AVS Forum banner
1 - 20 of 341 Posts

·
Registered
Joined
·
144 Posts
Discussion Starter · #1 ·
I probably should have posted this in the calibration forum but then thought it might be a welcome distraction for those of you anxiously waiting for your deliveries of new JVC's



I have a HD750 (RS20) projector and a Lumagen Radiance XS. I have been using the projector's CMS to correct the gamut and the Lumagen to fine tune the grey scale and gamma. I noticed some slight solarisation on some scenes and I wondered if the Lumagen CMS may produce a better result.


I used Calman 4.1 and calibrated both CMS's at 75% stimulus. The picture using the Lumagen CMS looked different and seemed more saturated compared to my calibration using the JVC CMS and also compared to the THX mode on the JVC.


I could not understand it so I then checked the CMS performance of both at 25%, 30%, 45%, 55%, 60% and 95% stimulus . They were quite close but I retouched the calibrations to make sure that they were as closely matched across the range of stimulus.


I still found that the images produced by the 2 methods were significantly different. To prove to myself that I was not imagining things or making measurement errors, I put up a colour bar test pattern and photographed it in both modes (see attached). As you can see the primaries and secondaries are almost perfectly matched.


In the process I noticed that the menu item "Color management" on the JVC menu looked quite different so I set about playing around with some mixed colours which then showed up significant differences especially with the orange middle bar.


This convinced me that I was not going mad, but I still wondered why there is such a difference. Even more importantly which one is correct?


I emailed Jim Peterson from Lumagen for his view on it


While waiting for a reply (very quick as usual) from Jim, I did some extended viewing of the two options ignoring the differences in colours. I was very pleasantly surprised by the improvement in the image when using The Lumagen CMS. The image had a lot more detail, was more 3D and looked like a thin veil had been removed.


I had previously noticed the benefits of the Lumagen scaling which is well documented but I have never seen anything written about the benefits of the Lumagen CMS compared to those used by display devices. The common recommendation seems to be to correct as much as one can in the display device rather than in the processor. From what I have seen, that advice appears to be very wrong.


Even before I got Jim's reply I had decided to stay with the Lumagen calibration but still wondered which colours were correct.


Here is his explanation:

Begin Quote:


One cannot properly correct for primary and secondary gamut errors in the HLS color space, which is what other consumer products attempt to do (or worse yet some try it in the YCbCr color space). It is not a linear color space. HLS color space is a "double cone" color space (it is drawn as a double-cone with bottom cone upside down with the other cone on top). So, conceptually, the problem with HLS is walking "straight" on the surface of a cone is not the same as walking straight on the surface of a cube.


To do gamut correction right one must use a linear-Gamma rectangular coordinate system. This is what we do, at the cost of a lot of gates.

All three variables (R,G,B) are linear in our implementation. This also gives us correct and independent control of the primaries and secondaries.


Please note that if you have "color decoder" errors (i.e. if the conversion from YCbCr to RGB is wrong) in the display -- as is often the case -- they need to be corrected in the YCbCr color space, which is why we have both Color, Hue and offsets in the YCbCr color space, plus our 3D linear-Gamma RGB color palette for Gamut correction.

End Quote


I am surprised that no one has reported this before and would be very interested if anyone else has noticed this or has an alternative view.


It also may be a justification for buying a RS40 and Lumagen instead of a RS50!



 

·
Registered
JVC RS500U, 2.35:1 125" screen, Marantz AV7705, 11.2 Def Tech speakers, Sunfire Theater Grand 400x7
Joined
·
3,277 Posts
Another benefit from paying less for the PJ and using that money towards an external CMS is that, like separate pre/pro in audio, you gain more flexibility. You're not likely to upgrade the VP, but probably will upgrade the PJ sooner and can go with a cheaper model then too.
 

·
Registered
Joined
·
5,668 Posts
In the past I have owned a Radiance and a HDQ. Great products and great support.


But, as respects the CMS on the JVCs, I have not been able to measure any shortcomings. There were problems with the CMS when it was first introduced on the RS20, but they were corrected in firmware. So, JVC's CMS functions properly and everything is as it should be. Did you upgrade your firmware?


Past JVCs show some solarization and there is no video processor that will eliminate it.


A Radiance has far better gamma controls than the JVCs do natively.


It is not illogical to do a RS40/Radiance pairing instead of going with an RS50 or above. It comes down to which is cheaper I guess. Either way, points can be made for either approach.


Do the Lumagens work with 3D?
 

