As many of you know, my Home Theater Geeks podcast is produced by TWiT, an Internet television network headed by tech journalist Leo Laporte. Among Leo's many admirable qualities is his refusal to review anything that he does not purchase as any consumer would, because this assures his audience that he is not being influenced by the manufacturer.
Leo Laporte's new Samsung KN55S9C OLED TV
Recently, Leo really put his money where his mouth is—he bought a Samsung KN55S9C OLED TV for $9000! I am very gratified that he asked me to calibrate it, which I did in collaboration with Robert Heron (dangled on AVS), noted technologist and co-host of the HD Nation podcast. In fact, we are two of very few AV journalists who have had the opportunity to play with this set outside of Samsung's facilities. A TWiT camera crew was on hand to record the proceedings, which is now a special episode of Home Theater Geeks.
I invite you to watch that episode to observe the process for yourself. However, while the camera was trained on the computer screen at some points, you can't really see the results of our calibration in full detail. So I'm posting them here, along with a summary of what we did and learned about this new TV technology.
Our primary calibration system included a Konica Minolta CS-200 chroma meter and LS-100 light meter, AV Foundry VideoForge test-pattern generator, and Spectracal CalMan 4 software running on an HP Pavilion dv8000 laptop. (I know, I know—I should be running CalMan 5 by now, but the laptop doesn't have enough memory, which I didn't realize until it was too late to upgrade for this project.)
As I tested the system in preparation for the calibration, the computer didn't recognize the CS-200. After trying everything I could think of—and everything Spectracal could think of—with no success, I called fellow calibrator Ray Coronado (RayJr on AVS), who came over and discovered that the USB connector was dirty! He cleaned it, and everything worked perfectly. Thanks Ray, you're a lifesaver!
Robert Heron also brought his calibration gear—an X-Rite i1Pro spectrophotometer, Sony Vaio VGN-NW160J laptop running CalMan 5, and a Raspberry Pi with customized OS so it can be used as a test-pattern generator with CalMan. Since OLED is a new TV technology, we were interested to see if the two color meters measured significantly different results.
We decided not to calibrate the 3D mode, which takes more time than we had. Besides, Leo doesn't much like 3D anyway, so he won't miss it.
After we reset the basic picture controls to their factory defaults, we selected the Movie picture mode, set the picture size to Screen Fit (no overscanning), and adjusted the picture controls using the test patterns from the VideoForge. The basic picture controls were pretty close to correct at their default settings, except Sharpness, which was way too high; there were obvious white outlines around the black lines on a gray background, a condition called ringing. We turned the Sharpness down to 5 (the default was 20), which eliminated the ringing.
Here is a snapshot of the pre-calibration measurements of the KN55S9C using the CS-200.
In the Movie picture mode (which uses the Warm2 color tone preset), the grayscale was not bad, as you can see in the snapshot above. The error in the grayscale rose with brightness—delta E was less than 3 at 50% and below, rising to just over 5 at 100%. A delta E of 3 or less is considered indistinguishable from correct to the human eye, so we did have a bit of work to do, at least in the upper half of the brightness range. All brightness levels were deficient in blue, while the bottom end had an excess of red, and the mid and high ranges had an excess of green.
As expected, the black level measured 0.000 on the LS-100; we could not see the screen at all when displaying a full black field. This is one of OLED's real strengths; I haven't seen blacks this good since the old CRT days.
The peak white level was over 60 footlamberts—yikes! I prefer 30 fL in a dark room, which is the environment we were working in, but other calibrators prefer it a bit brighter; for example, David Katzmaier of cnet calibrates to a peak white level of 40 fL. We settled on around 34 fL by reducing the Cell Light control, which is much like the backlight control on an LED-LCD TV. If Leo wants to watch something during the day or with lights on, he can increase this control without affecting the overall calibration.
With the color space set to Auto, the colors were very close to correct; all had average errors less than 2. Red had the largest errors—it was not bright enough and too saturated. But even so, the error was less than could be distinguished by the human eye, so we decided to leave the colors alone; getting red closer to correct was not worth the trouble and time it would have taken, especially since using the color-management system means starting in the set's native color space, which is way off the mark, as seen below.
The KN55S9C's native colors are way off from Rec.709.
Obviously, OLED can reproduce a much wider color gamut than is called for in the current video system, so it might be very appropriate for 4K/UHD, which could ultimately use a wider gamut. But for now, we returned the Samsung's color space control to Auto, which brought the colors back to where they should be for HD.
The post-calibration measurements were a thing of beauty.
After calibrating the grayscale using the 2-point controls, the errors at all brightness levels were well under 1. The lowest level was slightly blue, but otherwise red, green, and blue were very evenly matched. The KN55S9C provides a 10-point calibration option, but we decided not to use it because our results were well below the threshold of perceptibility.
