LCD and Plasma technology: it's imperfections detailed. Test pattern included. - AVS Forum
Forum Jump: 
Reply
 
Thread Tools
post #1 of 50 Old 10-16-2010, 01:56 PM - Thread Starter
AVS Special Member
 
8:13's Avatar
 
Join Date: Sep 2006
Posts: 1,250
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Basic Design

LCD uses a backlight.

Plasma is ionized gas: gas charged by electricity resulting in ionized gas.

Plasma Technology.

Between two plates of glass is a cell. The cell holds gas and phospor.

The dielectric is below the glass and holds the sustaining and scanning electrodes.
These electrodes discharge electricity into the protective m g 0 layer - the m g 0 layer sends electrons to the gas inside the cell ionizing the gas.

The charge from the m g 0 excites the gas in the cell, this sends ultraviolet light beams on the phospors; red, green, blue, that coats the inside bottom layer of the cell. When the phospor is lit up by the ultraviolet light, the phospors electron jumps up to a higher energy level and the atom heats up, then the energy level goes back down and this is when the phospor beams visible light back. The visible light is directly proportional to the number of photons randomly colliding with the coated surface.

When the phosphors energy level goes back down it beams visible light back towards the m g 0 and out from the glass towards the viewer.
The phosphors glowing time is called afterglow.
If the afterglow is still emitting when the next electrical burst is sent, the phospor didn't go to the energy level where it doesn't show light.

It acts similar to a seesaw (also known as a teeter-totter).
The ultraviolet light beams on the phospors, the phospors energy level goes up, then down, beaming visible light back towards the m g 0, instantly after the visible light dims the next ultraviolet light burst is sent raising phospors energy level again, and back and forth it goes: ultraviolet light, then visible light.

Phosphor persistence, also referred to as afterglow, may still show visible light when the other frames ultraviolet photons are sent to the phospor, this is what makes it show afterimage. The afterglow can last up to a minute, meaning the energy level of the phospor is still high enough it shows visible light.
In 3D TV's terminology afterimage is called Crosstalk: the image showing a persisting image of the previous frame.

LCD Technology.

The backlight goes to the diffuser that distributes the light across the screen evenly.
The diffuser light then goes to the polarizers which hold two electrodes, The polarizers align the light sent through the electrodes.
Inbetween the electrodes is twisted liquid crystal.

The liquid crystal is given a electrical charge making them untwist, and when they untwist they block light from the diffuser.
The LCD Drive twists one by one.
While twisted they send light from the diffuser to the colored filter on top of the electrode and below the polarized panel resulting in a red, green, blue color.

The problem is the way color is shown: The LCD drive circuit uses Dithering.

Frame Rate Control (FRC) Dithering uses 4 frames to show more color than the RGB sub pixel can show.

2 frames are shown with different colors and they blend to make a new color in the third frame and this color is held into the fourth frame, making the 120Hz framerate show at 30fps after FRC dithering.

You may think that since 240Hz LCD TV's use 120Hz circuits in 240Hz LCD TV's this is the cause of the ghosting.
But, if they used true 240Hz circuits instead of 120Hz the liquid crystal would have to twist twice as fast as for the 120Hz circuit, one at a time instead of in parallel. Using true 240Hz LCD drive circuit is not the solution. The solution is something called Parallel LCD drive circuit architecture.

Parallel LCD drive circuit architecture refreshes pixels in parallel instead of one by one, shortening the whole refresh cycle.
"Such architecture requires more circuitry and digital frame buffer with parallel IO...No need for dithering when parallel." - Mathew Orman

One of the three problems of crosstalk in LCD monitors is due to the LCD Circuit twisting one at a time, by twisting many in parallel the problem is solved according to Mr.Orman.

Right now the LCD Drive is 120Hz and twists the liquid crystal one at a time.
The parallel LCD Drive keeps the same 120Hz as before but twists many in parallel instead of just one at a time.
This way by staying at 120Hz the pixel charge is not making the liquid crystal twist twice as fast, which is what would happen if the LCD drive went from 120Hz to 240Hz.

The second problem is capacitance, which is defined as the ability to store charge.

