Panasonic 2010 plasma: Floating blacks - Page 2

The black levels fluxuate on mid/low APL scenes. I have not seen floating blacks during high APL scenes.

Thanks, but is there a predictable direction (up or down) of the black level shift with APL or does it seem random?

Cheers

Random, but repeatable

Not my best post on the subject, but it does get the idea across. This was eliminated in the 2009 1080p Panasonic models that I had, so it shouldn't be there on the 2010 models unless they added it back to help cover up the rising black levels.

The best way to tell if it is happening is to watch a black bar movie and see if the black bars are dimming and brightening as well as the content. With the old Panasonic models, the black bars would fluctuate noticably on most black bar movies, the 2009 models didn't have any fluctuation unless the black levels rose while you were watching it.

I notice this (if i'm understanding the phenomenon correctly) sometimes on my Pio 4270. Anyone know if this is the same thing? if so, its not that noticeable/troublesome to me usually.

Thanks, that seems to fit with the 20090021452 patent applicaton. It states that the black level will shift at 6% APL (gets darker below 6% and brighter above 6%). Also says that if a hysteresis characteristic is applied there will be two thresholds at 5% and 7%.

These particular APL thresholds may create a very random effect as average brightness will be difficult to differentiate in many of the scenes (ie - credits, mostly dark scenes with bright spots...etc)

Are you using Pure Cinema Advance?

They are not out yet.

So what is the patent's claimed advantage? Did I miss it in an earlier post? Surely they did for some other supposedly advantage. Not just to bug us.

The VT20's are out. At least at some BestBuy stores.

Very sneaky patent IMO My interpretation is this:

The panels 'normal' or 'natural' black level is the higher black level seen above 6% APL. This black level allows for stable high speed operation. When there is below 6% APL the panel switches to a lower black level to maximize contrast to the viewer at the expense of driving stability.

So, stated simply. IMO the advantage is to increase contrast when a lot of black is on the screen.

Note: According to this patent it sure does seem that Panasonic Plasmas 'natural' black level is either hidden in high APL scenes (floating blacks) or deferred (rising black levels).

Yep

Or clever. When you really need the black levels lower, you get it. As long as they don't process the snot out of the picture like LCDs, I guess a good thing. Or acceptable.

What I have stated to notice is not so much the grey bars on films or floating blacks,but that apparently no analog/scanned films really have that deep of black levels native to the source. Certainly so far most are not even close to the black bars for any source. And one of the worst was 2001 BR. God awful and on a Kuro too.

I vaguely remember from my 35mm slides film days noting that I never got true blacks on the screen. If I flashed up a totally unexposed slide, it glowed. And they use to rate films DMax. So even film, which is still way over 90% of most BR sources today can not really do a perfect black.

Are we chasing some holy grail by just looking a grey bars? It is a good test of the set, of course, and you do really want a perfect black, if 0 IRE. But if most of our sources today don't have it to show off our new digital set.

Of course, a totally digital source, like CGI or animation, could have perfect blacks. But if that source, I have found grey star fields and night skies even on an Kuro 101. Just seems they know people would not expect it. Hey, Panasonic CSR, my picture looks too black! What is wrong? Oh, it is featuere- for once. LOL.

Ooops, sorry. No it was meant as a test...

I haven't reviewed the patent, but based on your description, I can make a simple guess as to the motivation for the floating blacks method. Maybe you can tell me if this is close to the mark or not ...

To make things simple, lets assume that the drive electronics can drive any plasma cell with digital values ranging from 255 (max brightness) to 0 (min brightness). This range of 256 drive values represents the full dynamic range of the cell. One issue that would exist is it takes non-zero time to charge a cell from 0 to 255, or to discharge the cell from 255 to 0. In other words, the larger the voltage difference from one pixel state to the next, the longer it takes to get to the next state (i.e. it takes time to raise/lower the voltage on a capacitive load). This charge/discharge time affects how fast the display can run.

To see what that would mean in practice, assume for a moment that Panasonic's panels are capable of really dark blacks (e.g. dynamic range of 0 to 255) and that Samsung's panels aren't quite as nice in that regard and have relatively brighter blacks (e.g. dynamic range of only 32 to 255). If both the Panny and Samsung drive circuits were identical, this means the Samsung panel would run faster because the voltage change from darkest to brightest wouldn't take as long (if/when the need occurs in a frame change). The Panny on the other hand would be slower but have darker blacks. In this regard, there is a tradeoff between panel speed and dynamic range.

Taking my guess one step further, maybe Panasonic feels that a dynamic range for any particular APL level of 255-32+1=224 is good enough - even if the plasma cells can handle 0 to 255. This would be beneficial because the panel could run more reliably at a faster speed (again, smaller voltage swings means faster charge/discharge cycles). But maybe the clever engineers @Panny decided to employ a twist ... during high APL, plasma cells would be driven between 32 to 255, and during low APL, 0 to 223. The two APL regimes have the same range (so the display runs at the same fast speed) but the low APL regime appears to have very dark blacks (0 compared to 255 as opposed to 32 compared to 255). That seems like a big win except the absolute black levels in each APL range are different. So the circuit & firmware responsible for the range switching should be designed carefully to avoid the perception of floating blacks.

