Video Capture and Colorspaces
Video in TIB AVPortal:
Video Capture and Colorspaces
Formal Metadata
Title 
Video Capture and Colorspaces

Alternative Title 
Graphics  Colorspace

Title of Series  
Author 

License 
CC Attribution 2.0 Belgium:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor. 
Identifiers 

Publisher 

Release Date 
2016

Language 
English

Production Year 
2015

Content Metadata
Subject Area 
00:00
Demo (music)
Expert system
Basis <Mathematik>
Hecke operator
Motion capture
Mereology
Software maintenance
Sequence
Kernel (computing)
Causality
Figurate number
Color space
Videoconferencing
00:54
Metre
Discrete group
Point (geometry)
Distribution (mathematics)
Matter wave
Connectivity (graph theory)
Computergenerated imagery
Range (statistics)
Discrete element method
Area
Power (physics)
Number
Element (mathematics)
Type theory
Different (Kate Ryan album)
Interpreter (computing)
Core dump
Electronic visual display
Lie group
Sensitivity analysis
Standard deviation
Distribution (mathematics)
Graph (mathematics)
Infinity
Motion capture
Matter wave
Power (physics)
Digital photography
Arithmetic mean
Wave
Infinite impulse response
Phase transition
Order (biology)
Right angle
Radiant flux
Object (grammar)
Species
Videoconferencing
Color space
Freeware
Spectrum (functional analysis)
Impulse response
Row (database)
04:49
Motion capture
Videoconferencing
Color space
05:29
Tuple
Sensitivity analysis
Video projector
Matter wave
Texture mapping
1 (number)
Set (mathematics)
Mereology
Measurement
Mathematics
Core dump
Set (mathematics)
Daylight saving time
Videoconferencing
Matrix (mathematics)
Negative number
Electronic visual display
Color space
Imaginary number
Geometric quantization
Covering space
Area
Source code
Theory of relativity
Structural load
Point (geometry)
Bit
Maxima and minima
Motion capture
Food energy
Demoscene
Category of being
Digital photography
Arithmetic mean
Googol
Order (biology)
Computer science
Right angle
Quicksort
Linear map
Reading (process)
Spacetime
Point (geometry)
Functional (mathematics)
Observational study
Transformation (genetics)
Maxima and minima
Chromatisches Polynom
Discrete element method
Graph coloring
Theory
Power (physics)
Number
Machine vision
Peripheral
Term (mathematics)
Wellformed formula
Green's function
Gamma function
Plugin (computing)
Linear map
Form (programming)
Multiplication
Military base
Forcing (mathematics)
Heat transfer
Basis <Mathematik>
Color management
Matter wave
Transformation (genetics)
Cartesian coordinate system
System call
Sign (mathematics)
Loop (music)
Personal digital assistant
Übertragungsfunktion
Function (mathematics)
Calculation
Infinite conjugacy class property
Videoconferencing
Color space
Marginal distribution
Matrix (mathematics)
Tuple
15:54
Meta element
Context awareness
Group action
Video projector
Source code
Range (statistics)
Water vapor
Open set
Computer programming
Different (Kate Ryan album)
Core dump
Videoconferencing
Physical system
Fundamental theorem of algebra
Touchscreen
Aliasing
Gradient
Sound effect
Bit
Motion capture
Digital signal
Demoscene
Category of being
Arithmetic mean
Process (computing)
Order (biology)
Summierbarkeit
Spacetime
Directed graph
Aliasing
Point (geometry)
Slide rule
Addition
Electronic data interchange
Connectivity (graph theory)
Control flow
Device driver
Mass
Latent heat
Internet forum
Computer hardware
Energy level
Codierung <Programmierung>
Traffic reporting
Default (computer science)
Standard deviation
Graph (mathematics)
Information
Heat transfer
Basis <Mathematik>
Cartesian coordinate system
Limit (category theory)
System call
Frame problem
Graphical user interface
Word
Nonlinear system
Übertragungsfunktion
Function (mathematics)
Video game
Game theory
Color space
Family
Matrix (mathematics)
Standard deviation
Topological vector space
State of matter
Texture mapping
Multiplication sign
1 (number)
Insertion loss
Function (mathematics)
Mereology
Dressing (medical)
Usability
Mathematics
Matrix (mathematics)
Electronic visual display
Color space
Circle
Information
Geometric quantization
Logical constant
File format
Closed set
Point (geometry)
Range (statistics)
Price index
Band matrix
Interface (computing)
Linearization
Normal (geometry)
Website
output
Right angle
Ideal (ethics)
Resultant
Functional (mathematics)
Codierung <Programmierung>
Image resolution
Real number
Electronic program guide
Heat transfer
Chromatisches Polynom
Discrete element method
Theory
Hypothesis
Revision control
Machine vision
Causality
Green's function
Gamma function
Linear map
Graph (mathematics)
Color management
Transformation (genetics)
Device driver
Computer graphics (computer science)
Kernel (computing)
Geometric quantization
Pointer (computer programming)
Infinite impulse response
Computer hardware
Calculation
Transmissionskoeffizient
Videoconferencing
Electronic visual display
35:39
Email
Complex (psychology)
Presentation of a group
Greatest element
Video projector
Multiplication sign
View (database)
Range (statistics)
Set (mathematics)
Function (mathematics)
Shape (magazine)
Open set
Mathematics
Roundoff error
Type theory
Different (Kate Ryan album)
Analogy
Videoconferencing
Color space
Electronic visual display
