Panasonic full HD 3D plasma screens

[MfA]

I wonder how this is going to work. It's still only a 600 Hz subfield panel ... if you want to do 120 Hz for 60 Hz per eye (blech, I really wish they had standardized on 72 Hz per eye like cinemas) that means you can't do 24 bit color any more.

So presumably they will do 15 bit color with dithering in 3D mode? That's going to create waves ...

 

[homerdog]

that means you can't do 24 bit color any more.

Sorry for the noob question, but why?

 

[MfA]

Plasma's are essentially PWM ... each subfield is one bitplane.

600/120 = 5 bitplanes, times 3 for RGB is only 15 bits.

 

[Mize]

Plasma's are essentially PWM ... each subfield is one bitplane.

600/120 = 5 bitplanes, times 3 for RGB is only 15 bits.

welcome to 30 Hz

edit: so are all the 3D TVs in development based on shutter-glasses? LCDs are inherently polarized so you'd think there would be some uber-cool way to interleave the polarized glass at 90 degree rotations per pixel-column and then just double-up on the horizontal resolution

 

[Gubbi]

I wonder how this is going to work. It's still only a 600 Hz subfield panel ... if you want to do 120 Hz for 60 Hz per eye (blech, I really wish they had standardized on 72 Hz per eye like cinemas) that means you can't do 24 bit color any more.

So presumably they will do 15 bit color with dithering in 3D mode? That's going to create waves ...

It looks like a stop gap solution alright. I guess they are reusing their existing drive ICs. They need to either double the resolution or increase the subdrive frequency.

I'd opt for double the resolution since multiplexing two images on the one display at double the frequency is going to result in lower contrast ratio and false colours (the phospors only have half the time to decay, and you can have local divergence, black for right eye, white for left eye).

Cheers

 

[homerdog]

Plasma's are essentially PWM ... each subfield is one bitplane.

600/120 = 5 bitplanes, times 3 for RGB is only 15 bits.

Thanks, after about 20 minutes of google time this makes sense. :smile:

So each frame is split into 5 (or 10 for 60Hz material) subfields, each containing 1 bit of color information for each phosphor of every pixel. Is it hard for the television to do this math (are some better at it than others?)?

Edit: The math I am referring to is splitting a frame into 5 bitplanes.

 

[Mize]

This doesn't make sense.
By your logic a current plasma at 120 Hz / 60 Hz would have 2 bitplanes = 6 bits?

 

[ferro]

How exactly will this work when playing 24p content (e.g. 3D Blu-Ray)? 24 fps for left and right equals 48 fps, which would mean 600/48=12.5 sub-fields per frame. Maybe it doesn't need to be a multiple of 48? Or maybe it's 600 sub-fields for both the left and right frame, for a total of 1200 sub-fields?

 

[Gubbi]

This doesn't make sense.
By your logic a current plasma at 120 Hz / 60 Hz would have 2 bitplanes = 6 bits?

The subfield drive frequency of all current PDPs are between 480Hz and 600Hz. The more subfields, the more pulses and the higher the fidelity.

Cheers

 

[MfA]

I'd opt for double the resolution since multiplexing two images on the one display at double the frequency is going to result in lower contrast ratio and false colours (the phospors only have half the time to decay, and you can have local divergence, black for right eye, white for left eye).

I don't think phosphors are a problem, phosphors on modern CRTs already had decay times near a millisecond ... I assume these plasma displays have similarly fast phosphors, so you won't get crosstalk. You can double the brightness, but you can't get rid of the pixel gaps you get with interleaving.

They strobe, they don't sample and hold. LCDs have a bigger problem here, especially the 120 Hz displays ... if you have a 120 Hz LCD where a full image update (scanline order) takes 1/120 Hz you have literally no time to open the shutter for 60 Hz per eye, at any time there are always two images on the screen even if the response time is infinitely small. You need a 240 Hz LCD to get the necessary margins for a single image (and you need really bright strobing backlights, so you can just open the shutter inside the period where the image is stable).

 

[MfA]

edit: so are all the 3D TVs in development based on shutter-glasses? LCDs are inherently polarized so you'd think there would be some uber-cool way to interleave the polarized glass at 90 degree rotations per pixel-column and then just double-up on the horizontal resolution

The light coming out of the normal imaging LCD is not strictly polarized one way or another (and it fundamentally can't be, it uses that degree of freedom for the modulation together with the subsequent polarization filter). You need a second non pixelized LCD (and a QWP for circular polarization) to make it work, so it's more expensive ... you can buy monitors which do this though.

 

[Gubbi]

I don't think phosphors are a problem, phosphors on modern CRTs already had decay times near a millisecond ... I assume these plasma displays have similarly fast phosphors, so you won't get crosstalk. You can double the brightness, but you can't get rid of the pixel gaps you get with interleaving.

Panasonic (And Pioneer before they closed PDP shop) did claim they improved phospors.

I just know I have blue trails on my 7th gen Pio, and my brother has green trails on his 8th gen Panasonic, - mind you, only notice this with a test-DVD or by playing Gears of War. I'm guessing that image processing compares the incoming frame/field to the just displayed one and correcting for phosphor decay (the same way LCD correct for ghosting). Of course, it can do nothing when you have a white to black transition, - I've expect image quality to be considerable lower in some instances where divergence between the two stereoscopic images is large.

Cheers

 

[vazel]

Panasonic is improving phosphers to get rid of the green/yellow streaking. Google 'panasonic fast decay'.