Keohi HDTV



Powered by Keohi Web Design


Living with the Panasonic TH-42PD50U EDTV Consumer Plasma & the Samsung PPM42M5S EDTV Commercial Plasma Units
An Unofficial Head to Head Bout.

<< Back to Michael TLV Tips


Head to head and side by side.   In June of 2005, I decided I needed a casual viewing set for my living room to compliment the DLP projection set up in my home theater room down the hall.  I wanted a set to wash dishes by and the eat dinner by; one where image quality was not paramount.  I had worked on many plasma sets before, but never a Samsung unit so it perked my interest.

At the 2005 CES in Vegas, Samsung announced / boasted a major leap forward in plasma TV contrast ratios.  For what it is worth, these numbers have been slowly inching up year to year and generation to generation.  800 to 1 … then 1200:1 … then 3000:1 … now 4000:1.  Now here came Samsung with a unit boasting a whopping 10,000:1 contrast ratio.  I figured that I just had to see what that was all about and given my requirements for this particular TV, taking a chance with Samsung was an acceptable risk.  I had already calibrated plenty of Panasonic units and NEC units and Sony units so I knew of their quality.  This one would be my little experiment. 

I ended up choosing this particular Samsung unit because the price was right and I figured that I did not have any use for speakers and a tuner and other features that the typical consumer units offer at a cost premium that is unavoidable.  I just wanted a panel with a minimal footprint in the corner of the room.  I wasn’t even going to feed it cable TV.  It would only get fed from a DVD source all the time.  I did not have any higher aspirations for this unit.  Don’t think big so the fall won’t hurt nearly as much.   Keep telling myself that.

The Samsung plasma sits about 11 feet from my viewing position and it is definitely a bright unit with plenty of light output.   I put a small D6500 bias light behind the TV to illuminate the back wall and reduce eye strain.  This set does not come with an HDMI input, but it does have a DVI input.  It does de-interlacing of a 480i signal fairly well, but not perfect.  It does not pass all the Silicon Optic DVD torture tests, but does pass at least 50% of them.  I ultimately decided that I liked the component input operation the best and stuck with that. 

An interesting feature on this panel is something called pixel shift.  This is an orbiting image feature that shifts the image around by a user defined “up to 4x4 pixels” while you watch.  It is slow, but it is noticeable when activated.  It lessens the effect of image retention and thus burn-in by moving the image around.  So instead of accidentally leaving a static image on the set for two hours (let’s say), it becomes effectively like leaving the image over an area of 11 pixels rather than just one.  Something closer to only 11 minutes worth of a static image as a result of the shift function.  That’s the theory anyway.  Four months later, it still works and I have left plenty of paused DVD’s on screen sometimes for one hour at a time. 

The set also has another feature that “wipes” the image and helps to reduce image retention quickly.  It is a light to dark sweeping pattern that moves across the screen. 

Now a little bit about contrast ratios perceived versus real.  I didn’t at any time actually believe that the set could achieve Samsung’s 10,000 to 1 contrast ratio claim, but to humor myself, I set out to replicate their “fake” reading conditions to generate this ratio.  The absolute dark reading is taken from an inactive input.  The bright reading was taken from a 100% white window box pattern occupying about 20% of the screen face at 100% contrast.  (Don’t try this at home, it’s a bad idea.)  Well to my own surprise, I obtained CR readings of 12,000 to 13,500 to 1 from my color analyzer.   Now what happens when I do the CR test with the ANSI method of 16 boxes … 4 by 4 of black and white boxes totaling 16 at the properly grayscale calibrated levels.  1,050 to 1.  This was surprising.   Much higher than I usually get from any plasma set after a proper calibration.  But it actually tracks with this little working empirical rule of thumb I use.  When the manufacturer boasts 1,000:1 contrast, the real post calibrated numbers are almost always in the 10% range.  They say 1,000:1 and I end up with 100:1 using the proper testing methods.

A RPTV DLP unit from Toshiba claims 1,500:1 and I measure 110:1.  A Sanyo HD plasma set claims 3000:1 and I get 270:1.  It’s always in that 6% to 15% range.  Let’s say 10% and we’d be pretty close to reality. 

Strangely enough during the ISF calibration process, it was determined that the set actually operated better at a higher contrast / light output level than is typical for plasma units.  This improved the dynamic range of the image giving it the brighter whites and the blacker blacks.  The actual ISF calibration process was somewhat brutal though as it discovered all the warts of this set in what it can and can’t do.  The gamma settings just are not very good on this set and changing the gamma settings in the service menu did very little to help this.  There was a lot of unwanted discoloration in the grayscale sweep.  This is a black and white pattern that is supposed to transition from black up through white in block increments.  Under desired conditions, this is a black and white pattern devoid of color.  Here we have a major failing.  Portions of this sweep have hints of color … a bit of green here … and there.  And what is worse, the dark end where the dark grays inhabit are actually a dark purple rather than the desired grays. (Teeth grinding)  This is the result of poor gamma in the set and I can’t do anything about it.  I have to compromise.  End result is … grayscale tracking that is average to mediocre. 