Circle
Software framework
Information
Hard disk drive
Data conversion
Endliche Modelltheorie
Physical system
Touchscreen
Data storage device
Sound effect
Bit
Maxima and minima
Motion capture
Instance (computer science)
Term (mathematics)
Control flow
Arithmetic mean
Digital photography
Repository (publishing)
Chain
Order (biology)
Software testing
Right angle
Pattern language
Resultant
Spacetime
Point (geometry)
Functional (mathematics)
Link (knot theory)
Device driver
Heat transfer
Chromatisches Polynom
Graph coloring
Product (business)
Power (physics)
Latent heat
Internet forum
Software
Authorization
Energy level
Computer multitasking
Utility software
Software testing
Codierung <Programmierung>
output
Condition number
Data type
Default (computer science)
Game controller
Standard deviation
Multiplication
Military base
Surface
Projective plane
Planning
Ripping
Cartesian coordinate system
Limit (category theory)
System call
Frame problem
Maize
Loop (music)
Geometric quantization
Shader <Informatik>
Übertragungsfunktion
Function (mathematics)
Calculation
Video game
Transmissionskoeffizient
Videoconferencing
Color space
51:50
Uniform resource locator
Googol
00:06
OK I think of starts figure and cause basis so I'm er I go maintainer of to the city of Linux subsystem of the news kernel and so the 2nd part of last year I made got some questions about cost basis I realize that I didn't really knew what it was all about of both of and 2 books started researching experimenting with ions I in this is no expert but I know a heck of a lot more now and this talk is really going into to solve the backgrounds of its and especially what are the practical all sequences of what is going on and hopefully I can have a nice demo later as well what the hell is cooler
00:58
the starters that's of the 1st thing that you need to know about is what's called the spectral Bauer follow distribution so basically I'm looking at them on objects and around getting light of different wavelengths they hit my eyes and the spectral power distribution basically gives you for the wavelength is another meters of the visible spectrum how much power do you get In those wavelengths from the object are these are a few ABC there few standards light source standardized lies light sources and these are the our spectral power sounds great so that's that's physical the physical thing that support just he gets light of different wavelengths into your eyes with certain intensities the on your right does not have the what is it of 350 different so uh wavelength detectors we have to do with free short wave medium wave long waves usually called as and is not that they are specific to 1 wavelength they have overlap and they detect certain ranges but with different efficiencies 1 is means in practice phase that's but like with a certain spectral distribution distribution hits the eye and it is converted in basically the values that electrical intensities of electrical impulses 1 for each going so short wave mean were a long way and those 3 values are what are interpreted as or by you bring because you were you basically having this spectral power distribution can have many many infinite many different a ways you know the graphs but all usages 3 values it means that there is an infinite number of spectral power distribution that will give you the same value so to reproduce a polar polaron and displayed on a photograph you do not have to reproduce aspect about this and you just have to fake it so that's what your I didn't 3 intensities the QIC spot the same as duration of but the actual spectra politicians can be can build a totally different so it is a pure or is a pure illusion what you have on the displayed there has absolutely nothing to do with the original it'll freight the old point is the the the light emitting for whatever lands whatever is in their they creates the illusion of the same core nowdeceased cluster this pure biological Our is different a species can have different number of lights sensitive the elements in eyes with the record holder identified as the mentor shrimp is sensitive to 12 different costs so an RGB display that wouldn't work for them they would need another uh 1 with 12 different components core components in order to reproduce apologetic and see that is where the freed RGB discrete values come from this purely relate to work your biological II is doing I know of course the next step
04:52
this your brain his interpreting its and others can throw up really we things if you don't this is I think the strongest optical illusion that I know the maybe oldest but certainly i is 1 I can open see I see of orange Monday magenta of blue and green there however who agree on a
05:18
single by removing I have to go back
05:22
nobody believes that green blue no
05:28
single I have I have been
05:31
experimenting with this before just you know blocking out 0 just a little bit or that's I keep seeing it only 1 I'm 1 0 almost getting rid of the margin but then it's suddenly turns is cover slightly very slowly actually just terms changes scholar those back to it's an X I and I have not
05:52
