AVS Forum banner
Not open for further replies.
1 - 20 of 301 Posts

2,263 Posts
Discussion Starter · #1 ·
Thanks to the generosity of Ken Land (who let me use his brand new colorfacts while away on vacation) and my wife (who didn't kill me for spending over 30 hours with my TV, mostly late at night) I've managed to do a analysis and calibration of the HLM507.

My initial goal in this exercise was to achieve as close to 6500K color temperature from the HLM for as many inputs as possible from 0 - 100 IRE. After watching a large variety of material on different inputs, I realized my HLM was not producing a solid black. On all of the inputs black had some mosiquito noise in it, which even though I couldn't see from my viewing distance I wanted to remove. The summary of my results is that I have been able to get a good color reproduction with no noise in blacks. A side benefit seems to be that the "sunburn effect" has gone away, but I reserve the right to change my mind on that until I can do some more viewing.

My equipment list is:

HLM507W - release 108, service date Sept 4, 2002

Pentium III 850 MHz with ATI Radeon All-In-Wonder 7200 as DVI and PC input source

Panasonic DVD-RP62K as Component 1 480P, 480I and S-Video1 input source

All testing was done at night time to minimize ambient light impacting the results. I initialy place the colorfacts monitor in the center of the screen, but later moved it to the upper left hand corner to minimize the impact of the hotspot in the center of the screen.

I decided to start working on the DVI input since it is the shortest path to the DMD chip, and the fully digital path should eliminate the most rounding errors of any inputs. I started by returning the service menu back to the factory defaults. I decided to analyze the out-of-box performance of the set to see what it looks like to the average user. I used colorfacts to analyze the color output in dynamic, movie, and standard mode with cool1, cool2, normal, warm1 and warm2 color tone. I measured the contrast ratio (Full On/OFF and ANSI checkerboard), temperature and RGB historgrams for each combination. The following table lists the results of the contrast and color temperature reported by color facts. The color temp is the average of the color temperatures from 0 - 100 IRE taken in 10 IRE steps. An asterisk next to the color temp indicates that the temperature clipped to either 0K or 13000K. The colorfacts appears to have accuracy issues in low light situations. Measurements below 10 IRE I don't trust and between 10 IRE and 20 IRE are suspect.

\t\tOn/Off\tANSI\tColor Temp\t


















Warm2 is the closest to 6500K so I decide to use it for further testing. I will attach pictures in follow-up posts inlcuding the CIE chart, RGB historgram , Temperature histogram and Luminace histogram for Dynamic, Warm2. As you can see blue dominates as is the case for all of the color tones. Also notice that green dominates in the low IREs, which has been noted by others (such as Tom Roper). It has been assumed that it was a source issue, but I found that all inputs have a dominant green in the low IREs which was something else I worked on eliminating.

Next, I used a DVD to tune the Custom mode for best results and then took the same measurements for each of the different Gamma settings in the service menu. From these, I decided that Gamma0 would yield the best results for my exercise. See the RGB levels photos later. Notice that Gamma3 seems to have the flatest response, however it's Luminance histogram came out to almost 1.0, which gives a very poor resulting image. Gamma0 gave a Luminance of around 2.5 and the highest red levels. Since red is the weakest color, I decide to use the gamma setting which emphasized red the most. Gamma0 also returnd relatively flat RGB levels across IRE levels.

After all of these measurements, I decided to proceed with Gamma0, Custom, Warm2. After many hours of tweaking in the service menu, I managed to get results I liked. However, twice in the last week, I've had to re-tweak the settings. I suspect it's because the bulb was still "breaking in" The first change happened between 72 and 82 hours, the second around 120 hours. The final results I will present are for a bulb life of 135 hours. I hope to rent the colorfacts from Ken in 6 months or so to see if my settings still hold.

The colorfacts is a great tool and if Samsung had been more conscious of the end user I could have dialed in the results much faster. Unfortunately, the service menu always comes up in Dynamic Cool1 so I would often have to make a change in the service menu, then go check the results (after remembering to change back to Custom Warm2) and then repeat. If the service menu would come up in the last mode then I could have used the colorfacts continuous measurement mode to get immediate feedback on my adjustments.

So, now for the results. Here's a table of what I was able to achieve for each of the inputs:

Gamma0 Custom Warm2\t\t\t\t

\tOn/Off\tANSI\tColor Temp\t






As you can see, DVI has the best contrast results. I believe this is due to the shorter signal path. As the path gets longer I imagine the noise goes up and the contrast goes down. I was not able to measure the checkerboard contrast via DVD due to time constraints, but I suspect the results would be around 140:1. Note, the PC results were not completely tweaked out because of time constraints.

I will include pictures of the RGB historgram , Temperature histogram and Luminace histogram for each of these inputs in follow up posts. I'm attaching my service menu and user menu settings that got these results here.



hlm507_service_settings_tuned.zip 4.5283203125k . file


1 - 20 of 301 Posts
Not open for further replies.