Lumagen as a pattern generator RGB video out, auto,Null>>>eecolor LUT 1 (null from LS)>>>VT60 pro 1>>> Six feet high speed HDMI cables all the way. When simply viewing Blu Ray or DVD the path is Oppo BDP 95 (4:2:2)>>>eecolor (2 thru 6 Lut)>>>VT60.
With both the Radiance and LS there have been several updates from the last time I tried this.

As for anything in the signal path (possibly) distorting results, if that is happening there is very little I can do to change it. Non the less the profile readings are taken from the display and then corrected/optimized with LS software. That being the case the only wild card is my Oppo, but that is unlikely a cause of distortion as long as I keep any processing in the Oppo to a minimum (digital out).

Mike I can't stress enough the pre profile settings in my VT60 as being very important..

smile.gif
ss
Edited by sillysally - 8/8/13 at 4:12pm
post #227 of 253
Quote:
Originally Posted by sillysally View Post

Non the less the profile readings are taken from the display and then corrected/optimized with LS software.

LS is taking the readings from the probe and if a LUT is active, it adds the offset of the LUT, that is why all readings in your second 17p are distorted, IMO...

But, the ultimate test is very simple:

did you upload the second LUT to the eeColor and then ran a LS QP WITHOUT an active LUT ?

if so, can you post results ?

Thanks !

- M
post #228 of 253
Quote:
Originally Posted by Iron Mike View Post

LS is taking the readings from the probe and if a LUT is active, it adds the offset of the LUT, that is why all readings in your second 17p are distorted, IMO...

But, the ultimate test is very simple:

did you upload the second LUT to the eeColor and then ran a LS QP WITHOUT an active LUT ?

if so, can you post results ?

Thanks !

- M

No I haven't but I will if that helps you. I am not a big believer in using a QP except to get a ruff idea of whats happening with a pre profile setup.
I hope you got the part about not using the Video Conversion when making the active LUT.
If from what Steve says about the internal algorithms in LS for this process is correct, and I have no reason to believe other wise it should work.
Its my understanding that the Video Conversion is just to overcome a weird 'error' (or operational design) of the eeColor, and as you know the Video Conversion is not done until if ever (in this case) the LUT is made. So the first LUT's in this case is made without the use of Video Conversion, only the second LUT is done using Video Conversion. Therefore keeping what the meter is reading for the profile the same on all accounts.

Always a pleasure.

ss
Edited by sillysally - 8/8/13 at 5:03pm
post #229 of 253
Quote:
Originally Posted by sillysally View Post

When profiling the path is, Lumagen as a pattern generator RGB video out, auto,Null>>>eecolor LUT 1 (null from LS)>>>VT60 pro 1>>> Six feet high speed HDMI cables all the way. When simply viewing Blu Ray or DVD the path is Oppo BDP 95 (4:2:2)>>>eecolor (2 thru 6 Lut)>>>VT60.
With both the Radiance and LS there have been several updates from the last time I tried this.

ss, during profiling and when the Luamgen is the pattern source (RGB out), after this entire process is complete, do you have all your sources outputting RGB or 4:2:2 (before the eecolor)?
post #230 of 253
Quote:
Originally Posted by turbe View Post

ss, during profiling and when the Luamgen is the pattern source (RGB out), after this entire process is complete, do you have all your sources outputting RGB or 4:2:2 (before the eecolor)?

turbe, yes once I remove the Radiance and load the LUT into eecolor, I connect the Oppo in its place and set to 4:2:2. The Oppo is my only source for viewing video.
btw, I do leave the Radiance in the chain to check its 21 point step grayscale pattern in both RGB video and 4:2:2 after I have loaded the LUT into eecolor, once I check the 21 point step GS pattern I remove the Radiance and add the Oppo.

ss
post #231 of 253
I take it during the LS profiling process, the VT60 has RGB being inputted, correct?

Have you tried doing the complete LS profiling process with the Radiance as the pattern generator, outputting 4:2:2?
post #232 of 253
Quote:
Originally Posted by turbe View Post

I take it during the LS profiling process, the VT60 has RGB being inputted, correct?

Have you tried doing the complete LS profiling process with the Radiance as the pattern generator, outputting 4:2:2?

Yes RGB Video (16-235) about 11% window patterns are inputted to the VT60.