The Liquid Crystal when it shows Color, is twisted. Then it untwists and Black Frame Insertion is shown. If the Twist showed very bright color then the Liquid Crystal holds charge past the BFI.

The Third Problem is the Liquid Crystal is too heavy, and when twisted to show color has enertia that has to be overcome when the Black Frame Insertion is shown next.
If the weight of the Liquid Crystal is lowered it will twist faster when showing color, and showing the black screen.

So the Liquid Crystal itself needs the properties of low weight and low capacitance.

3D Test Patterns.

Pattern one is looked at with one eye closed, and the open eye looks where the pointer says to look for that eye.
What's being pointed at is black and white bars. Look at the color of the black or white bar and compare it to the IRE scale of white to black squares next to it.
The black IRE square the black or white bar looks like is the amount of crosstalk your TV has.

Pattern 2 is only supposed to have white in one bar at a time, anymore than One white bar and your seeing crosstalk.
You'll need to either guess if the bar is too wide or use a dslr camera white the shutter speed at 1/1000 and take a photo of the screen white it's playing the 3D test pattern.

Both patterns are in side by side, 16:9 ratio, left eye first.

Pattern 3 is in 8x8 and 16x16 pixel squares.
It is in side by side 3D, left eye first, 16:9 aspect ratio 30 seconds long.
It has these three frame rates per eye: 50, 59.940, 60.

These are easy to use. Watch the 16x16 videos first, and look for the bars to become lit up wrong.
The bars aren't supposed to stay lit up when the white square isn't in it.
The white square is exactly 16x16 pixels, it's a square, so if you see a long white bar inside the pattern it's showing a error in the TV.
Or maybe you'll get color in the boxes, or maybe the boxes will get partially filled after the white box left it, these are all errors that it's not working properly.

I made these tests for the TV manufacturers to use so they could implement motion compensation, or the soap opera effect 120Hz TV' have.
But I don't mind at all if regular people use it to test 3D TV's in the store.
You don't need a camera to see the errors, you don't need to wear the glasses that 3D TV's use. Just put the pattern into the TV from a USB stick and watch the entire 30 seconds and look for any errors I told you about.
If the pattern is a square as it moves there is no crosstalk, if it's too fast to see then use the dslr camera method pattern 2 uses.

Link to test patterns

____________________

This next test pattern is 1080p, 16:9, side by side left side first, 1 minute long.

To use these to test for ghosting/crosstalk. Wear the glasses and watch the videos in 3D on your TV.
Then, close one eye and see if the video is showing two images of the same pattern.
A double image would look like a rectangle rather than a square.

NTSC

High Contrast version.
High Brightness version.

PAL

High Contrast version.
High Brightness version.
____________________

Quote:
Originally Posted by icester View Post

You can confirm it by playing a movie that has side by side configuration and contains left image with text "left image" and right image with text "right image".

By pausing the movie you can look through your DLP-Link shutters glasses with right eye closed and see if the image says "left eye" Then do the pay /pause sequence several times and see if you always get "left image" text when your right eye is closed...

Mathew Orman

Link to quote.

Link to test files
16:9 aspect ratio, Side by Side Left side first, 1080p resolution - squished

Also these test files are good to test for ghosting.

While playing the test files close one eye and see if the text is doubled.
Ghosting is visible if "Right Image" and "Left Image" text is being shown at the same time with one eye closed.
You wear your 3d active polarized or passive polarized glasses for this test.


There is new, and then you are new.
This is a moral of the bears and their cereal.
8:13 is offline  
Sponsored Links
Advertisement
 
post #2 of 50 Old 10-17-2010, 08:45 AM
 
icester's Avatar
 
Join Date: Apr 2010
Posts: 1,175
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by 8:13 View Post

Basic Design
Both Direct and Reflective display technology "LCD and Plasma", display 3D images using reflection.
The direct technology means the reflection is held for a very short time, but it is still reflective so there is a afterimage when compared to a 3D DLP display.

The technological problem is then how to make the technology use reflection with no afterimage?

The problem is solved in direct Plasma Technology using better phosphor material so reflection is held for a very short time.
So the better the phosphor material the shorter the reflection is held.