seen a convincing explanation of this are some people there are a few ideas that people have 3 as the single 1 that I believe doesn't seem to fit the it's it's an
06:04
amazing 1 I'm so whatever so this is where basically saying core is a very big plugin but you know computer scientist we do have to try and reproduce so you you were caught some seen on video like this 1 you want to reproduce later on the display it would be nice if I wouldn't look they're purple or blue or whatever other sort of it costs roughly reproduced in the cases we which I would look really we it if that wasn't the case in so all space is all about how to reproduce course on the moral basis is called the CIA see the CIE XYZ based on studies done in the 19 twenties Our where they actually try to figure out how this the i seek or how it on the sensitivities of the i and they basically figured out how to reproduce specific wavelengths using 3 different primary light sources I this is called perceive the scene the CI in RGB color 4 the mother of all almost basis and C. is still commissioning the nation of their commissions for everything including all they they do with 4 Our in practice this turns out to be fairly and we'll be called space for calculations of examples of negative values which I don't understand our how that's possible but in practice I don't care and they have a transformation that takes the RGB color space into what they call the x y z base where everything is all nicely positive I linear in the an easy to calculate is the only thing is you can also the the primaries are imaginary so the actual light sources that it's based on the google fixes this panel physically possible ah but is just a linear transformation to a matrix from RGB to the White the X Y Z and is really all you need to know what it's and this is the foundation of all caused they all course bases are defined in relation to get 1 and it is really cool I think of that this is based on studies done by a pool of 17 people for over 85 years ago as quite I think it's quite amazing but what they did and is still valid today I good to know is that the wine describes luminance xn sets are all components and they're just let us and if you think that this confusing well wait for it it gets much more confusing what I use often this XYZ lowercase where XYZ case are populated like this when a user load more as capital why lowercase x y it's don't blame me the nice thing about this 1 this gap keeps your original wife then fully minutes and the lower case X and y based on this formula they our values for chromaticity would basically is these are the costs with maximum intensity yeah so poor is both the polar itself and then you have how bright it is of why is roughly the brightness luminance and X and Y lower case very are values that's specify color and 3 if you're reading about this then you have to pay attention to where its capital or lowercase and sometimes people don't do it quite right so if you think it's confusing it really is constant a pool of space mean steps so K and RGB right you have values for red green and blue but what's except the pore over at me now there are many many calls of rats which 1 do you mean which which is the core or few rats light source and what is the core a few blue rats light source and the green ones because you need to know that in order to accurately reproduce the so you have 3 lands in this place projector or centers in a camera and those when you want to reproduce a car you need to like them up with enough intensity that you think the i have to receiver seeing a specific or In this you have different always for your light sources you will need to trigger them in different ways so all of space is actually saying OK what what are my colors what are my light sources I need rat 1 green 1 and the blue 1 but exactly what color regrets green and blue is and that is defined in usually just x lowercase x and y because the apple wise women and so we can discard that's that has nothing to do with the polar search and basically so you get a pose space defined by which core of rats which call me was called loop and they basically say if you falls in given displayed a rats at the value at the maximum 6 2 5 5 0 0 4 on the typical RGB display that is that's the call that together get you also need what's called a white points assess what quality white gets avoid Journal all Friedland at maximum well you know if you have RGB for what you should get whites the you have many many things all so you can actually blame is that to specify specifically which white 1 and usually is what's called the the 65 which is a standardized it's supposed to mimic daylight the Our but photographs apparently used the 15 which is slightly different I don't know the details of it is just numbers that courage and if you you end up with a matrix have there used to be physical properties so the old Paciotti's there where the cost of false force in the the SO Trajan would give you these days they tend to be translated try you displayed to whatever actual light sources you have any color space that is defined can be your form this are the moral all calls basis through a 3 by 3 matrix with application the it also mean and you also