On the color decoding end, they got the red just about right on this set so no red push at all and the green color decoding is fairly close.  You can’t fix this anyway. But at least it is very close. 

I have joked that I need to set up a proximity detection switch for this plasma set.  It should be set to five feet and the moment I come closer, the set would shut off.   It sort of sums this up.  Don’t even bother trying to look at this set up close.  There is far too much posterization in the image structure.   Gradations are just poor and the dark stuff gets a purple tint to it.  But things start changing as you step back beyond five feet and the set turns itself back on.  (The proximity detection feature is a joke … it does not really exist.)   With all its flaws seen close up, they all vanish at the seating position except for the purple blacks.   The image looks very watchable and very nice.  

As an aside, I decide to confirm for myself the always on going debate between ED plasma sets and HD plasma sets.  The conventional wisdom says that at a certain distance back from the TV, the additional resolution advantage of the HD plasma set becomes diminished and both plasma sets end up looking the same.   For this test, I took my Toshiba 44NM84 DLP RPTV which uses a 1280 by 720 HD2+ chip and hooked up my HD signal generator to it displaying a multiburst pattern.  The pattern is effectively a number of vertical strips black/white/black/white/black and so one.  Each band of black is one pixel wide … so as long as I can see both the black and the white; I am seeing the full 1280 lines across the width of the screen.

The experiment uses that pattern up and I start to walk backwards from the image … one foot … two feet … and so forth.  My goal is to determine at what point I can no longer differentiate between the black and white lines.   At that point, the pattern simply looks gray to me rather than distinctly black and white.  This happens at about 7.5 feet back from the set.  At this distance, I can only see lines of two black pixel widths and two white pixel widths … which is the equivalent of seeing an image made up of 640 lines of horizontal resolution.  640 lines by 720 lines yields 462,000 effective pixels of detail in this HD image that has 922,000 pixels to start with.  Suddenly the difference between a HD plasma set and an ED plasma set in terms of detail is significantly less then where we started out. 

Conclusions drawn based on the results of this experiment.  The draw of higher definition resolution is significantly diminished when one intends to sit further back than eight feet to watch a 42” Plasma set or any other technology for that matter.   This is the case for anyone with 20/20 vision which is the vast majority of the population.  Obviously there will be a small number of people with eagle vision and claim they can see the additional resolution, but they are in a minority.

Does this test mean that the images will look the same after eight feet?  Of course not.  There will always be other factors beyond resolution that account for the image looking different.   The scaling algorithm used in HD sets versus those in ED sets are different since the image has to be scaled to fit differing panel sizes.  Do not mistake a “different look” as being better due to improved resolution.

Back on track again and so in September of 2005, I added the Panasonic TH-42PD50U to my home for some initial testing.   The set is stationed in the living room sitting right next to the Samsung plasma unit which has already been properly ISF calibrated to look as good as it’s ever going to get.  The two sets are about two feet apart. 

The Panasonic plasma unit is somewhat different this year compared to the previous years in that the grayscale adjustments typically found in the service menu of the TV were removed this year.  There is no direct access to grayscale adjustments on the consumer line of these plasma sets for the 2005 model year.  Is all lost as a result of this?  No.   Panasonic officials indicated that they felt the warm setting in the TV should be close enough out of the box.  Initial tests with the set freshly out of the box seemed to confirm this claim.  Yes the warm setting is indeed in the ball park of where D6500 is supposed to be.  It tracks within 300K of the correct setting.  Please bear in mind that grayscale calibration is very product specific.  Depending on the DVD player you use, there will be variance in the grayscale seen on screen.  It could be anything from negligible to a significant change.   It would not be wise to blindly assume that the TV’s grayscale will be accurate based on any DVD player you hook up to it.

Panasonic staff has indicated that the change was an oversight of some sort and that the grayscale controls would be restored for the 2006 model year.  The plasma sets were not the only ones to lose the ability to do proper grayscale.  The 2005 DLP RPTVs also lost that ability this year.  Well, all in not lost for the consumer plasma set though.  Through the hard work and digging of fellow ISF calibrator Gregg Loewen at Lion Audio/Video Consultants (Lion/Loewen … get it?), a solution to the grayscale issue was discovered.  Contrary to what the Panasonic people had said to others in the industry, the grayscale could indeed be adjusted in the current models.  It just took a bit more work to get there.  The secret was a special service remote control that was necessary to access the white balance adjustment part of the TV. 