Isolating how to improve Plasma technology is the next problem.

Plasma Technology.

Between two plates of glass is a cell.

The dielectric is below the glass and the dielectric holds the sustaining and scanning electrodes. These electrodes discharge electricity into the protective m g 0 layer - the m g 0 layer sends electrons to the gas inside the cell towards the address electrode.
The charge from the m g 0 causes a beam of ultraviolet photons, and this beams on the red, green, blue phosphor that coats the inside bottom layer of the cell.

It is then that light from the red, green, blue phosphor that beams back towards the m g 0 and out from the glass towards the viewer.
The phosphors glowing time is called afterglow, this is followed by the dimming of the glow.

If the afterglow is still emitting when the next electrical burst is sent, it didn't dim the glow fast enough.
After the glow is dimmed the next electrical charge is instantly sent to the phosphor again, then, it takes a bit of time to send enough ultraviolet radiation to get the phosphor to send light again.
So in a sense the Plasma Technology cycle is reflective in nature, Reflective in that the ultraviolet photons beam into the phospor which in turn beams back out to the m g 0, and may hold a memory of the previous ultraviolet interaction.
If the afterglow stopped emitting the rgb beam immediately instead of slowly then there would be no memory, but the phosphor needs to be able to do that

The thing then is to ask if the afterglow stays lit up when the other frames ultraviolet photons are sent to the address electrode? This is what makes it show crosstalk, muddying up the clean picture.

LCD Technology.

The backlight goes to the diffuser that distributes the light across the screen evenly.
The diffuser light then goes to the polarizers which hold two electrodes, The polarizers align the light sent through the electrodes.
Inbetween the electrodes is twisted liquid crystal.

The liquid crystal is given a electrical charge making them untwist, and when they untwist they block light from the diffuser.
When they remain twisted they send light from the diffuser to the colored filter on top of the electrode and below the polarized panel resulting in a red, green, blue color.

The problem is changing colors - The color is shown by untwisting, then retwisting.
If the 3D picture is using one color and the other eye is then shown a different color, the twisting and untwisting isn't fast enough to prevent the first pictures color to totally untwist and so block the color of light it was showing.
What happens is like with plasma, the second color is shown even before the first color has faded resulting in a unwanted light being seen.

Test Pattern.

I made a test pattern. If you use it and see 4 bars instead of two then your tv is too slow to play 3D content without ghosting/crosstalk.
The pattern should wiggle, but only two bars should be visible.

If you put the test on a memory stick you can test your tv at the store before you buy it to see if you see two or 4 bars wiggling.

I think you have few mistakes:
The only reflective technology is DLP and Mirasol's IMOD.

Note, all the artifacts you get on LCD monitors running at 120Hz are due to color dithering as a basic technology is not capable with full color delivery unless time sequential dithering is used.

Mathew Orman
icester is offline  
post #3 of 50 Old 10-17-2010, 02:21 PM - Thread Starter
AVS Special Member
 
8:13's Avatar
 
Join Date: Sep 2006
Posts: 1,250
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Would spatial dithering fix the problem?

For those who don't know what "Frame Rate Control (FRC) / Temporal Dithering" is it lets more colors be seen by having one frame show one color, then the next frame shows a different color and the third frame is the percieved new color previously unavailable. this new color is held into the fourth frame.

Spatial dithering is the one frame has secondary color pixels in a area being different color and when blended with the main color pixel and the third color is seen.

So with the explaination for 120Hz given by icester, even though the frame rate is 120Hz after frame blanking (with 3D TV's having frame blanking), four frames are used to display one color. So then we divide 120 frames by 4 = 30fps and this is how many frames are in the color made using (FRC) color dithering.

I've edited the first post. Give it a look and see if it's right, please.
The way I've described Plasma is reflective and this helps understand the afterglow. Maybe I could've worded it differently but it seems to work fine.


There is new, and then you are new.
This is a moral of the bears and their cereal.
8:13 is offline  
post #4 of 50 Old 10-17-2010, 02:41 PM
 
icester's Avatar
 
Join Date: Apr 2010
Posts: 1,175
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by 8:13 View Post

Would spatial dithering fix the problem?