hear about all this discourse space can only represents so many costs and that almost placed him more as both really true all space is just at the Wesley matrix multiplication so in theory all cost spaces can represent the same course are the problem comes that's certain color spaces will have course where you certainly have negative values and you can't have minus 10 rats it's not physically possible for the calculation you can still retrieved your original policy but once you go to the physical world if you get negative values or values above 1 there that off and that is what will leave its you amount of course that you're a physical device can reproduce the debt that is actually gets very confusing in the beginning that you don't really understand what it is but yes X Y Z so typical exploited caused by these that can reproduce all costs but the Newman I can see every all tuples space can do the same but want to go to the physical world then you will have the detections your original space defined here there all linear so if you double the value of rats you will get the rats that is physically twice as intense as as it was written the I like year is not linear so what they what every course Bayes defines is what is called the transfer function and fortunately is also called got 1 that's not rederive name a transfer functions are way all the problems I the power function of usually the 1st part is linear and then follow function takes over and it is usually a lot more of our In June with what the eye can see so what i is more sensitive with differences in certain areas and the of the in order to tell that the core is actually specify its uh after the power function has been applied the use of quotes while sometimes use quotes quite often they don't because people have no idea about densification and adjusting RGB but
15:54
it really is Argentina quote meaning that you know there's been a transfer function here 1st they have to ways to BCI are also called YUV also incorrectly but again all historical so that is you're right for RGB so why why you why cities be sure for those who don't know frequently why is all this RGB is red green blue but for video you quite often 1 luminance and then to or components the eyes more sensitive to luminance so you're luminance resolution is actually higher and i is less sensitive to or core resolutions of all resolutions actually lower so it's a more efficient way of to reproduce the scene which means that you save bandwidth which makes it hard for manufacturers will happen i is also why you typically don't do that for you know you DB green because this yeah close to your screen and you can actually see the effect of Y CBC artist is known as the goal of fidelity is not as good as it should be but when you have moving video you want to see if no clue to convert from RGB to ICT sioc again you have a matrix and color spaces they actually define which matrix you should use or make because the different ways of doing it why C B C are are contrary to what 90 % of all documentation says is not a color space it's different encoding for all how to improve efforts that that is what the Apollo space dust that tells you what those 3 values actually mean to what call or do they should they end up I I Seidel time I had the same uh wrong idea my had started out with this and clearly most people writing documentation datasheets make the same mistake yes sir nice facts about open GL if you use step by default it expects linear RGB so without the transfer function since most pictures that gets J. back whatever they are all with the transfer function so RGB should be quoted you need to tell that to open GL if you don't do it's the images misinterpreted just for fun so this what you have the you have the original based Daniel I want to go to all space and I will get to that in the next slides that you if you want to use this the 3 between the matrix then it passes through a transfer functions so all the red green and blue components they are converted by function to a new value linear RGB then it can go through 3 by 3 matrix to nonlinear why cities there is really is no linear why since large Nevins and in the end lost 1 this step you quantize it's so you go from this floating point value soon say 16 to 235 and you just got any values outside of the that's where you basically go to the physical world it's everybody still follow it's got a lot of you know you know this is noticed this states how where it is linear RGB at the time of the the story of of the know of this How come that the just frescoed little bit to the whole space itself I will be the 1st 1 person and property questions so as RGB this the best known because that's the default for any graph at work that you do Our a standardized why this standards a so all of space is defined so 1st of all is defined by which all primaries do use so what what does it mean if say red and green and blue to which cost the theory days reproduced and it's also called chromaticities same name for that the chromaticity source as arginine are defined and are identical to recommendations I should probably get go through these as RGB graphics base as an pt