Panasonic Calibration Controls

[fluxo]

I've had a look at the manual (VT30) and the calibration controls are very poorly
explained. Is there any other documentation from Panasonic that describes the
controls more thoroughly?

Some of the things I'd like to know:

In what order are the various controls applied to the input?

What are the units or scaling factors?

Where the manual says of R-Saturation "Adjusts the saturation of red area" what
does that mean exactly?

I presume it means rotating colour vectors that correspond to the "most red" point
for a certain IRE value (i.e., moving the vertices of the triangles that are
sometimes shown in chromaticity diagrams). I would also guess that some form
interpolation is used for secondaries etc., but where is the algorithm or
mathematical formula for that described?

Cheers.

 

[Badger0-0]

Adjusts the saturation of red area

I'd assume it adjusts the saturation of red, simple.

And you're looking way too deep into things.

Just my opinion

 

[Bumtious]

I've had a look at the manual (VT30) and the calibration controls are very poorly
explained. Is there any other documentation from Panasonic that describes the
controls more thoroughly?

Some of the things I'd like to know:

In what order are the various controls applied to the input?

What are the units or scaling factors?

Where the manual says of R-Saturation "Adjusts the saturation of red area" what
does that mean exactly?

I presume it means rotating colour vectors that correspond to the "most red" point
for a certain IRE value (i.e., moving the vertices of the triangles that are
sometimes shown in chromaticity diagrams). I would also guess that some form
interpolation is used for secondaries etc., but where is the algorithm or
mathematical formula for that described?

Cheers.

The manual explains it quite well, as the manuals for calibration are 3 to 4 times the size of the Panasonic instruction manuals and are complex.

What you need to be looking at is how to calibrate a set.

Basically, you start by turning all the excessive processing off, then you set your black level, then you set your white level, then you set your gamma controls and greyscale by adjusting the basic white balance controls followed by the advanced white balance controls going back and forth yo check your white and black levels.Next you adjust the colour management, primary colours and secondary, then go back and check your greyscale and gamma is still intact.

You do this with software and measuring meters. When you have white and black levels right, gamma at a decent 2,4 curve, greyscale nice and flat, and colour gamut point hitting rec709 colour space points then you are done.

Thats the process basically.

No instruction manual will explain this.

You may want to learn how to do this yourself in which case search for ISF or THX calibration, and buy the software and meters needed to measure what you are doing.

Or you could engage a qualified calibrator to do it for you. It takes 4 hours plus to do properly.

You cant do it by eye by just throwing settings at it.

 

[fluxo]

I'd assume it adjusts the saturation of red, simple.

Well, red is always 100% saturated, isn't it? If it weren't it would be some
other colour, e.g., pink, magenta, off-red, whatever.

 

[Bumtious]

Well, red is always 100% saturated, isn't it? If it weren't it would be some
other colour, e.g., pink, magenta, off-red, whatever.

No it's not, if the mix of colours is not right, greyscale and gamma not right then nothing will be right, which is indeed the case with every set that comes out of a box.

What is little known. The colour green is not used in broadcast. Only red and blue, to save bandwidth, red and blue make green so if the mix of red and blue is not right then green will be wrong which is the case almost 100% of the time.

 

[fluxo]

The manual explains it quite well.

What you need to be looking at is how to calibrate a set.

Basically, you start by turning all the excessive processing off, then you set your black level, then you set your white level, then you set your gamma controls and greyscale by adjusting the basic white balance controls followed by the advanced white balance controls going back and forth yo check your white and black levels.Next you adjust the colour management, primary colours and secondary, then go back and check your greyscale and gamma is still intact.

You do this with software and measuring meters. Thats the process basically.

No instruction manual will explain this.

You may want to learn how to do this yourself in which case search for ISF or THX calibration.

Or you could engage a qualified calibrator to do it for you. It takes 4 hours plus to do.

You cant do it properly by eye.

I realize one can twiddle controls this way and that way until it looks
about right, according to measurements taken. But what I'm trying to
grasp is what the numbers in the menus actually mean. For example,
if R-Saturation is -1, how much less saturation is there?

In other words, that 1 less, is it 1% less, or 1 degree less between the
greyscale and red vector, or what exactly?

If someone were to ask you your height and you said "27 high" they
might then wish to know what units of measurement you were using.
Without that, the number is not meaningful, and I hate things that are
not meaningful - but perhaps that's just me!

So, in short, I'd like to know what the numbers in the menus mean.

 

[fluxo]

No it's not, if the mix of colours is not right, greyscale and gamma not right then nothing will be right, which is indeed the case with every set that comes out of a box.

I didn't mean the colour as it is displayed. I meant the colour as it
might be defined.