The average gamma was about 2.1 both pre- and post-calibration, which is just a bit low. But when we nudged the gamma control by one click, it jumped to 2.4. I prefer a gamma of 2.2, so we reset the control to produce a gamma of 2.1.
The i1Pro grayscale measurements were similar to the CS-200, though not exactly the same.
After calibrating the set with the CS-200, Robert's i1Pro achieved similar grayscale results; for example, both instruments measured a similarly shaped drop in gamma from 70 to 90 percent, though its reported average gamma was slightly higher (2.18 versus 2.11). The i1Pro measured slightly elevated green throughout most of the brightness range and slightly higher errors, barely exceeding a value of 3 from 80 to 100%.
With CalMan 5, Robert was able to perform a saturation sweep.
Robert also performed a color-saturation sweep, which CalMan 4 can't do. As you can see above, the saturation of most colors was a bit low at brightness levels below maximum. Compared with the CS-200, the i1Pro measured slightly larger errors in the colors, especially blue, and you can see the slight excess of green it detected in the RGB Balance graph. Finally, the i1Pro measured a slightly lower peak white level of 30.885 fL. But all in all, these differences were minor, so we didn't worry about them.
After all that, it was time to watch some movie clips. The black bars on letterboxed images simply disappeared, even during very dark scenes—it was a beautiful thing. Robert and I both thought the shadow detail in clips such as the opening scene in Master and Commander was not as good as it could be, but we didn't want to raise the gamma any more, so we raised the brightness control by three clicks, and the shadow detail improved quite a bit.
Colors were gorgeous—rich and vivid without being garish. Skin tones in Samsara and Life of Pi looked completely natural, as did the colors of animals, sky, water, earth, and foliage.
We also looked at some of the video-processing tests on the Spears & Munsil HD Benchmark test disc, setting Leo's PS3 to output 1080i. The TV did a great job deinterlacing both video and film-based content, with nary a hint of artifacts.
Like LCD TVs, the KN55S9C offers frame interpolation to reduce motion blur; Samsung calls this feature Auto Motion Plus. Of course, frame interpolation also causes the dreaded "soap-opera effect," which makes film-based content look like it was shot on video. Fortunately, the Custom setting provides two controls: Blur Reduction and Judder Reduction, which control how strongly frame interpolation is applied to video- and film-based content, respectively. To avoid the soap-opera effect on film-based material while sharpening motion blur on video-based content, we left Blur Reduction at its default value of 5 and turned Judder Reduction down to 0.
Also in the Custom mode of Auto Motion Plus is a control called Clear Motion, which inserts black frames between the actual image frames in order to reduce motion blur. We didn't find this to be dramatically effective—except to reduce the light output of the screen—so we left it off.
In addition to his calibration gear, Robert brought along a Leo Bodnar LagTester to measure the input lag for gaming. With game mode off, the input lag measured 180.1 ms—way too much for good gaming. With game mode on, it was still a poor 81.6 ms. Clearly, the KN55S9C is not a gaming display.
My only complaint is the curved screen. Unfortunately, Leo has the set on a piece of furniture that places its center above eye level when seated on the couch, and the TV's stand tilts it back slightly. This made the screen's curve obvious to me, even sitting dead center. The image was not rectangular—the top and bottom bowed inward, which looked weird.
Another issue I have with the curved screen is that it encourages a solitary viewing experience. Ideally, you should sit directly centered on the screen at a fairly close distance. Anyone else is in a compromised location. Granted, the off-axis performance of OLED is not degraded as it is with LCD, but the geometry of the image is strange. And if there is something reflected in the screen, like a light, that reflection is smeared out, appearing larger than it would on a flat screen. You can see this effect in the photo at the top of this article—the video light is reflected near the top of the screen. Finally, you can't easily wall-mount it. I really hope that Samsung and other manufacturers make a truly flat OLED sooner than later.
Other than the curved screen, the Samsung KN55S9C is a spectacular display. Yes, 55 inches is relatively small by today's standards, and $9000 is way beyond most folks' budget. But if you've got the dough and you want the best picture quality currently available—and you don't mind the curved screen—the KN55S9C is truly an exceptional TV.
My deepest appreciation to the folks at Spectracal and to Ray Coronado (RayJr) for his help with troubleshooting the CS-200-to-computer issue and with creating the CalMan 4 graphics used in this article. He is a CalMan wizard.
Are you a professional ISF expert or hobbiest? Which type(s) of tv sets do you work ?
A big DITTO for me also. I am thinking, when an OLED hits 65+ inches, that is when I start looking.
8.3 Surround, protected by APC H15
Speakers: SpeakerCraft: 6 AIM8 Fives & 2 TIME Fives used as Wides
Woofers: SVS PB13-Ultra (1), HSU MBM-12 MK2 (2)
Samsung UN78JS9500, mLG OLED55E6P
Samsung UBD-K8500 UHD player, Denon 4311CI, Harmony 1100, Harmony Elite, 9 Noctua cooling fans
Louder is NOT better!