Yes I have tried 4:2:2 and 4:4:4 in place of RGB Video. My guess is because of how eecolor works with 4:2:2 processing at 10 bit then outputting at 12 bit, I get better results when running a large profile using RGB Video window patterns. However imo its more of a subjective choice.
And its my understanding that a cube is based on RGB when using LS.

ss
post #233 of 253
You may want to use Ted's LS disc in the Oppo (Oppo for patterns and outputting 4:2:2 / 4:2:2 inputted to the VT60) to run a LS verification as well..

Edit: Since your regular content source is 4:2:2
Edited by turbe - 8/8/13 at 7:33pm
post #234 of 253
Did you miss the part about the Radiance outputting 4:2:2 and 4:4:4 along with the same size pattern that ted is going to use in his upcoming Blu Ray disc.
Also if I have the Radiance set for outputting 4:2:2 the VT60 is inputting 4:2:2
Also you may want to read the part about how eecolor handles 4:2:2

For me the only difference using Ted's disc and the Radiance as a pattern source, is there going to be any difference outputting RGB video as a digital signal.?
post #235 of 253
I am going by your replies to..
Quote:
I take it during the LS profiling process, the VT60 has RGB being inputted, correct?
Quote:
Yes RGB Video (16-235)

anyways, you may want to check that and I think you will want to (4:2:2 vs RGB in the display)


.
Edited by turbe - 8/8/13 at 10:23pm
post #236 of 253
Quote:
Originally Posted by sillysally View Post

No I haven't but I will if that helps you. I am not a big believer in using a QP except to get a ruff idea of whats happening with a pre profile setup.

yeah, the QP's are really useful for pre-cal and post-cal verification - naturally I prefer a full 17p for LUT creation, although it has been demonstrated that for very linear displays (Pro monitors) a QP can be enough...

in any case, once you load that second LUT into the eeColor and make it ACTIVE IN THE EECOLOR and then run a QP (with no other LUT active in LS) to verify the LUT, it would really surprise me if the second LUT is better than the first... IMO, it can;t be

the second will be way off (since it is based on the raw TV signal + the first LUT and the first LUT is now not in the path anymore), the first will be as good as it gets...

but let me know, I've seen stranger things !!! smile.gif

Thanks !

- M
post #237 of 253
Mike here is what I think you want.
QP-LUT 2, slider set at 1, delay 0.5 and LLH, loop back, primary's only, no active LUT,

imo what this points out is a full 17^3 profile needs to be run. These results are fairly similar to the 1st full profile run.
The one strange thing is with the QP tells me Gamut is 99%. The Gamut as I said from the second full (active LUT) reported 96%. Also note the compression in the Cube.

Maybe you can ask Steve why this is, because imo there are to many differences for this just to be caused by eecolor box.

.

ss
post #238 of 253
and now post a QP (with the same probe settings) of the first LUT active in the eeColor, again with NOTHING active in LS... the first LUT was hopefully done from a 17p profile with NOTHING active in LS (?)... if so, this QP will show the best calibration possible for your TV...


you don't need to run a 17p to verify the LUT calibration, that'll be complete overkill, QP's are perfect for this

the Gamut (96% | 99%) doesn't mean anything, don't worry, but the second LUT gamut is off because you had the first LUT active in LS which distorts everything...
post #239 of 253
what do you mean by "slider set at 1" and "loop back"... did u mean CLOSED LOOP ?

also, run QP with primaries + secondaries - no need for LLH with K10-A, takes 5 times as long with no benefit...

with the i1D3 the LLH is crucial...
post #240 of 253
Quote:
Originally Posted by sillysally View Post

Its my understanding that the Video Conversion is just to overcome a weird 'error' (or operational design) of the eeColor, and as you know the Video Conversion is not done until if ever (in this case) the LUT is made. So the first LUT's in this case is made without the use of Video Conversion, only the second LUT is done using Video Conversion. Therefore keeping what the meter is reading for the profile the same on all accounts.

SS, We use VideoScale for eeColor because the eeColor's LUT Table Memory Stores the LUT there @ 0-255 (0-1024 @ 10bit). When you are using Lumagen as a Pattern Generator the Patterns are generated @ 16-235, so with the VideoScale, Lightspace with a complex algorithm is moving the Black from 0 to 16 and the 255 to 235 and this is interacting to the whole LUT Data.

For Lumaget 3D-Cube Calibrations , VideoScale is not required because Lumagen's LUT Table Memory Stores the LUT there @ 16-235 already.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Display Calibration
AVS › AVS Forum › Display Devices › Display Calibration › ee/ColorBox, LightSpace, and 3D LUT Calibration