Radiance 3D LUT (5x5x5 Cube) Calibration with CalMAN 5 - Page 3

True, but how different is the player from standards? My guess is not that much.

There is no way to do a cube without an automated pattern source, it's just not doable. Even if someone made a disc with all 125 swatches having to use IR would take a 20 minute quick and easy processes and turn it into a 2+ hour ordeal.

If you are going to spend the money on a lumagen why not also go get a blu-ray player that is bit perfect. I believe there are some cheap panasonics that work in the $100 range.
http://www.hometheaterhifi.com/blu-ray-players/blu-ray-players-reviews/panasonic-dmp-bdt210-blu-ray-player-for-the-home-theater/page-4-on-the-bench.html

@ sotti: I agree, and IMO I would go for the Oppos, I was just asking about the possibilities of the workflow to squeeze the last little bit of accuracy out of there...

@ joshua: since you did not state your point, I read the article... up until page 5 of the Spectracal statement, that article was exactly saying what my point was, that source devices can distort the signal and you would need to compensate for that by calibrating with patterns from the source device - but the Spectracal statement on page 5 of that article seems to say the opposite, so I am confused whether I am missing or misunderstanding a fundamental part of the calibration logic...

I mean you can only accommodate for device specific issues when you analyze (reference) data from that specific device... !? Or not ? ;-)

Thanks.

- M

Analyzing the data is easy.

Correcting it is much harder.

If your blu-ray player is off by say 3%, so instead of 55% you get 58%, do you adjust 55% or 60%? If you turn 55% down so that the 58% swatch is now dead on, but you've dropped the 50% swatch because it lived inbetween 50 and 55% so droppping 55% adjusted your 50% point as well.

I believe the lumagen does have some correction per input, and that would be the place to correct a funky component. The only way you can get things right is by starting at the display and calibrating everything to reference working backwards. That means doing the radiance grayscale and cube with the internal patterns, then correcting for a funky device on the input adjustments into the radiance.

@ sotti: Does the Radiance use one cube and one grayscale for all inputs or is it possible to do a separate grayscale and cube for each input ?

Thanks for the info and explanation.

- M

I'm not sure on the exact capabilities of each radiance, but they have several different types of configurations. The grayscale and cube live in a CMS, and at least on the XD I play around with I have 8 CMS that I can map to any input/output, setup to auto switch and do all kinds of crazy things.

Hi Mike, I posted the article because source devices can and mostly do distort the video signal. The source device should be considered in the video calibration process, but only as the last step of the video chain calibration and only once you've calibrated everything in the video chain in front of that source (display, then processor, then source). As Joel described, if you are not fixing the video chain closest to the display before working your way backwards during the calibration process to the source, you may end up making incorrect adjustments. The goal of the calibration is for the entire signal path to be optimized from source to display, and not considering each piece equipment that has adjustment in that chain is greater potential for error when only correcting at the source (blu-ray/DVD player).

I hope this helps.

Sotti & Joshua: thank you for the explanation.

Maybe I am still misunderstanding this and then please just ignore this answer - but your answers only make sense to me if you have to share grayscale & cube calibration settings across multiple inputs. THEN you would want to calibrate with the best reference patterns from the Lumagen or some other reference signal generator.

I don't have the Lumagen yet, but I thought maybe it was possible to have a full blown calibration configuration (grayscale, gamma, gamut etc.) for EACH INPUT, ergo for each specific device (that's why I initially asked). No general grayscale or cube configuration is shared across the inputs. Every input has their own complete calibration. In my current understanding - and it's probably wrong again - if you run a complete calibration for a specific source signal device you get the most accurate calibration for that specific device on that input.

If you're generally calibrating the Lumagen with the included patterns for your TV (grayscale & cube) and then use that general calibration for all inputs (ergo all other signal sources), you might be off (maybe just a tiny bit, if even as sotti said before) for some signal sources and you would have to tweak a little bit (as sotti said). But if you can have a full blown calibration for each input device, then obviously you would consider that, right ?

I do understand though, that the integrated patterns in the Lumagen are gonna be very accurate for almost all inputs and sources and it will be a much faster auto calibration using Calman, and probably makes more sense to spend 10 mins on the calibration then hours and hours with maybe little to zero improvement... I was just generally elaborating on the possibilities and options of the Lumagen using Calman...

Thanks again !

- M

You can have separate calibrations for every input.

The issue is actually being able to calibrate that input. If you have a truly funky source, it'll never work right, even basic old school 10 point grayscale, if the points don't line up, you'll be fighting with the fact that the adjustment points don't line up with the patterns.

Believe me this is a huge issue for AutoCal and something we run into internally testing hardware.

Where multiple calibrations do work is if you want to target SMPTE-C primaries for DVD and rec.709 for BD, or one cal for a gamma of 2.0 for daytime and a gamma of 2.4 for nightime. On the radiances that have dual outputs, you can drive two different displays from the same source and have them both calibrated.


But you are right, if you wanted your DVD player to look perfect and it was off, the only way to get it 100% right would be to use it as the pattern source. But with a cube calibration that would be prohibitively time consuming, after your first attempt you'd be running to the store to grab one of those BDT210s and just forget about it. Not to mention nobody has the right 125 patterns on a disc to begin with, so it's all a bit of a moot point as you simply can't do it at all. Even then you wouldn't want reference patterns, you need patterns cooked for the exact way your player is skewed.