Alan 70 ever that this standard definition DVD was based recommendation 709 is highdefinition color space is just give a little bit of context all the standards that will refer to 1 another so you get really really if used very quickly this thing that data 701 by definition under 70 and standard definition as RGB useful computer graphics the fact that there are 3 4 very common applications it already an indication of confusing everything is that the 1 B bloody mess so for as RGB the chromaticities are identical to recommendation 7 I had to go back to that chromaticities they are going here so those defined which red green and blue on lands or sends which you are using so what what course art sensitive to exactly and they are really they are defined in the standard In order to go from the or regional so the horizontal pure perfect specification of polar so Hollywood encoded in that color space to have 3 by 3 matrix 3 was the matrix is calculated for the chromaticities that are part of the standard again chromaticities basically what color red green and blue do use the projector or display and what call whites do you get if you turn them all at the same time as RGB however so the chromaticities are the same as highdefinition TV break perhaps they're interchangeable you think note or not both so they decided to have different transfer functions so again normal pictures and graphs computer graphics kernel nonlinear arguably they're the only nonlinear RGB was planned for a future in between follow which is not the same between as RGB and http no idea why they didn't do that I as RGB is almost always best the standard name says redgreenblue so you rarely use the Y CBC yard encoding again this just a calculations is just a different way of representing caloric has nothing to do with the chromaticity and there is actually 1 defines as wise and the so it's it's I think is used in some game like for months or computer graphics not going into detail there then you have standard definition of space so that's the old style all the company's also called as in the as speedy scene on the transfer function is the same as that for high definition the but it has slightly different chromaticity so the actual if you want to reproduce it you would have slightly different red green and blue cost I how to encode from RGB to why C R is standardized by the DT 6 all ones that have lots of people before we have very few of doing any sums video you heard about that 1 the BT 601 standard just defines how you go from RGB to why these it is not a costly it says nothing about you could make its purple and the for a bold circle chromaticities you also use usability 601 to convert from RGB to it's very often used called color space is not usually if it's called like that and really an alias to the standard the there isn't as an alternative encoding that you can use so by default 601 they will and goats in the range from roughly 16 to 2 55 there is 1 that will use the full range from 0 to 255 called extended family with encoding so by definition cost space and it is actually very nice standard because it defines everything in just 1 standard as so what you have in practice if you have standard definition TV the right following get standard and you want to display its it'll say a high definition output the then the Y C B C or encoding by the BCR encoding is different between 2 dresses or saying but 2 why cities SNA encoding different so you have to compensate for that if you want to haven't as RGB image and display
25:47
highdefinition TV everything is the same except for the bloody transfer so if you do not compensate for that you will get will cost are below chromaticities between standard definition highdefinition or for that matter as on the as I've mentioned here they are different however the differences are mind you I will actually show it later that what differences are if it's basically not almost undetectable so nobody cares about that but you will see the difference between the Y. C B C or encodings and the transfer function there are a few of us that are in use Adobe RGB I'm not going into too much detail there BT 20 20 that's an interesting 1 because that is the 1 that is going to alter our highdefinition TVs for k the the standard is there but as far as I'm aware the equipment is still in the research on the I'm there might be some cinema systems out there that can actually do it but I'm I don't know if they exist definitely way my pay grade level by 1 thing 1 of the things at the center says that you only use it for the people are of and it's or more per component Our BT 2020 gives somewhat slow larger government so the primary such chosen in such a way primary cause of chosen such that it can reproduce many more polished than the existing all of space physically reproducing so apparently you can get many more on especially very highly saturated and almost growing costs are next to impossible to reproduce with the existing cause basis the that 1 should be able to do it in to be able to do much better job I've seen some articles in the press about gets and seems to be really call of some all the ones uh 240 and that was basically for