·
Registered
Joined
·
99 Posts

Quote:
Originally Posted by Lawguy /forum/post/19601763


Do the Lumagens work with 3D?

The Radiance does have 3D support and even though still in beta it already has some very impressive functionality (3D image stretch/shift for anamorphic lenses, pretty much full CMS/enhancement processing etc.) Many people seem to be a bit annoyed with lack of for example vstretch for 3D in JVC DLA-X projectors so below you can find explanation what the Radiance currently can do with 3D.

http://www.convergent-av.co.uk/forum...9974f998472298


And here's the latest on what new is planned for 3D in near future, nice stuff there too IMO

http://www.convergent-av.co.uk/forum...9974f998472298


Sorry if the links don't jump to correct message, I did not find a way to link to a specific message in the Lumagen support forum. But those threads will give one loads of info about Radiance usage in 3D age.


Being a Radiance XS owner and a JVC D-ILA'oholic getting a RS40/X3 to replace my current HD100/RS2 is a no-brainer.
 

·
Registered
Joined
·
174 Posts
Interesting. I have an RS35 and a Radiance XD. I've used the JVC CMS under the same assumption of use the display controls first. As you've pointed out, when the primaries are correctly set, an actual picture looks under saturated, so I just bumped up the saturation.


I also briefly had a Sony VW85 before, which has Hue and Saturation controls, not Lightness though. It's Normal gamut setting was apparently on the button for REC709, but it looked under saturated on a real image.


At work there are a couple of Samsung LCD TV's, 55 inch A650's. Their CMS uses RGB rather than HSL, and the colours are good. I wonder if the CMS controls offered are an indication of their accuracy?
 

·
Registered
Joined
·
3,866 Posts
I'm wondering how my VideoEQ Pro sits in with this: I've not got any built in CMS (HD350) to compare it to, so can only rely on what my calibration told me was correct (rented i1Pro, so should be good). FWIW I find the image less saturated than without the VEQ, but then the HD350 is excessively oversaturated anyway...
 

·
Registered
Joined
·
144 Posts
Discussion Starter · #7 ·

Quote:
Originally Posted by Lawguy /forum/post/19601763



But, as respects the CMS on the JVCs, I have not been able to measure any shortcomings. There were problems with the CMS when it was first introduced on the RS20, but they were corrected in firmware. So, JVC's CMS functions properly and everything is as it should be. Did you upgrade your firmware?

That is the whole point of this thread.


Both aprroaches measure correctly but real world images look different.


This is because we all only measure 6 points. As my test shows (and Jim confirmed) the method used by JVC gets the primaries and secondaries right but has problems of varying degrees with the colours in between.


I am using the upgraded firmware in the projector
 

·
Premium Member
Joined
·
6,249 Posts

Quote:
Originally Posted by Htnut2000 /forum/post/19604437


That is the whole point of this thread.


Both aprroaches measure correctly but real world images look different.


This is because we all only measure 6 points. As my test shows (and Jim confirmed) the method used by JVC gets the primaries and secondaries right but has problems of varying degrees with the colours in between.


I am using the upgraded firmware in the projector

What would be interesting would be comparing the two approaches at more than 6 color points. If only it were easy.....
 

·
Premium Member
Joined
·
2,787 Posts
In addition to the Video EQ comparison question, does anyone know how the DVDO Duo CMS works, i.e. in RGB or HSL space? I know the interface on the Duo uses an HSL type system, but I don't know if it is based on an RGB system for its calculations.
 

·
Registered
Joined
·
144 Posts
Discussion Starter · #10 ·

Quote:
Originally Posted by Geof /forum/post/19604675


What would be interesting would be comparing the two approaches at more than 6 color points. If only it were easy.....

I agree that we need more standardised colours and targets on the CIE chart.


Creating the the colours is not difficult, in my case I just used Microsoft paint. The problem is knowing where those colours should sit on the CIE chart.


I did another test today using the desaturated primary colours from the Lumagen with interesting results. All of the points measured using the Lumagen were correctly located on the line between the white point and their respective colour.


In the JVC case, green was located correctly but more saturated, red had similar saturation but was shifted towards orange and blue was quite far off the line and a little more saturated
 

·
Registered
Joined
·
21,464 Posts
I have had a Lumagen for years and really like it. Lumagen is a great company, zillions of upgrade software and added features over the years and at no cost to the end user. The 3D upgrade is the first software upgrade they are charging for and even then it is included in the free software upgrades, its just this time, you need to buy a key to use the 3D features. AVS is a Lumagen dealer so if at all interested, and after visiting the Lumagen web site, give me a call at 240 876 2536.
 