I am THX certified, and I have attended parts of the ISF calibration course, though I'm not officially certified by ISF. I work on all types of TVs and projectors.
I really hope this tech works out in the field as it sure has promise. However, it will be interesting to see how this set holds up after a year. It would be great Scott if you would do a follow=up in a year to see if there is any noticeable premature aging of the blue phosphor as well as any evidence of burn in especially with news and sports tickers and letterrboxed bars.
Yeah, I hope to do exactly that! I'm also very curious to see if the blue-longevity problem has been solved. (BTW, OLED does not use phosphors like CRT and plasma; rather, the OLED material emits light directly upon electrical stimulation.)
Louder is NOT better!
I would say the OLED's picture quality is the best I've ever seen, even Kuro. But the VT60 and ZT60 are no slouches by any means. In a side-by-side comparison, I'm sure the OLED would have deeper blacks, but by themselves, their picture quality is superb.
first, very nice and interesting writeup!
Also, Imagic did a good job of photo documenting this issue with the curved reflective screens in his earlier piece on one of the recent OLED displays.
Great write up scott, how far were you sitting from the screen? did you notice any motion blurr?. Must of had fun playing with that samsung :D
Hsu Research HB-1 MK2
Hsu Research Hc-1 MK2
Hsu Research VTF3 mk4 12" subwoofer
is there any CMS adjustments on tis tv?
it seems to need them to adjust the low luminance values for Red and Blue.
im surprised that this Oled didnt have a higher peak White than 205cd/m2?
im using a peak white around 280cd/m2 on my Sony 55" W9 even in a low lit room to get that "feels like you are there" feeling from the contrast.
high peak white really adds the 3D feeling,most in darker scenes when the difference between white and black get so high.
but it all depends on how good the tv is to show details in between.
also great to see what Quantum dots can do to the gamut compared to OLED.
they really can match it
the colors on your OLED in the wide gamut mode is on pair with the Sony W9 with Quantum Dots.
only the Green color is slight cleaner on the OLED
Edit: one tip is to use the Native wider gamut colors for watching DSLR photos with ADOBE RGB profiles.
(use Adobe RGB in the Camera to Camera raw or what ever you use)
do that and you will see a big difference compared to SRGB.
photos comes to life in another way
also you should get some of the SONYs Mastered in 4K discs with xvColor data and try them with the Native colors on your OLED.
with those discs you get correct colors for the wide gamut.
I do kind of miss my plasma's blacks, sometimes, but then I look at my massive 100 inch image and say, nah, I'll pass. But I am definitely curious about OLED, I'm dying to see one of these in person. Personally, I think curved screen makes sense only for scope projectors, or perhaps massive sized theaters. Flat is just better, for 99% of uses, especially at that size.
You, Leo, and Robert are the best out there to help consumers learn and know about consumer electronics.
From all of us here at AVS please for us Scott give us a shout out to Leo and Robert.
Yes I can remember a long time ago when Leo was on Tech TV thems are the good old days.
I am sure there are other professionals in your field and around the world but you 3 at top notch in our book that is in the public's eye.
A far cry from the days where you had to wait 5 minutes after you pull the knob on the old black and white TV where you see a white dot appear and hoped that you got a good picture, or ? hoped that one of the vacuum tubes didn't go on the frits while watching or else dad would have to go to the local Reads drug store and test the tube on those fancy tube store testers.
JBL Pro Cinema/Crown power/Sony SXRD
Moreover, with Game Mode the input lag is dramatically better, virtually unnoticable for me. And the fact that the screen of the Pioneer is flat, thus making it more suitible for off angle viewing.
What is the word on reliability? Any indication how this display will hold up over a few years of use?
Certainly I am excited about the black levels of OLED, but the KRP's black level of 0.0005 ftL is impressive in its own rite.
Would the quality difference be night and day in your opinion?
Thanks for the write up.
That's great to hear it's beaten a Kuro (MLL) - it's only taken five years and a completely new/different technology to achieve that goal though
Being a keen gamer the big time lag is a concern... But me personally would advise anyone worried about the OLED not being suitable in that regard to not worry as it will get better and better with each generation that passes.
Hopefully by the time in a few years when i pick up a U-HD OLED to replace the trusty 500M as the main display the lag will be at least at level pegging with the Pioneer in that regard.
Nice review it's just great that it's finally all happening this new technology and to know that there is going to be a genuine successor to the KRPM in a flat panel sense (when these come too) and something worthwhile to save for.
Hey, Scott. I have my S9 on order and this is good news! Nice work! I guess this proves that present calibration equipment is perfectly fine for OLED TVs.
I have to ask you, since OLED does not use phosphors, there should be no need for a break-in period before calibration. Is this correct?
Also, do you recall the setting that you used with Cell Light control to get a peak white level of 34 fL?
|Samsung Kn55s9c Curved Panel Smart 3d Oled Hdtv|