about 10 years in place for the early highdefinition equipment has been superseded by the recommendations of results Ntsc your original color space nobody uses it anymore read superseded by as the ones at the end the same for policy comes again going to the same almost space up and what so I told you didn't put that chromaticities all standard definition I definition are slightly different for this almost detectable in practice the the companies the studios are going completely to 7 9 the problem with highdefinition you get signaling through the EDI D and flew in 4 French francs to my or called 7 channels for display water that tells you what color space use standard definition TV you do not have that information so if you would like for example playback and real programs that was not in your original post based you wouldn't be able to know that so that that read that information has to come from outside which space standard definitions but in practice a I I'm going to be so far I have limited the full range so we had know chromaticities what redgreenblue the you have yet to transfer function then you have why should BCR encoding and a loss 1 is called quantization so we go from floating point to actual values they have to have 2 possibilities 1 is full arranged in order is limited range the RGB typically I 0 2 2 5 5 words quantization why B are are typically limited range but you have the L 2 notes as well so you have limited range RGB I have a full range twice a BCR I've seeing this I've never seen death but it's still life who with considerable both unfortunately a AC y so you look to the spec and look how good a and the whole spaces and quantization of all the good stuff i is really complicated but this is I think in 1 slide more or less what you with you have to take care of so if you as a transmitter if the sink receiver can all do a support a lighting for friends and all those for example DVI connector or something like that or you can also handle why CBC after the old for ideas of basically anything but our standard definition or highdefinition resolutions you use fullerenes as RGB just what you would expect if you're sending out standard definition or highdefinition so know ball or something you something you have to use limited range as RGB I have no clue which idiots came up with that or why you would ever want to do that that is the standard so if you do it wrong you're likely to get wrong policy unfortunately some equipment never implemented this so they would for example expect full range anyway it's another and total mass so if you have a citywide uh please just filling in for frames and just give all the right information I have not heard if anyone knows why they came up with using limited range are to be here let me know because I have no clue why would ever want to do that anyway if you are in a more same situation and you actually have in info frames and again you can signal that for graphics computer computergraphic useful range centered definition limited arranged as beauty once itself then the definition of places highdefinition recommendations it's support all of additional formats including the new would later say to my standard in groups to new ultrahighdefinition format as well as what can signal if you want to implement this and I've been actually been doing that for the world lost 2 3 weeks in Linux drivers the it's not easy but but luckily this is my 1 minute guides definition used as highdefinition you that anything else is that your your just that it's a very good people so what the problems all this subversive all highly confusing names are inconsistent so you you know you you get to calculate calculating C with the main thesis so we go through the literature and everybody has a slightly different matrix I applications most applications ignored is completely full range limited influence quantization usually ignored or don't bring correctly I transfer functions are seen to be a really bad boy when it comes to hardware registered more so go from so the no haste 87 604 I've been looking at the states to my receiver if you give it highdefinition TV and you want to convert it to RGB to show you displayed it's users to reward it it just ignore sold transfer so you end up with an S. RGB picture that is not using the as RGB transfer function but the 1 from recommendation 7 so it's actually role but the good news is that support for oldschool spaces they have been added the input of report 19 it I for displayed so that this Corvus Aksumite 2 . 0 as far as I can tell completely for DisplayPort there are a few missing mostly because this but what standard basically says nothing about it they say they are dead but they do not give me any pointers to standard so what it means if anybody has experiences DisplayPort and these whole spaces these with no the only 1 that I could figure out what the BCI B 3 which is used as seen last and as a prop stand for that it could be added political this it anyway so as what it for now some resources some books 1st 2 is very practical this goes to make a lot more likely a lot more into the theory of all and physical and biological and everything related to it 1 website found very useful goes a lot into the mathematics going from 1 color space of more of the calculations work the and the video phonetic specification this is the very latest version where websites if you go to the chapter on cost basis i've completely written it's and the fairly extensive and practical guide to how calls basis work 4 questions I am