With that in hand, I could fully calibrate the grayscale on the set.  It would seem that the TV can actually accommodate maybe six or more separate grayscale settings.  

Panasonic has decided to leave out the color decoder controls on this unit and that is unfortunate.  Fortunately, the default positions of both color and tint are close and the red decoder seemed to be set up correctly and was not over saturating red.  No red push.  The Green decoder could use a bit more tweaking on the tint side, but there is nothing that can be done on board. 

The set calibrated well with the Toshiba DVD Recorder feeding the 1080i signal via HDMI input.  The component 480p signal was being sent to the Samsung.  Test patterns indicated that resulting images were comparable with the HDMI having a slight detail edge.  This cannot be seen on real live material though.

A side note here, the Panasonic does not have any anti-burn in features built into the set unlike the Samsung with the scrolling pixels.  The Samsung also operates with much less heat generation than the Panasonic. 

Optimized contrast ratio on the Panasonic was about 580 to 1 which is lower than the Samsung, but consistent with my 10% guideline.  Overscan was set at about 2.5% on all sides where as the Samsung had a bit more on the top and bottom.  Resulting image made the people look skinny on the Samsung due to the slightly uneven geometry. 

The Panasonic does not hold black at black between two different patterns for setting brightness and this is unfortunate.   I don’t think this is a weakness in the power supply as it usually is on CRT based sets.  While in the service menu, the TV can be made to display perfectly solid blacks between the two patterns.  There must be some software issue in the TV that is making the uneven black level.  If we set the brightness based on the 50% white test pattern, the dark pluge pattern ends up too bright and the blacker than black bars show up.  If we set it up for optimal brightness with the darker pluge pattern, the 50% white pattern ends up too dark and crushes the black detail.  Compromising for this weakness, it is better to see black detail and have the image slightly on the bright side than lose detail and have things too dark.  

The Samsung does not have this black level problem, but the dark end goes to purple.

Optimized, the Panasonic has a far more linear grayscale with much better gamma than the Samsung.  There is no discoloration in the black and white materials.  They look black and white and there are no hints of green or purple creeping into the image.  The set has a very pleasing image with great looking blacks.   From the 10 foot seating position, the image looks effectively like a CRT display and that is about all we can ask of it for now.   

The pixel structure of the Panasonic is smaller than that of the Samsung.  This results in larger gaps in between the pixels meaning more of a screen door effect.   It is possible to get about one foot closer on the Samsung without seeing the pixel structure compared to the Panasonic. 

Up close, the Panasonic image tends to kick the Samsung image around quite a bit.  All the flaws of both sets are revealed close up and the Samsung has many more flaws than the Panasonic.  If we had to watch TV at four feet away, then the Panasonic would be very watchable albeit with the screen door effect very visible.  I don’t think I could watch the Samsung for very long.  It would be too punishing given that posterization effect.  My Subjective close up rating would be as follows:

Panasonic: 7.5 / 10

Samsung: 4.5 / 10

But a funny thing happens to an image when you start to back up from the image.  At the normal seating position of 10 to 11 feet, the two sets start to look more similar than different.   The grayscale looks surprisingly close except for any items with dark material.  The purple dark end starts to creep into the image.  The Samsung actually starts to look sharper than the Panasonic.  There is a bit more “snap” to the image and the light output is higher.  It is hard to figure out why this is happening, but the subjective normal seating distance rating would now be as follows:

Panasonic: 9.0 / 10

Samsung: 7.5 / 10

You have to juggle the purple dark end of the Samsung with the softer image from the Panasonic.  I was very surprised that the Samsung did as well as it did at the normal seating distance test.  You don’t want to turn the sharpness up on the Panasonic because that adds an undesirable ringing effect to the image. 

Well, I already made my choice with these sets since I got both of them.  I thought that I would be relegating the Samsung to the upstairs, but because it did not fare quite as poorly as I thought it would, it may stay put in the living room for now.  Sharper images from farther distances always have some appeal.  

 (I still have some sneaking suspicion that my particular Samsung unit might have some problems with its image processing, but I can’t put my finger on it and because of the way I purchased it, I have little ability to send it all the way back to the US on a hunch.  The Consumer units that I see in the stores just don’t seem to exhibit many of the weaknesses I see in this unit.  But bear in mind that the demo loops running in the stores are almost always with bright material making it hard to tell.)

That’s it for this time.

Michael TLV @ The Laser Video Experience
Lion Audio/Video Consultants Network

Comments? Post it here.