OK lets all get ready the Radiance Cube 125 3D LUT beta firmware is going up later today from Lumagen. CalMAN 5 Home Video is going RC "release candidate" later today or tomorrow as well. So those with a Radiance and CalMAN 5 should have a fun time this weekend. The update keys for CalMAN 5 Home Video should also be going out today via email.

One of the reasons CalMAN 5 Home Video is being made as a release candidate is I’m sure tweaks will be made to the Radiance firmware and maybe CalMAN code once in the hands of more people. But even as a release candidate for Home Video CalMAN 5 has been out in the wild for some time now in the commercial markets.

Every 3 to 4 years comes along a significant approach to how we calibrate displays today marks one of those landmarks.
Edited by derekjsmith - 9/13/12 at 1:06pm

It's up - http://www.lumagen.com/testindex.php?module=updates

Radiance Cube 125 3D LUT beta firmware released into the wild!

CalMAN 5 Home Video RC is going through some final testing overnight and will be up in the morning. The CalMAN 5 keys will also start showing up in email boxes about the same time.

Hey Joel I have a question about "pre calibrating" the projector or display prior to running the cube autocal. Sorry this seems to be beaten to death here, but I dont fully understand what you mean. Are you saying to take the Radiance out of the chain and calibrate first using Calman 5, a meter, and the projector's / display's native contrast, brightness, and gain / offset controls only? Or to do a manual calibration with the Radiance in the chain (and projector at "default" settings) before running an autocal session and the cube? I dont think I understand how that differs from just running autocal from the start with the radiance in the loop, since that autocal session would calibrate grayscale itself.

You don't need to take the radiance out of the chain, just reset it's calibration controls or use reference patterns. Then do all the normal stuff, brightness contrast, cuts and gains. If you have gamma control in the projector it would be fine to manualy do that as well.

After you get that basic calibration done, then you move on to calibrating the grayscale, then the cube.

You're working backwards trying to get everything as perfect as possible.
Display Controls -> Lumagen Grayscale -> Lumagen Cube

The Cube workflow we'll be releasing today or tomorrow will walk you right through it.

In fact the Radiance has all the patterns you need to pre-calibrate a display.

As we are now released from NDA, I just wanted to post a copy of my last (as in most recent, not the very last I hope:)) post in the Calman 5 beta forum. I'm sure results will vary depending on each display and user, but as far as I'm concerned, I couldn't be happier. And those who know me know I'm not easy to please... Here we go:


Just wanted to say thanks again to Spectracal and Lumagen for this breakthrough. I've just done a new calibration with a 2.3 target (my usual target in my bat cave) and the results are nothing but spectacular on screen.

It is by far the best my rs45 has ever looked, both regarding color accuracy and dimensionality/shadow detail.

Having both the right chromaticity and gamma at every level is just spectacular, especially in scenes with low light.

I went through a series of clips I know really well, and the level of detail is simply incredible. Some colors which used to suffer from the necessary compromises in a manual calibration are fully restored.

Well done again!


Ps: I'm using a Lumagen radiance mini with the latest beta f/w, an i1pro2 as a reference meter and an i1display Pro as a field meter. I do a basic calibration before hand, especially setting the gains at 109%, but all the hard core work is done automatically in 45mn by Calman pushing 2 buttons.

Sent from my iPad using Tapatalk HD

This display could not reach full G & R saturation. Otherwise, very nice....

Do you have "before" data?

Here's some Pre Data:
Personally I think the Color Checker chart is much more informative than sweeps.



Here's my post data.




C6 profiled to i1 Pro on a Panasonic VT25

Nope. I was eating dinner during the grayscale calibration and when I got back to the computer the connection with the Radiance had been lost and I had to restart the program. It might be the new FW because it didn't happen with the previous versions.

Looks good, Buzz! Nice work.

The color checker is cool....

Still not perfect but a very nice improvement!!

Mmmmm. Color Checker isn't in my workflow. Where do I find it?

That's pretty damn close to perfect.

Foliage which is one of the worst offenders about this far off at a dE 2000 of 1.5 looks like this

I added it, because I like it.



Here's the layout, with design mode you can import/export layouts.

Colorchecker.zip 12k .zip file

Unzip the .clfx file, then in design mode you can right click a layout category and select import layout.

Here is a LG that has a very odd gamut and not so great gamma:

Pre:


With CalMAN 5 press two buttons and 36 minutes later you have:

Very impressive.

Yeah but I'm ARI....Anal Retentive Impaired.

Seriously it looks damn fine.

This is going to spark some business for Lumagen and you and your competitor.

Same calibrations but using Joel favorite Color Checker charts:

Pre:


Post: the black dot in the middle of no where blue is absolute black


Same display but calibrated with the traditional method of WRGBCMY CMS with the Radiance: to fix RGBCMY with CMS we ended up with under saturating all the rest and actually made it worse (look at those DeltaC's)


The Color Checker charts by the way are based on Pantone color checker targets used by most professional publishers
Edited by derekjsmith - 9/13/12 at 7:28pm

Thanks, Joel. It works like a champ. With the revised workflow I LUTed my 65VT50 and the results are excellent. After first watching some reference content I added a Darblet between the Radiance and the plasma. 40% Darblet is now doing what 55% used to do. I'm presently watching today's Sprint Cup race I recorded and am pleased to report I'm watching the best picture I've ever seen. Kudos, guys.

Question before I drop ~2000 on a Radiance Mini without my wife knowing....(you all know how dangerous that is.....)

Can it fix the bad color tracking of a Sharp Elite?