For those who don't know what "Frame Rate Control (FRC) / Temporal Dithering" is it lets more colors be seen by having one frame show one color, then the next frame shows a different color and the third frame is the percieved new color previously unavailable. this new color is held into the fourth frame.

Spatial dithering is the one frame has secondary color pixels in a area being different color and when blended with the main color pixel and the third color is seen.

So with the explaination for 120Hz given by icester, even though the frame rate is 120Hz after frame blanking (with 3D TV's having frame blanking), four frames are used to display one color. So then we divide 120 frames by 4 = 30fps and this is how many frames are in the color made using (FRC) color dithering.

I've edited the first post. Give it a look and see if it's right, please.
the way I've described Plasma is reflective and this helps understand the afterglow. Maybe I could've worded it differently but it seems to work fine.

What would fix the problem would be simply parallel LCD drive circuit architecture. At the moment they are refreshed one by one pixels. Such architecture requires more circuitry and digital frame buffer with parallel IO. The reason no one is implementing it is the cost and the fact that 3D is just a marketing gimmick at the moment.

Mathew Orman
icester is offline  
post #5 of 50 Old 10-17-2010, 03:48 PM - Thread Starter
AVS Special Member
 
8:13's Avatar
 
Join Date: Sep 2006
Posts: 1,250
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
This parallel architechture would improve dithering performance, not eliminate it I gather as better drive circuitry doesn't change the Panel Color Depth (6 bits per RGB sub pixel I guess).
So if it improves dithering, would this still be FRC or spatial dithering?
It would have to be spatial dithering since the glasses are 60Hz and if FRC dithering were used the glasses would be showing the frame twice as fast as the TV showed a frame.


There is new, and then you are new.
This is a moral of the bears and their cereal.
8:13 is offline  
post #6 of 50 Old 10-17-2010, 04:02 PM
 
icester's Avatar
 
Join Date: Apr 2010
Posts: 1,175
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by 8:13 View Post

This parallel architechture would improve dithering performance, not eliminate it I gather as better drive circuitry doesn't change the Panel Color Depth (6 bits per RGB sub pixel I guess).
So if it improves dithering, would this still be FRC or spatial dithering?
It would have to be spatial dithering since the glasses are 60Hz and if FRC dithering were used the glasses would be showing the frame twice as fast as the TV showed a frame.

No need for dithering when parallel. Also, Sharp is already implemented wider (actually thicker or richer) color spectrum using RGBY four component pixel architecture. Still only at 60 Hz.

Mathew Orman
icester is offline  
post #7 of 50 Old 10-17-2010, 05:40 PM - Thread Starter
AVS Special Member
 
8:13's Avatar
 
Join Date: Sep 2006
Posts: 1,250
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
For every time you increase the LCD drive circuit architecture Hz, say from 120Hz to 240Hz, you half the pixel charging time.
Meaning the pixel must twist and untwist twice as fast as before.

Parallel LCD drive circuit architecture takes 120Hz LCD drive circuit architecture and makes it 240Hz. But this still needs the Liquid Crystal to twist and untwist twice as fast.


There is new, and then you are new.
This is a moral of the bears and their cereal.
8:13 is offline  
post #8 of 50 Old 10-17-2010, 06:17 PM
 
icester's Avatar
 
Join Date: Apr 2010
Posts: 1,175
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by 8:13 View Post

For every time you increase the LCD drive circuit architecture Hz, say from 120Hz to 240Hz, you half the pixel charging time.
Meaning the pixel must twist and untwist twice as fast as before.

Parallel LCD drive circuit architecture takes 120Hz LCD drive circuit architecture and makes it 240Hz. But this still needs the Liquid Crystal to twist and untwist twice as fast.

No,
at the moment they have to be biased to twist one by one. But if in parallel then the whole refresh cycle would be shortened.

Mathew Orman
icester is offline  
post #9 of 50 Old 10-17-2010, 07:48 PM - Thread Starter
AVS Special Member
 
8:13's Avatar
 
Join Date: Sep 2006
Posts: 1,250
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Ive edited the first post Mr.Orman. Please look to see if it's correct.