35:43
trying to perceive that works to give a little bit of give a demonstration of what what exactly does that mean when you do something wrong with corn the but I think you see if I can make this work but so what I'm started now is the vivid driver those wearing my previous presentation a lot more about it is basically a the difference driver that is generating the image and you can use it to its it's emulating heart the nice thing about it is that it can emulate so any color space so putting a lot of effort to get so you can actually say generates pictures with what you have here as RGB color space but is which you get is a standard course almost all of our although quite they actually if you look here in the past but on other it is the color space conversion of our life paid special attention at all of course used here can be represented accurately or without going out of range in all the all the color space if you would go from to say for example 100 % Boulevard where maximum values of new green whatever then there might be out of range in all and that would suck up everything so if you want to compare costs want to see the effect of different color spaces you need to have go a step Our reproducible by all the older so let's do so and keep this as the the reference so this is how I of course remember this I have absolutely no clue what for a production system is of the project so I'm not I don't know if the score follows from you know they're probably not correct and all what I'm interested in is to see the effect of doing something wrong this course so I just need to see the difference so I am I have 2 instances of this the just framework I'm going to know what the next 1 will have the bottom 1 is to reference that is what it should be on the book the top 1 is where I'm going to make changes Our 1st if just scorespace nothing happens what you see is very slight flickering sometimes right and so Q freeforall 2 which is a test utility I use that application that that reproduces that makes a picture it's understands all these calls and it so big shaders inside open GL that will convert from whatever it receives to as RGB which is the output so although figuring that you see is when is is it is getting signal that there is a new color space but it is still this plane 1 frame is old 1 so that's also the next frame it will have been pulled up and it's using right shapes and this too conclusions so changing goals bases here makes no effect we but need you need to make an small change here I just remember why you feel that it's correct find that OK so what I can do now this so let me send out as RGB but I cannot tell so here I program the driver to send cost this Pacific all of space and in this step this is what the application is used by different of to detect so do whatever the brighter gives you but I can say no no no no I want interpret it as highdefinition TV now you see a difference right and this is the difference caused by different transfer functions because the transfer function is different between as RGB and it 2 and from the authors of the normal and our so difference between recommendation of knife in his RGB so if you do not take that into 2 into account you get the wrong color I will never see this you know if you just watching a video or whatever you will not see this if you only see it if you put it next to 1 another or perhaps if you're an artist and you really experience the scoring you can can detect these things the mean if you're a photographer if you're a seeing a mile and a cameraman you work with coral they are we have on circle of people from General Motors you know they make new models and they they really want to know what all you powers it's important to them that is derived for if you do not take this into account you get the reproducible polar I don't think so what happens if you that is as bad as as RGB further certainly 7 9 let's say standard definition and so what so what standard conditions no 1 the sounds screen I'm saying use the wrong why CBC our conversion the so this is now converted using I this the on will do something about the differences are a fairly small you can see from early in the green and here that you don't have call cost that is the effect that you get if you use the wrong Y C B C R to RGB encoding is actually a bit more stronger uh feasible here on the left if you that need and this right and the text book this is I I can show is that it was
42:51