·
Registered
Joined
·
5,668 Posts
Quote:
Originally Posted by Htnut2000
As my test shows (and Jim confirmed) the method used by JVC gets the primaries and secondaries right but has problems of varying degrees with the colours in between.
The problem is that you are assuming that the Radiance is correctly handling the different levels of saturation and that the RS20 is not. The truth is that neither is probably correct but that both are reasonably close.


In the RS20 calibration thread, this issue was discussed. Tom Huffman and others measured saturation and found most CMSs, including the Radiance, were lacking in their ability to track saturation.


See here .
 

·
Registered
Joined
·
4,940 Posts
Some weeks ago I loaded the new firmware into my RS20 and calibrated the CMS to compare it to my Radiance CMS. I also concluded that the Lumagen was superior and the JVC created wrong tints for some colors although both CMS measured closely with the primaries and secondaries. I turned the JVC CMS off and went back to using the Lumagen CMS only.
 

·
Registered
Joined
·
5,668 Posts
Quote:
Originally Posted by mhafner
Some weeks ago I loaded the new firmware into my RS20 and calibrated the CMS to compare it to my Radiance CMS. I also concluded that the Lumagen was superior and the JVC created wrong tints for some colors although both CMS measured closely with the primaries and secondaries. I turned the JVC CMS off and went back to using the Lumagen CMS only.
How did you conclude that the Lumagen was better?
 

·
Registered
Joined
·
4,940 Posts
Quote:
Originally Posted by Lawguy
How did you conclude that the Lumagen was better?
By watching the same material with both CMS on and off. If one CMS adds tints the source material does not have and the other CMS neither (instead of only correcting the gamut) I have my answer.
 

·
Registered
Joined
·
5,668 Posts
Quote:
Originally Posted by mhafner
By watching the same material with both CMS on and off. If one CMS adds tints the source material does not have and the other CMS neither (instead of only correcting the gamut) I have my answer.
I have never before heard of a Tint issue before. Tint is being added to the whole image? Please explain.
 

·
Registered
Joined
·
4,940 Posts

Quote:
Originally Posted by Lawguy /forum/post/19607547


I have never before heard of a Tint issue before. Tint is being added to the whole image? Please explain.

What I meant was that some colors got a tint, e.g. their hue changed when it was not supposed to.
 

·
Registered
Joined
·
144 Posts
Discussion Starter · #18 ·
I believe that there are two different aspects to what I see. The first is that undoubtably the image looks much better using the Lumagen CMS. I have tried to describe it but I probably do not have the right words. One comment from someone who is very familiar with my setup asked if I had changed the lamp without knowing that I had made the CMS change


In terms of the colors I am not entirely convinced.


I now find the JVC looking quite undersaturated with most material but also think that the Lumagen looks over saturated on some material.


What we really need is more measurements that will objectively explain the difference that can be clearly seen. Clearly measuring six points is not sufficient
 

·
Registered
Joined
·
5,668 Posts

Quote:
Originally Posted by Htnut2000 /forum/post/19612629


One comment from someone who is very familiar with my setup asked if I had changed the lamp without knowing that I had made the CMS change.

Differences in gamma can definitely have that effect. Lumagen's gamma controls are awesome. It is very easy to dial in any gamma that you want. JVCs controls are okay. You can get where you want to go but the journey is not the easiest.
 

·
Registered
Joined
·
1,347 Posts
The update to the RS-20 firmware should have taken care of the CMS and inter saturation points errors you are referring to. At least that is what I remember from the old RS-20 thread.


This is something I commented about with the Epson CMS error, and I developed my own calibration method as a results. (to compensate for the error, as I have to live with it since there was no firmware update from Epson). Here is my calibration thread, and it may describe something similar to what you are seeing, at least in principal.
http://www.avsforum.com/avs-vb/showt...7#post16166537


I asked about the CMS implimentation in the RS-20 thread after finding my results, and sure enough, the RS-20 had problems with it's CMS color saturation tracking as well. This is the reply from Tom Huffman to my inquiry that Lawguy provided the link to. JVC owners are fortunate that JVC spent the time to fix and improve their CMS. I'm not so lucky with my Epson, although I have reduced the visual impact of the error.


Perhaps you could perform the same measurements I did, and see where the lower saturation points end up. Be aware that the saturation points on the REC.709 AVS-HD disc assumes a gamma of 2.22 for the points to be spaced evenly.


Dan
 
1 - 20 of 341 Posts
Top