There is new, and then you are new.
This is a moral of the bears and their cereal.
8:13 is offline  
post #10 of 50 Old 10-18-2010, 04:47 AM
 
icester's Avatar
 
Join Date: Apr 2010
Posts: 1,175
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by 8:13 View Post

Ive edited the first post Mr.Orman. Please look to see if it's correct.

Again, the Plasma technology is not considered as reflective.
All reflective or transmissive type modulators require a light source. Plasma is the light source of UV radiation which is converted to visible light using RGB phosphors.

Any after image generated by plasma is due only to phosphor persistence.

Here is the definition:

'The property of a phosphor that determines its ability to emit light for a time after the stimulus has been extinguished. Persistence may extend to over a minute.'

Mathew Orman
icester is offline  
post #11 of 50 Old 10-18-2010, 10:42 AM - Thread Starter
AVS Special Member
 
8:13's Avatar
 
Join Date: Sep 2006
Posts: 1,250
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
The Plasma section has been edited. Thank you for your help Mr.Orman.


There is new, and then you are new.
This is a moral of the bears and their cereal.
8:13 is offline  
post #12 of 50 Old 10-18-2010, 10:57 AM
 
icester's Avatar
 
Join Date: Apr 2010
Posts: 1,175
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by 8:13 View Post

The Plasma section has been edited. Thank you for your help Mr.Orman.

You welcome Mr.8:13:-)

Mathew Orman
icester is offline  
post #13 of 50 Old 10-18-2010, 11:50 AM
Member
 
banduraj's Avatar
 
Join Date: Jul 2010
Posts: 142
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
I assume you mean the pattern should display 2 bars without glasses to be able to view 3D without crosstalk?
banduraj is offline  
post #14 of 50 Old 10-18-2010, 12:43 PM - Thread Starter
AVS Special Member
 
8:13's Avatar
 
Join Date: Sep 2006
Posts: 1,250
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Quote:
Originally Posted by banduraj View Post

I assume you mean the pattern should display 2 bars without glasses to be able to view 3D without crosstalk?

Yes, I've edited it so that's clear.


There is new, and then you are new.
This is a moral of the bears and their cereal.
8:13 is offline  
post #15 of 50 Old 10-18-2010, 02:49 PM
Member
 
banduraj's Avatar
 
Join Date: Jul 2010
Posts: 142
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Both of these files fail to play on my Samsung UN55C7000 (FW 1030.1) due to an unsupported audio codec. My guess is that there is no audio in these files and it does not like it.
banduraj is offline  
post #16 of 50 Old 10-19-2010, 01:50 PM - Thread Starter
AVS Special Member
 
8:13's Avatar
 
Join Date: Sep 2006
Posts: 1,250
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Quote:
Originally Posted by banduraj View Post

Both of these files fail to play on my Samsung UN55C7000 (FW 1030.1) due to an unsupported audio codec. My guess is that there is no audio in these files and it does not like it.

I've made new test files. Please try these. The link is in the first post. This is the second update since yesterday.


There is new, and then you are new.
This is a moral of the bears and their cereal.
8:13 is offline  
post #17 of 50 Old 10-19-2010, 03:22 PM
Member
 
banduraj's Avatar
 
Join Date: Jul 2010
Posts: 142
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
These new files also fail to play on my UN55C7000. This time it is due to an "Unsupported Resolution".
banduraj is offline  
post #18 of 50 Old 10-20-2010, 04:44 AM - Thread Starter
AVS Special Member
 
8:13's Avatar
 
Join Date: Sep 2006
Posts: 1,250
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Quote:
Originally Posted by banduraj View Post

These new files also fail to play on my UN55C7000. This time it is due to an "Unsupported Resolution".

I looked at the problem your having and read the manual and tried different encodes and patterns. And then after all that I remembered what I read in the Samsung manual. USB media is played using something called Media Play. 3D is disabled when you use Mediaplay and when you plug media into the TV using USB your using Media Play, so the TV doesn't accept 3D from USB sticks.