For example, one way of defining "red" is as a vector <R,G,B> in
which G and B are both zero and R is non-zero. Desaturation of that
colour occurs when either or both of G and B are made non-zero. But,
in that case, the desaturated colour is no longer "red" (by definition).

 

[Bumtious]

I realize one can twiddle controls this way and that way until it looks
about right, according to measurements taken. But what I'm trying to
grasp is what the numbers in the menus actually mean. For example,
if R-Saturation is -1, how much less saturation is there?

In other words, that 1 less, is it 1% less, or 1 degree less between the
greyscale and red vector, or what exactly?

If someone were to ask you your height and you said "27 high" they
might then wish to know what units of measurement you were using.
Without that, the number is not meaningful, and I hate things that are
not meaningful - but perhaps that's just me!

So, in short, I'd like to know what the numbers in the menus mean.

They don't mean anything, they are simply a numerical value that will differ from set to set due to the wide tolerances of mass produced components.

 

[Bumtious]

I didn't mean the colour as it is displayed. I meant the colour as it
might be defined.

For example, one way of defining "red" is as a vector <R,G,B> in
which G and B are both zero and R is non-zero. Desaturation of that
colour occurs when either or both of G and B are made non-zero. But,
in that case, the desaturated colour is no longer "red" (by definition).

I really don't understand what you are saying.

Red does have defined points on a CIE graph, but the only way to place it there is to adjust greyscale, gamma and a 3D CMS control and measure it, and assuming the set actually has the range and correct implementation of a 3D CMS system.

Red co-ordinates on a rec709 chart are

x-0.640 y-0.330 Y-0.213

 

[fluxo]

I really don't understand what you are saying.

Red does have defined points on a CIE graph, but the only way to place it there is to
adjust greyscale, gamma and a 3D CMS control and measure it, and assuming the set
actually has the range and correct implementation of a 3D CMS system.

Given that you're having to assume it, that would suggest you don't know how the CMS
is implemented. Which is my point really.

There are these controls in the menus, but because Panasonic haven't explained what
they do, nobody really knows how they work. So people will mess about with various
settings and by trial and error and taking measurements produce a decent result. But
the scientific way, the efficient way, would be to understand how it works and get to
that result faster, without so much of the trial and error.

 

[Bumtious]

Given that you're having to assume it, that would suggest you don't know how the CMS
is implemented. Which is my point really.

There are these controls in the menus, but because Panasonic haven't explained what
they do, nobody really knows how they work. So people will mess about with various
settings and by trial and error and taking measurement produce a decent result. But
the scientific way, the efficient way, would be to understand how it works and get to
that result faster, without so much of the trial and error.

Now I think I get your point

It will cost far too much money to implement and set up in that way, a calibrator knows what to do with them as they have been trained.

I'm not aware of any set at any cost weather in the domestic, professional, post production, gaming or medical industry that does not need to be calibrated properly.

You have to measure the sets light and colour output and mix them correctly to industry defined standards to achieve the correct result.

Saying that, the VT30 CMS controls do not work properly, this years VT50 does.

 

[Scooby2000]

Given that you're having to assume it, that would suggest you don't know how the CMS
is implemented. Which is my point really.

There are these controls in the menus, but because Panasonic haven't explained what
they do, nobody really knows how they work. So people will mess about with various
settings and by trial and error and taking measurements produce a decent result. But
the scientific way, the efficient way, would be to understand how it works and get to
that result faster, without so much of the trial and error.

Sorry but he does know how it works, you don't seem too. There's no magic numbers and as said all sets are different given the cheap components used. Calibration is about making the right adjustments with the controls using measuring equipment.

 

[Bumtious]

Sorry but he does know how it works, you don't seem too. There's no magic numbers and as said all sets are different given the cheap components used. Calibration is about making the right adjustments with the controls using measuring equipment.

I don't thing he means me Gaz. I think he means generally implemented.

Point is the controls are implemented to work in a particular way, and this year they do what they were intended to do. The CMS controls have a brightness element, a saturation element and a tint element. There are 18 controls just in CMS that need balancing to work as one with each primary and secondary colour moving up and down, side to side at the same rate on the right plains when you adjust the colour control. Any deviation and it screws the picture up.

What he doesn't seem to grasp is that the only way to balance them all is to measure what you are doing.

Even if they used military grade components, the final adjustments would still need to be measured. Best way to do that is in situ. Would cost a bloody fortune to do it on the production line.

 

[Scooby2000]

I don't thing he means me Gaz. I think he means generally implemented.

Point is the controls are implemented to work in a particular way, what he doesn't seem to grasp is that the only way to balance them all is to measure what you are doing.

Even if they used military grade components, the final adjustments would still need to be measured. Best way to do that is in situ. Would cost a bloody fortune to do it on the production line.