Another potential advantage of this method should be that power consumption would be lower by not using the full dynamic range of the cell. Given California's new HDTV power regulations for 2011, this would help keep Panny in a large market.

Anyways, this is just a guess, but we have exactly these same issues in my line of work (design of solid-state memories). Your use of the term "drive stability" made me think of this analogy.

Best.

I had a 75u that only shifted during high APL scenes, but my G20 does like you mentioned above. I've tried to lower my brightness on both THX and Custom modes but it seems to make little difference. Neither does the AGC control or Gamma control.

You said you were able to adjust it out of your set, is it possible that some sets may be affected more than others?

I have come across several HDNet programs that have elevated blacks in the source as well. Typically on my PDP (141 monitor) to reach MLL I need to drop the brightness to -2. However, with these HDNet programs I need to drop it to -20+ at times. I think D-Nice called this phenomenon 'black pedestal' or something along those lines.

Cheers

Your on the right track by considering time. However, PDPs voltage remains relatively constant throughout the 255 levels. It uses PWM (varies the "time-on") to create different gray levels. Because PDP relies so heavily on time, anything that limits time will limit both the number of gray levels and the peak brightness of the PDP. Some parameters that limit time in PDP operation:

1 - resolution (more pixels to address means less time to emit light)
2 - discharge delay (time lapse from voltage application to actual discharge)
3 - Length of initialization period (initialization does not contribute to gray level and thus eats up time)
4 - # of initialization periods

Reduction of #2-4 above opens up a lot of time which enables improvements in brightness and # of gray levels. Specifically, reduction of #2 allows for high speed addressing and reduction of #3-4 creates lower black level as well as enabling higher brightness and more gray levels.

In the case of floating blacks only 1 intialization period is used below 6% APL and 2 intialization periods are used above 6%. So you are correct that "time" is saved by using the floating black system. However, I still see the main advantage as increasing contrast (better blacks) at low APL.


MLL (produced by the initialization period) is not part of the 0-255. MLL can be raised or lowered via voltage changes or # of periods without changing the 0-255. In other words MLL is not a gray level and is not intended to be part of the signal. It is unwanted background light.


Good thinking, when only 1 initialization period is used there should be a small power savings.


Cheers

I think you're being too generous

Its clear to me now a PDP only resembles a computer memory in that there is row/column addressing. I'd like to learn more about how PDP's work, especially the drive process (sort of like the process of writing to a memory). I found this nice set of slides:

http://www.docstoc.com/docs/24683452...anel-Principle

But there is no text so it leaves one guessing. Are you aware of any similar resources I could study? I suspect a lot of what I want to understand are covered in slides 1-21. If I can understand the timing diagram in slide 21, I should have a foundation to move on to the subfield drive topic, etc.

And BTW, it sounds like initialization is similar to pre-charge in a dynamic RAM where prior to writing, the memory cell spends some time "getting ready" to be written to ( a capacitor is being charged). Perhaps initialization is getting rid of charge from a previous cycle in preparation for taking on a new value in the next cycle ... but that's just a guess.

Yes there are two row electrodes and one column electrode. There is no TFT control in PDPs. Control of the pixel is achieved through dielectric memory charge (wall charge).


Slide 21 shows the typical cycle the PDP goes through to generate a single subfield. The example they give is a very old conventional style where each and every subfield contains an initialization step that wipes all charge from every sub-pixel on the panel. Then to write image data, each row of sub-pixels is addressed and individual sub-pixels are set into an "on" or "off" state by creating wall charge only in the cells to be "on" and not creating wall charge in the cells to be "off". Once this is achieved the entire panel is sent AC voltage pulses and only the cells previously set in the "on" state will emit light. It then repeats this cycle about 8-10 times per frame with each sustain cycle having different amount of AC pulses to create PWM.

Check out this post on the Panasonic rising black thread:
http://www.avsforum.com/avs-vb/showt...3#post17831713


In the case described in the pdf you are exactly right. In general the initialization step (also called reset step) serves to equalize the wall charge in every pixel (in this case it means erasing all wall charge to zero) to prepare the array to be written with image data. The 2nd important reason for the initialization step is to generate seed electrons into the discharge space (called priming) which enables the gas to be discharge quickly when voltage is applied in the write step and sustain step.

I didn't realize how important that is until I read your other post more than a few times.

At first I couldn't understand why shorter initializations result in lower black levels than longer initializations. I understood zeroing out the voltage across the cell by bleeding out charge from the subpixel must be a necessary step (so you're always starting from a known state). But I thought the more time spent doing this, the "better" the reset would be. Then I remembered your last sentence above. If a longer initialization also means more time priming, then more seed electrons are going to be introduced. I'm guessing more seed electrons will make it easier/quicker to increase current across the subpixel later on. But maybe this will also create more electrons that can be excited to create phosphor-knocking UV photons? In which case the black level might not be as low as it would have been had a shorter initialization (and slower "turn on") been used? I do remember the PWM signal in the sustain phase is responsible for establishing grey level, but that must be on top of a DC level established by the seed electrons?