Also the resolution the TV goes into 3D mode automatically is 1280x1470, or frame packed. And then further down in the manual it says the decoder in the TV only accepts up to 1080p, so 1440 is too much pixels and that's why the Samsung TV won't play 3D from the USB stick because it needs frame packed 720p and it can only decode up to 1080p, it can't decode that many pixels.
In the process I have remade the frame packing 3D patterns and the stacked vertical patterns. They are linked to in the first post.

You need to play the patterns probably from a PC since it will play the resolution these files are at, or maybe play them through a in-store 3D blu ray player that accepts USB. Powerdvd 10 plays them fine.

Samsung Manual quoted in three jpeg pictures.


There is new, and then you are new.
This is a moral of the bears and their cereal.
8:13 is offline  
post #19 of 50 Old 10-20-2010, 04:59 AM
Member
 
banduraj's Avatar
 
Join Date: Jul 2010
Posts: 142
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by 8:13 View Post

I looked at the problem your having and read the manual and tried different encodes and patterns. And then after all that I remembered what I read in the Samsung manual. USB media is played using something called Media Play. 3D is disabled when you use Mediaplay and when you plug media into the TV using USB your using Media Play, so the TV doesn't accept 3D from USB sticks.

Also the resolution the TV goes into 3D mode automatically is 1280x1470, or frame packed. And then further down in the manual it says the decoder in the TV only accepts up to 1080p, so 1440 is too much pixels and that's why the Samsung TV won't play 3D from the USB stick because it needs frame packed 720p and it can only decode up to 1080p, it can't decode that many pixels.
In the process I have remade the frame packing 3D patterns and the stacked vertical patterns. They are linked to in the first post.

You need to play the patterns probably from a PC since it will play the resolution these files are at, or maybe play them through a in-store 3D blu ray player that accepts USB. Powerdvd 10 plays them fine.

Samsung Manual quoted in three jpeg pictures.

OK. I do remember 3D being disabled in Media Play. And all digital video files are played in Media Play. But I have taken side-by-side video files and played them on my Samsung. You just need to turn on 3D when the video starts and select the correct 3D mode. I will try that later this evening when I get home.
banduraj is offline  
post #20 of 50 Old 10-20-2010, 05:56 AM - Thread Starter
AVS Special Member
 
8:13's Avatar
 
Join Date: Sep 2006
Posts: 1,250
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
I made the test patterns 720p resolution to be sure it's not too large when you test them later on. Thank you for helping me make the patterns run by replying so far.


There is new, and then you are new.
This is a moral of the bears and their cereal.
8:13 is offline  
post #21 of 50 Old 10-20-2010, 07:56 AM
 
icester's Avatar
 
Join Date: Apr 2010
Posts: 1,175
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by 8:13 View Post

I made the test patterns 720p resolution to be sure it's not too large when you test them later on. Thank you for helping me make the patterns run by replying so far.

Here is another test patter for ghosting test:

http://www.stereoforum.org/viewtopic.php?f=16&t=53

Mathew Orman
icester is offline  
post #22 of 50 Old 10-20-2010, 09:56 AM
Member
 
banduraj's Avatar
 
Join Date: Jul 2010
Posts: 142
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by icester View Post

Here is another test patter for ghosting test:

http://www.stereoforum.org/viewtopic.php?f=16&t=53

Mathew Orman

Do you want to translate that for us non-German speaking forum members?
banduraj is offline  
post #23 of 50 Old 10-20-2010, 01:09 PM
 
icester's Avatar
 
Join Date: Apr 2010
Posts: 1,175
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by banduraj View Post

Do you want to translate that for us non-German speaking forum members?

The test pattern is in English and since a while back the Google has built in translator.
I do not speak German.

Mathew Orman
icester is offline  
post #24 of 50 Old 10-20-2010, 02:49 PM
Member
 
banduraj's Avatar
 
Join Date: Jul 2010
Posts: 142
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by 8:13 View Post

I made the test patterns 720p resolution to be sure it's not too large when you test them later on. Thank you for helping me make the patterns run by replying so far.