the not this is when you mess up for range limited range and this is not an the projector or messes things up it's much but the visible my level that is right switch it around but were inch that's a bit yeah now this means that visible the visits to write that's if you if you had a limited full quantization incorrectly and invest effect get now 1 final 1 and I wanted to show I know that you see but yeah but this if I called it where it right it's getting very complicated to see what you're sending which you receiving and how you interpreted and what that this year but so these are the differences between the chromaticities standard definition I definitions as the set is it's it's next to impossible to detect and I'm going back and forth you see very slight differences there the and here and In practice these all next to impossible to think most people just ignore that 1 so with the this tool you can you can use this to moralist detect differences in what what what is the effect of a get it wrong to really see a store chain off your your poaching chain is correct you need to do a lot more but at least this gives you an idea of what if I do it wrong whatever the results I very nice as well this Gehenna there's a generator inside here and also in view of Linux people CTL 2 the that test pattern generator can generate all those different color space so you can actually feed it straight into u AC my transmitter loop at to NAC Myers Seaver and compare which is sending in which of getting back to check whether your chain is correct but you never messing up New York all spaces i is actually what I've been doing in Boston it's work just make sure that's and you know or is it messed up somewhere along in the link I I hope is gives a little bit of an idea of the complexities that you are getting too if you have to deal with all the QC photo tool is publicly available in the future of his repository it has all the shaders that we use to do the calculations this pattern generator is also available at the this again it is all the calculations to send out right policy so if you're interested in depth and you want to work with a you can just look it up I spent a lot of time trying to get this right because the only way to check if your heart risk correct to 1st have to write this tools I would feel that I just scratching the surface uh also feel that I still really don't know everything there is to know what all but the least I hope this is a little bit of in an impression of the single and I'm opening up the floor for questions the the and and then we get more all and the like they do a ball spectacle space needs to define so the question is order entry any standards that that defined white point a proper standard doing this scorespaces will define the red green and blue primaries ends the white points if it doesn't define a 1 point its multiple space of kind of like I find that I would do it by the way Bob I realized forum are correctly is in as MPEG1 70 and standard so it's not in the BTC if you're looking into BED 601 standard as well but those that's not only about that of the so that sept 170 m my of all of I will tools standards with so I'm pretty sure it's is there all I know it's there the the and what was the what would you the whole of the it that I not on 2 or more question is why do you need to white point at all the white point is necessary to I so you have you have these 3 courses but you basically need to know if you put all costs the sale of the like bolts let's let's make it really physical if you give maximum power to all the white belts you should get the widest cooler but which 1 is it exactly the main also you know you if you put maximum to it some might actually shine a little brighter than the other so you might come up with a slightly greenish white were slightly bluish white you don't want that so the white points is actually telling you effectively it's telling you how much power you should give to its primary to achieve that white that specific white collar goddamn multiple ways how you could describe why points but I think that is the most practical so you actually it's a way of finetuning your call was based so that if you ballots to go to maximum power where do you end up and a way of describing it is that it gives you the relationships or on how much what the maximum values are for each call maximum politicking give I took me i it's it's complicated and eventually to me also quite a long while before I OK this idea of what it meant I think it's fairly poorly described in practice but again the easiest way for me is just to say if I want to give 2 5 5 2 4 5 2 5 5 to my display what White should end up and that translates to how much power should 3 light sources gets to get there this way I can describe the is a technical advantage using limited range of forwards the short answer is no our Y. is limited range use it basically comes from again the oldstyle analog systems where you need it had to because all analog so sometimes you're yours electrical signal would actually dipped below your minimum or goal both you maximum so you need had room to accommodate it basically alleges he things crawling all see although I know that some decoders and goes for example they can have rounding errors will you effectively can end up in roughly the same situation the use of the I up and I think that's another 1 in the no more questions which we understood this whole how understandable that could could try find what is called the lights that made be honored it and