I would appreciate knowing what is happening from a gas perspective in the three phases. Again, I'm somewhat unsure and guessing again:

Initialization - Apply voltages across subpixel to bleed out charge. Then apply voltages to inject "seed" electrons and produce +/- charges on opposite walls. Gas is now a low-energy plasma with ionized atoms and electrons? Or perhaps the gas isn't a plasma yet, we've only managed to inject/tunnel some electrons across the dielectric barriers?

Address - Increase voltages and current across subpixel, gas is presently a plasma, introduce more carriers into the plasma until electrons jump their valence shells. Electrons now produce UV photons upon returning to lower energy states. UV photons strike phosphors and emit visible photons.

Sustain - Continue applying voltage (at lower level?) to continue electron excitation. I'm not sure why there are multiple sustain cycles within a single subfield cycle. Could this be some sort of power-saving step given that glowing has a some kind of inertia (glows for awhile, then decays)?

Finally, I was curious in your other thread, you say as the panel ages, the voltage needed to initialize must increase. Is this due to some fatiguing mechanism in the dielectric?

Hopefully after all of this, I will have a big picture and understand how the "floating black level" issue fits in.

Best

Yes, well more accurately, the presence of seed electrons reduces the discharge delay. Ionization of the gas becomes easier (less voltage required) and quicker (less delay between voltage application and ionization).


The seed electrons are created via a secondary electron effect. The top two electrodes are coated in MgO which has a very high secondary electron coefficient in Xenon/Neon. When the cell is discharged, high energy ions impact the MgO and electrons are ejected into the discharge space. The black level is visible because during the initialization step the gas is discharged (ionized) creating UV radiation. The strength of this discharge determines the brightness of the black level.

Now, if priming (seed electrons) is very efficient and long lasting then there is less need to prime and thus the black level can be reduced by lowering the voltage or by reducing the number of initialization steps per frame.


Yes, but if there is no image data sent to a pixel, there is no PWM, but the pixel still must undergo initialization to keep up the priming. This is why you see a black level MLL no matter what is on the screen.


The initialization is usually a ramp pulse that creates a small discharge (ionization) in the townsend region. This charges/discharges the dielectrics and at the same time generates secondary electrons from MgO as mentioned above.

The Address step selectively discharges (ionizes) to create enough wall charge for the sustain step.

The sustain step applies a AC pulse train because each pulse phase creates a discharge followed by quenching and then reverses polarity and repeats. The longer the AC pulse is applied the more discharges occur and the brighter the light emitted.


They do not explain it in the patent. However, I speculate that the MgO material is sputtered and deposited onto the phosphor in a form that is less electrically active (lower secondary electron efficiency). This will increase the voltage required to discharge the gas, especially in the initialization step.

BTW, I love talking about this stuff so thanks

Cheers

Hi all,
we too, in Italy, are fighting with this new floating-black feature. It has been reported by some owners on the G20 official thread on our biggest AV forum. It looks like there is no way to get it sorted. Really disappointing during "night" watching movies.

Knowing very well the other feature (rising blacks - we have it too) and the official answers from panasonic reported by CNET, I think this one could be the result of the new system panasonic is using on these 2010 series to apply different voltages to the panel. My 2 cents.

Hi
Just my feedback from Portugal we also have de same problem, some G20 owners reported de Floating blacks problem.

I have seen random reports of floating blacks on some Panasonics for at least the past 3 years here at AVS. Maybe even longer. It's not like it's a new issue or unique to the 13G line.

True, however they did eliminate it in the 12g models (which also happen to be the ones that the rising blacks were first spotted on by normal people) maybe they thought that by putting it back in then nobody would notice the rising blacks.

Why Panasonic can use #1 initialization period instead of 2 with APL below 6%? Because if the pixels in the previus cycle were dark then with only 1 period you can reset correctly the pixels? Instead, if the picture was too bright then you need more time for reset?
Another question, if the picture is dark and ALL pixels don't use some sub-fields (es. sub-field 7 and sub-field 8) then the initialization period of these sub-fields is totally supressed?
Thanks.

The priming effect created by the initialization is transient but still long enough to allow for the below 6% APL case. Like I said earlier, I think the reason they switch to 1 initialization period below 6% APL is solely to increase contrast. However, the panels natural/stable operation required 2 intialization periods.


I'm not sure.


Not quite. REAL BLACK DRIVE has an initialization period in every subfield no matter what but there is a catch. See below:

REAL BLACK DRIVE (Panasonic)
1 or 2 "all cell" initialization periods (every subpixel is initialized no matter what)
All other subfields use "selective" initialization periods (only pixels that have previously generated a subfield)


CLEAR DRIVE (Pioneer)
A single "all cell" initializatin period
Contiguous subfields