OK. The 720p patterns play. However, I don't think it's producing the effect you are looking for. If I switch it to 3D mode while the pattern is playing, I only have the options of 2D->3D, over-under or side-by-side. My guess is, if you are going to want your pattern to work on Samsung TV's, it's probably going to have to be in side-by-side format.
banduraj is offline  
post #25 of 50 Old 10-21-2010, 03:51 AM - Thread Starter
AVS Special Member
 
8:13's Avatar
 
Join Date: Sep 2006
Posts: 1,250
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
I made new test patterns from the pattern shown by icester. The pattern I got is the same but from mtbs3d.com. The pattern I have as the source had no hblank, and was not 1080p. So I made the image 1080p and added a hblank of 280 pixels, so it's in frame packing 1080p and side by side format.

When I made the previous test pattern using that source I thought it had a hblank as I just glanced at the image, but I was wrong and it was a honest mistake. These test patterns work when played from powerdvd 10.

I don't think it will play in a 3D TV from USB, but it might. The resolution might be too big.

I don't know the resolution to make the pattern so it'll play from USB.


There is new, and then you are new.
This is a moral of the bears and their cereal.
8:13 is offline  
post #26 of 50 Old 10-21-2010, 04:50 AM
Member
 
banduraj's Avatar
 
Join Date: Jul 2010
Posts: 142
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
OK. I'll give these new ones a test when I get home. I tried the ones you posted from the site icester posted and the 720p side-by-side ones work, however, they do require glasses to test.
banduraj is offline  
post #27 of 50 Old 10-21-2010, 05:25 AM
 
icester's Avatar
 
Join Date: Apr 2010
Posts: 1,175
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by 8:13 View Post

I made new test patterns from the pattern shown by icester. The pattern I got is the same but from mtbs3d.com. The pattern I have as the source had no hblank, and was not 1080p. So I made the image 1080p and added a hblank of 280 pixels, so it's in frame packing 1080p and side by side format.

When I made the previous test pattern using that source I thought it had a hblank as I just glanced at the image, but I was wrong and it was a honest mistake. These test patterns work when played from powerdvd 10.

I don't think it will play in a 3D TV from USB, but it might. The resolution might be too big.

I don't know the resolution to make the pattern so it'll play from USB.

Sorry,
but that is not going to work.
The frame packed format is only related to HDMI 1.4 digital video transmission and those are low level hardware protocols.
Your test clips do not work in both stereoscopic player and media player.
All what you need to do is the content layout format.
The sync and blanking as well as packed format are created in hardware and in case of the PC it is done by nVidia graphics card.
The best is to use: frame sequential 1920 x 1080, dual stream 1920 x 1080 or horizontally or vertically compressed
frame compatible layouts. Doubled full frame in horizontal or vertical layout will not work well with regular video codecs.
There is an instruction on how to create stereoscopic media file on 3dtv.at

Mathew Orman
icester is offline  
post #28 of 50 Old 10-21-2010, 08:29 PM - Thread Starter
AVS Special Member
 
8:13's Avatar
 
Join Date: Sep 2006
Posts: 1,250
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
I've updated the test files using the pattern Mr.Orman linked to. I tested them and they work fine. But I would like to here what you saw when you tried them.

I learned how to make the test files, it required three hblank black strips: one on either side of the two pictures and one in the middle separating the pictures.

The actual pixel size of the hblank determines the parallax. So if the pictures are the same except from a part that changes color, as was this test pattern, the hblank makes them too far apart, so to fix this you would crop the picture so when the hblank adds parallax in 3D mode the picture is in zero parallax.

The test pattern didn't have any hblank and wasn't in 1080p or 720p 16:9 ratio so I had a lot of work to do along with testing.

Now that I know how, I can make a new one of my own design. Maybe another one that moves across the screen one pixel at a time at 68 pixels parallax, side by side like these test patterns.
But this test pattern learning experience took a long time. I think I'll just leave it alone after this and only show the one pattern that's there right now.

