Comments on: Sony's LED-based LCD claims 'best' crown

The Sony KDL-55XBR8 is officially the best LCD TV CNET has ever tested. It can't overcome the picture quality of the best plasma, however, and it still costs a mint.

[puremind]

Why did the reviewer calibrate to a gamma of 2.07 and 40Fl?

With such low peak luminance I find myself straining my eyes, like when I am reading a book in a poorly lit room.

I would have preferred a gamma of with 2.2 with less bright details (the reviewer himself aknowleges this issue in his review, mentioning shadow details being "a tad too bright" compared to the Kuro.

I think the calibration issues are due to the fluctuating gamma of LED-LCD's. You can afford a higher gamma on LED Models, because test patterns overestimate black crush on this type of LCD.

Also, I found a German test that find the Samsung winning the competition between the two LCD's. Could it be that the US market receives different Samsung products? I believe they have a different processor, software and are produced in different places (Mexico for the US, Slovakia for Europe).

I find it a shame that the US apparently does not have access to the best of what Samsung can offer.

[puremind]

Also it occured to me that perhaps the Sony would have fared better against the Kuro with proper gamma clibration, especially regarding black levels, making it the overall winner.

And finally it occured to be that in the Geek Box, both the Samsung was the only TV of the 3 that received good everywhere, when both the Sony and the Kuro had failed areas. This made me doubt the final ranking between those 3 TV's.

It seems that the ranking has more to do with the unrealistic calibration targets (which favour plasmas, as LCD's are optimized for peak luminance higher than 40Fl).

I find it a bit restrictive to calibrate to a limited luminance level, as the forces a non-perfect gamma on most LCDs and artificially reduced ANSI contrast. I think each TV should be allowed to show the full spectrum of its capabilities. The only targets that should be set in my opionon are accurate gamma, color temperature and color hue/saturations. What the black and white levels end up being then shows the dynamic range and hence the quality of the TV.

Many people inluding myself complain that they can't use the CNET settings for real life viewing, as those are aiming more at placing all TV's on an equal footing. IWhile this may be desirable for better comparison, I think the approach and assumptions used greatly affect the outcome.

I certainly appreciate the CNET tests for the wealth of information in them. But personally I would prefer a format similar to the famous UK testers (Vincent and David, I am sure people will know which web site I mean - I am not sure I am allowed to name web sites here)

[katzmaier]

Hello puremind,

I appreciate the feedback. I definitely understand that, especially for LCDs, the ceiling of 40FTL results in CR that's way beneath their real-world contrast capabilities, and in some cases can cause less-than-ideal gamma results. I am also open to persuasive arguments for changing my test procedure, if it's warranted. But in this case I'm not so sure it is.

As you've noted, the CNET calibration targets are designed to facilitate a same-footing side-by-side comparison. To change that and calibrate for ideal gamma for every TV would in my opinion make those comparisons much less valid. The target I set is, as I say, designed for a completely dark room to minimize eyestrain -- a brighter picture may threaten to cause that sort of strain, and cause other problems associated with higher contrast (afterimages & glare, excessive blooming on LED displays, etc). It may be worth noting here that standard brightness in a movie theater is 12FTL, and I doubt most moviegoers equate that picture to reading a book in a poorly lit room.

That said, I have no doubt that if Sony's engineers had, they could achieve a gamma closer to the ideal of 2.2. In fact, my post-calibration gamma you cited is closer to that ideal than any of the gamma presets in the TV's Cinema mode (which, for the record, I measured at High=1.675, Med=1.769, Low=1.88, Off=1.991). The TV's max light output in default Cinema mode is also 43 FTL, in case it matters, which says to me that Sony's engineers agree that my target is at least in the ballpark (or maybe it just means they sent me a fixed-for-CNET sample...who knows?).

One final point is the basic premise that any TV can get really bright; equalizing the max light output in our calibrations takes that factor out of the equation and focuses more closely on a TV's ability to reproduce great blacks. IMO the dynamic range does not dictate the "quality of the TV" to the nearly same extent that black level reproduction does.

Addressing your other points: As the review mentions, blooming and especially off-angle issues were the principal reasons the Kuro beat the Sony. And the Geek Box is a contributor to the final score, but it's by no means the arbiter, especially compared to black levels and other factors not covered by the Box.

David

PS: As for other web sites, I do know what you mean, and anyone can post (appropriate) URLs. http://www.hdtvtest.co.uk/ does an excellent job.

[puremind]

Hi David. Thank you for the clarification.

I agree that black levels is an important component in a dim environment and 40FLT is also the recommended white point under low light conditions.

But what about daytime viewing? Surely a white point of 50FLT or more would be desirable. Calibrating all TV's with a perfect gamma to draw comparisons (e.g. regarding the achieved white point) is just as interesting. Using both approaches would put LCD's and Plasmas on an equal footing and allow them both to excel at what they do best. This would also provide plasma users with better settings for daytime viewing, as unlike with LCDs where backlighting is usually enough to adapt for lighting conditions, plasma users have a harder time.

Finally, as a reader, I have a couple of suggestions:
1- Include white balance charts (the geek box color temp indicator doesn't indicate at which luminance levels irregularities happen)
2- Include gamma charts so that people have a better representation of what the calibrated settings were aiming at
3- Systematically include black level measurement for the dim calibration and white level for the bright room calibration.
4- Provide On/Off and ANSI contrast for both calibrations
5- Show CIE charts. Since you have some very good equipment, it can help people with lesser equipment do a better color space calibration.
6- Include input lag measurements for both game mode and the calibrated setting.

You have such great equipment and there is so much worth sharing with your readers! Most importantly, having this information would allow them to pick the right TV based on their individual viewing preference rather than the reviewer's. Some people don't care about viewing angle, place an emphasis daytime viewing which cancels out blooming, or want to use their TV as a PC monitor, or couldn't care less about standard definition or scaling as they use an external processor.

Your tests are informative and it nice that you do state your preference (something sometimes lacking on hdtvtest.co.uk, where no clear ranking exists) but one reason people like hdtvtest.co.uk is that they give you all the information you need to make your own decision, in a very transparent way.

I am sure CNET as an editor has additional constraints but hopefully some of the above can be implemented. Many people would be delighted. I sure would.

[katzmaier]

Hi puremind,

Sorry for the late reply, but I'd just like to say I appreciate your suggestions. I plan to implement more updates to CNET's HDTV testing procedure in 2009 and am exploring the possibilities of including charts of the data that I currently collect, as well as formal post-calibration ANSI and full-raster contrast ratio tests. I won't make any promises, and nothing's going to change this year, but I always appreciate input like yours and will do my best to improve our TV reviews.

David

[puremind]

Thank you David, this is great feedback. I have now order a KRP500A (European Elite Model) due to my own testing. The black level and hence gamma fluctuation on the LED LCD's is a bit disturbing and my Samsung set had a frame drop issue in 24p, which has been reported by a number of users.

[allenstewart]

I've had the 46XBR8 for a few days.

I concur with your review but I have noticed that in comparison with my 40XBR1, the screen (while not as glossy as the Pioneer or Samsung) is a bit more shiny.

To me, it makes it seems more like I'm looking at the picture (as great as it looks) through a window. My wife thinks I'm crazy- does anyone else understand what I'm talking about?

Thanks.