Oh I really should be sleeping.

As you say though you need to set the controls and test it's the only way, it's the same with cars, medical kit, everything needs to be calibrated to a degree and tested, first thing I was told when learning electronics, test your testing equipment lol.

 

[fluxo]

Sorry but he does know how it works, you don't seem too.

I'm sure ******** knows many many things I don't know, but this specific
thing is something I don't think either of us knows.

There's no magic numbers and as said all sets are different given the cheap components used.

I never claimed there were any "magic numbers". But there is, I would
suggest to you, a program running on the chip that deals with the whole
CMS thing. And that program would run in an entirely predictable way,
if one only knew how it run.

I think we need to try a little thought experiment. Suppose I have a light
with a digital dimmer switch. That is, I turn a dial and an LED display
shows a number. If the number is higher then the light is brighter. If it
is lower, then the light is dimmer.

Now suppose I want the brightness to be at a specific level. I could turn
the dial up a little and take a measurement. If the light is too bright, I could
then turn the dial back a little bit and take another measurement. And this
is repeated over and over until the brightness is just right.

That is one way of doing it. But suppose I know that the LED shows
the number of Watts of electricity supplied to the light. Knowing how
lights work, knowing what the numbers mean, one could avoid trial and
error. I would just need to take an initial measurement and then I could
calculate how far the dial would have to be turned.

Either method will produce the right outcome. But one method is more
informed and more efficient than the other.

 

[fluxo]

Oh I really should be sleeping.

Can't disagree with that. Strewth! What on earth are we doing here at
this silly hour?

 

[Bumtious]

I'm sure ******** knows many many things I don't know, but this specific
thing is something I don't think either of us knows.

I never claimed there were any "magic numbers". But there is, I would
suggest to you, a program running on the chip that deals with the whole
CMS thing. And that program would run in an entirely predictable way,
if one only knew how it run.

I think we need to try a little thought experiment. Suppose I have a light
with a digital dimmer switch. That is, I turn a dial and an LED display
shows a number. If the number is higher then the light is brighter. If it
is lower, then the light is dimmer.

Now suppose I want the brightness to be at a specific level. I could turn
the dial up a little and take a measurement. If the light is too bright, I could
then turn the dial back a little bit and take another measurement. And this
is repeated over and over until the brightness is just right.

That is one way of doing it. But suppose I know that the LED shows
the number of Watts of electricity supplied to the light. Knowing how
lights work, knowing what the numbers mean, one could avoid trial and
error. I would just need to take an initial measurement and then I could
calculate how far the dial would have to be turned.

Either method will produce the right outcome. But one method is more
informed and more efficient than the other.

Fluxo

That is one light going up and down

We have six colour mixing a whole spectrum. These colours go up and down, side to side and front to back.

Up have to measure them all.

A CMS chip in your TV costs $2. A CMS chip in a Lumagan processor costs over $100. A chip in a £50,000 Runco projector even more. They all need adjustment cuz NON are military grade. Even milatary grade medical monitors need adjustment.

If you have a recording of the Wimbledon men's tennis final in 3D, don't put it into 3D. Watch it side by side. Each "camera" has 2 cameras.

NON of the cameras used had the same colour. The left camera on the base line had different gamut from the right. The service line camera was the other way round. Not one matched each other.

I can easily demonstrate this.

CMS calibration is seriously hard to do.

In your analogy, even if the switch is saying xx watts how do you know the light is working the same as the one next to it.

So even if the chip is outputting perfect red, it is highly probable the path the signal takes to get the screen will interfere with that perfect red, and then you have the voltages applied to the screen and the screen itself.

You are assuming everything is identical.

Well in mass market production, it's miles away from identical.

 

[fluxo]

In your analogy, even if the switch is saying xx watts how do you know the light is working the same as the one next to it.

So even if the chip is outputting perfect red, it is highly probable the path the signal takes to get the screen will interfere with that perfect red, and then you have the voltages applied to the screen and the screen itself.

You are assuming everything is identical.

Well in mass market production, it's miles away from identical.

Indeed, the lights might vary. That's why an initial measurement was
taken using method two. But I'll happily concede it was a simplified
and perhaps not very realistic example.

Still, engineers are often able to predict outcomes with some accuracy
even when there are many variables involved.

When engineers design aircraft, for example, they don't glue on wings of
different shapes and sizes until they find a plane that doesn't crash. They
model these things in computer simulations.

Maybe it is true that a CMS is so fiendishly complicated that no accurate
predictions can be made when it feeds its output into components that
vary so much. It's pretty hard for me to say without knowing exactly what
the program does!

But it's past 2am, so if you say that, I'll happily accept it as fact