November 8, 2007 - Last Thursday’s Display Daily on the presentation of LG’s new G Platform line of PDPs at FPD International generated some interesting responses, and most of them concerned the credibility of the claimed luminance and color bit-depth of these displays. LG was specifying luminances of 1000 and 1500 nits, depending on model, and a 16-bit color depth.

Ken Werner
Senior Analyst and Editor

Let’s say immediately that these luminance figures, at least, are not far from what LG’s competitors are claiming, so the feedback I received is really about PDP specs in general, not just LG’s. As far as bit-depth is concerned Samsung SDI claims 13 bits and Panasonic claims 16-bit processing.

A notable e-mail on all this came from a senior video quality engineer at a Silicon Valley start-up:

I was wondering if you would care to comment on a particular issue of interest to us and many consumers I believe. Namely, the luminance specifications used by plasma manufacturers. It is well known that due to overall power consumption limitations plasma monitors dim the picture for a full white flat field and raise the luminance output as the white area gets smaller and smaller. Case in point, we purchased a Samsung plasma professional monitor model # PPM50M6H rated at a peak output of 1,300. Using standard instrumentation in the form of a calibrated PhotoResearch Spectrascan PR-650 and the DisplayMate battery of test images, we measured 52 nits for a full white field in the out-of-box dynamic picture mode and 240 nits for a white square covering only 0.20% of the total picture area with the rest of the picture being driven black. We’re very well aware of the lack of standardization in the industry regarding manufacturer’s specifications of luminance and sequential contrast, but this is bordering on the verge of misrepresentation. Why is it that the industry has not agreed to a regulatory oversight by a reputable and well-respected agency, such as VESA or NIST for example, so that everyone measures the same agreed-upon quantities in the same repeatable manner under controlled laboratory conditions?

Before answering the video quality engineer’s question, I thought it would be interesting to see if his measurements were representative of those obtained by other sophisticated observers, and what those observers think of the situation. I went to Ray Soneira, President of DisplayMate Technologies and a globally recognized expert in display evaluation, and consultant Pete Putman, President of ROAM Consulting, who frequently evaluates flat-panel TV sets. (Incidentally, both Ray and Pete are Analysts at Insight Media.)

Ray responds:

There are two distortions in claimed plasma peak luminance, which I have mentioned in my "Display Technology Shoot-Out" article series, now on http://www.displaymate.com/shootout.html.

1. For plasma displays the manufacturer’s spec is often for the luminance without the anti-reflection absorbing layer installed, essentially the peak luminance of the naked glass.

2. For plasma displays peak luminance values specked by manufacturers are frequently for a 1 percent window. Heat dissipation significantly cuts that for higher APL [average picture level]. The data and text below are for a NEC plasma panel from Part I of my "Display Technology Shoot-Out" article:

212 cd/m2 5% APL
133 cd/m2 25% APL
81 cd/m2 50% APL
53 cd/m2 100% APL

For most computer applications the APL is rather high (because word processors and spread sheets, for example, use a peak white background) but for most video applications it is relatively low (because the images are generally dimmer and are colored, not gray or white). As a result, plasma displays are generally used for video and are unsuitable for most computer applications.

The display industry is experiencing a period of significantly increased spec exaggeration, particularly for Peak Luminance, Contrast Ratio, View Angles (for LCD), Response Times (for LCD), Color Gamut, Number of Screen Colors, intensity bits per color, etc. This is due to increased competition and is true for most industries (not just the display industry). Personally. I think a lot of it should be characterized as fraud rather than specsmanship.

Incidentally, I don’t believe the 16-bits per color spec. I am pretty sure that involves spatial and temporal dithering over and above pulse width modulation.

Plasma is digital like DLP, and requires pulse width modulation to generate an intensity scale. To generate 16-bit color at 60 Hz requires a switching frequency of 3.9 MHz for each primary color. Last time I checked (2 years ago) plasma switched well under 100 KHz. I’m sure it has improved, but I’ll bet it is still under 200 KHz, so the 16-bits per color requires all sorts of spatial and temporal averaging (dithering). That will generate visible artifacts, and not just in DisplayMate test patterns.

And Pete agrees:

This isn’t any surprise to people who work with displays on a regular basis. The current limiting circuits in plasma kick in once power consumption reaches certain manufacturer-specified levels. They are concerned about overdriving the plasma and prematurely aging it, also there are power consumption issues to deal with.

I ignore all of the manufacturer’s claims for brightness and contrast and just do my own measurements. Case in point: I just finished a set of measurements on a new plasma, where I calibrated for best grayscale image and gamma.

Brightness measured 50 nits with a full white window. That increased to 163 nits with a small white window. With a 50/50 black/white pattern, brightness measured 83 nits.

I would be astounded to see an average (ANSI?) full white measurement of 400 nits from any plasma, It would probably be engulfed in flames at that point.

And the answer to the video quality engineer’s question concerning why manufacturers don’t measure according to realistic standards? They have no motivation to do so. We saw with the old InfoComm Projection Shoot-Out and the SID Display Technology Showcase, that after a certain point manufacturers see no benefit in being measured on a level playing field.

Would any PDP manufacturer want to put a luminance of 100 nits on the box for a $2000 TV set? Not likely.

And there were some independent efforts to establish independent specking of projectors after the Projection Shootout collapsed. Not only was there no industry interest, but building the infrastructure required to obtain hardware for testing, make measurements, and get results back to the manufacturers in time for them to use them in advertising and packaging - all done with complete secrecy - would have been both daunting and expensive.

What’s the solution? For personal use, believe your eyeballs. If you need reliable measurements, you now know of two places where you can get them. If you want to make your own on a regular basis without buying very expensive equipment, there are economical solutions, such as DataColor’s Spyder2 colorimeter.

Reply to the author