Here's the avs code I used in case anyones interested in that kind of stuff:

Code:
loadplugin("C:\\Program Files\\AviSynth 2.5\\plugins\\DirectShowSource.dll")

left=directshowsource("C:\\Documents and Settings\\Jeremy\\Desktop\\1080p\\Simple\\30\\Left.mp4", audio=false, 29.970).crop(70, 0, 0, 0).addborders(70, 0, 70, 0)
right=directshowsource("C:\\Documents and Settings\\Jeremy\\Desktop\\1080p\\Simple\\30\\Right.mp4", audio=false, 29.970).crop(70, 0, 0, 0).addborders(70, 0, 70, 0)

return stackhorizontal(left, right).converttoyv12()
Note that the source this was used on was made into a 30 second 30fps 1080p clip already. The addborders makes it 3d compatible and the crop makes it zero parallax.

With no cropping the addborders for the right side only has one 70: (0, 0, 70, 0), this along with the left addborders makes the video 3D compatible.
The larger the number the larger the parallax, but it needs to be at least 45 pixels I think. I had problems using only 40 pixels so I used a larger number.

I'll attach the pictures I used to make the test video linked to in the first post. When I got the one picture from mtbs3d I used ms paint and made these two pictures.

 

pattern 2.zip 197.31640625k . file


There is new, and then you are new.
This is a moral of the bears and their cereal.
8:13 is offline  
post #29 of 50 Old 10-22-2010, 03:47 AM
 
icester's Avatar
 
Join Date: Apr 2010
Posts: 1,175
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by 8:13 View Post

I've updated the test files using the pattern Mr.Orman linked to. I tested them and they work fine. But I would like to here what you saw when you tried them.

I learned how to make the test files, it required three hblank black strips: one on either side of the two pictures and one in the middle separating the pictures.

The actual pixel size of the hblank determines the parallax. So if the pictures are the same except from a part that changes color, as was this test pattern, the hblank makes them too far apart, so to fix this you would crop the picture so when the hblank adds parallax in 3D mode the picture is in zero parallax.

The test pattern didn't have any hblank and wasn't in 1080p or 720p 16:9 ratio so I had a lot of work to do along with testing.

Now that I know how, I can make a new one of my own design. Maybe another one that moves across the screen one pixel at a time at 68 pixels parallax, side by side like these test patterns.
But this test pattern learning experience took a long time. I think I'll just leave it alone after this and only show the one pattern that's there right now.

Here's the avs code I used in case anyones interested in that kind of stuff:

Code:
loadplugin("C:\\Program Files\\AviSynth 2.5\\plugins\\DirectShowSource.dll")

left=directshowsource("C:\\Documents and Settings\\Jeremy\\Desktop\\1080p\\Simple\\30\\Left.mp4", audio=false, 29.970).crop(70, 0, 0, 0).addborders(70, 0, 70, 0)
right=directshowsource("C:\\Documents and Settings\\Jeremy\\Desktop\\1080p\\Simple\\30\\Right.mp4", audio=false, 29.970).crop(70, 0, 0, 0).addborders(70, 0, 70, 0)

return stackhorizontal(left, right).converttoyv12()
Note that the source this was used on was made into a 30 second 30fps 1080p clip already. The addborders makes it 3d compatible and the crop makes it zero parallax.

With no cropping the addborders for the right side only has one 70: (0, 0, 70, 0), this along with the left addborders makes the video 3D compatible.
The larger the number the larger the parallax, but it needs to be at least 45 pixels I think. I had problems using only 40 pixels so I used a larger number.

I'll attach the pictures I used to make the test video linked to in the first post. When I got the one picture from mtbs3d I used ms paint and made these two pictures.

Please note, that ghosting is also dependent on phase.
You should have another test pattern which have full height. This way you can detect ghosting throughout entire screen.

Also calling black level pixels as hblank (horizontal blanking) is confusing becasue of the video low level signal definition.
I suggest to use just names like 'black or white stripes'.

Here is an instance where the name 'hblank' is used correctly:

http://zone.ni.com/devzone/cda/pub/p/id/1078

Mathew Orman
icester is offline  
post #30 of 50 Old 10-22-2010, 04:54 AM
Member
 
banduraj's Avatar
 
Join Date: Jul 2010
Posts: 142
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Just to be clear. How do I use these to test for ghosting?
banduraj is offline  
Reply 3D Tech Talk

User Tag List

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off