00:34 Raster Images (mosaic) VS Vector Images (just math) 05:37 Definition of Colors (visible wavelength spectrum); Eye Receptors (cones); colors do not exist universally, they're just our perception 11:14 RGB: The three color values in our visual system (Red, Green, Blue). We can represent them in the [0,1] range; color space 15:34 Low-Dynamic-Range (LDR) vs High-Dynamic-Range (HDR); how to store/represent them 22:26 Raster Image Storage (interleaved VS separated channels); popular file formats; image compression; animation 50:16 Raster Image in Memory (interleaved, scanline order, swizzled order - typically Z-shaped) 55:57 Questions / Closing thoughts
Thanks very much for sharing your lectures, Cem. Your lectures are engaging, and give more insight into real applications. I am preparing to use math now I have not used since my early days in electrical engineering. Look forward to more of your explanations!
If we were to find the area under the curve for the graph at 11:48 (Sensitivity in our eyes for specific colors), what would that area under the curve represent?
The area under each curve would correspond to the expected sensitivity to a flat light spectrum (you can think of this as white light, though the definition of a white light spectrum varies and it is not always flat, see: CIE Standard Illuminant D65).
Is there an example of what is "beyond white" for HDR value storage? I'm having a hard time picturing what this translates to. I know my monitor has a brightness setting, so I can make all colors brighter or more intense. Can we represent these HDR only values on the color space cube at 15:12?
well it's a month late but they'll appear white, from what I understood, think in terms of intensity. Consider the situation where you're looking at the sky on a very bright sunny day. Think of the view you see as a 2d image (like you're looking at a photo) and the light that enters your eyes as colour data of this photo. Your eyes will try to compensate for excessive brightness by constricting the pupil to reduce the no. of photons that enter the eyes, you can approximate this to dividing the intensity of all colours by a fixed value. In this case you'll be able to distinguish between the sun and the sky. Now consider for a moment that the data is in ldr form. So all bright light will be capped at 1 (not preserving the data in this case). You won't be able to distinguish between the sky and sun. They'll look white to you. I found this very good image that reflects just that: vfxdoc.readthedocs.io/en/latest/shaders/img/LDR-HDR.png The image doesn't show if your pupils do constrict in the ldr case, this would be simply reducing the brightness of the overall image, you won't get back the details to distinguish between sky and the sun. I hope this makes some sense :) I edited the original image to what a similar brightness adjustment would look like: imgur.com/a/nsvWGkG
Would be great to add a slide showing the use of HDR keeping the values that are out of the visible spectrum when editing them. i.e. two images side by side HDR vs LDR before and after adjusting the gamma/b&c values where you can see how the LDR clips when it runs out of data while editing. Good example would be an indoor shot looking out a window where depending on the light level you either lose all the outside data or all the inside data.
Really great lecture, thanks so much Cem :) Been a programmer for past decade and only recently discovered a passion in programming graphics. Wish I found this sooner! Does the University of Utah offer other online graphics courses such as specific courses in path tracing/ray marching/ray tracing and so forth? Are there preliminary math courses for graphics at all at your university specifically or is Calculus, Analytical Geometry and Linear Algebra enough to get a good understanding of graphics?
Just to clarify, this is not an online course. This is a course for which I have recorded lectures and I've decided to share them publicly online. I plan to do the same for my other graphics courses, as time permits. Obviously, we have more courses at the University of Utah and not all of them have recorded lectures.
@@cem_yuksel Thanks for the clarification and again thanks for sharing the lectures! Look forward to the rest of your content/lectures as and when you release them :) Will take a look at the university's website. Take care :)
In the HDR we are splitting the color into 64K values. So we have more levels of each RGB colors. Doesn't that mean if we do color manipulation , we still do color value clipping in HDR? Because just like the LDR (256 levels), if we multiply the colours in HDR by 2, we need 128K levels to represent the result, but we only have 64K. Someone explain this, please.
Hi…thank you for sharing the videos of your lectures. Would it be possible to share the slides too? Would love to take notes on them. Love these high quality lectures.
![Lp. Сердце Вселенной #60 РОЖДЕНИЕ ЛОЛОЛОШКИ [Финал] • Майнкрафт](http://i.ytimg.com/vi/YoR0pAV9FVQ/mqdefault.jpg)
00:34 Raster Images (mosaic) VS Vector Images (just math)
05:37 Definition of Colors (visible wavelength spectrum); Eye Receptors (cones); colors do not exist universally, they're just our perception
11:14 RGB: The three color values in our visual system (Red, Green, Blue). We can represent them in the [0,1] range; color space
15:34 Low-Dynamic-Range (LDR) vs High-Dynamic-Range (HDR); how to store/represent them
22:26 Raster Image Storage (interleaved VS separated channels); popular file formats; image compression; animation
50:16 Raster Image in Memory (interleaved, scanline order, swizzled order - typically Z-shaped)
55:57 Questions / Closing thoughts
@memgen-憂鬱頭 You're welcome! I didn't finish all the videos, though. I hope to do it eventually. 😅
Godsend!
Thanks very much for sharing your lectures, Cem. Your lectures are engaging, and give more insight into real applications. I am preparing to use math now I have not used since my early days in electrical engineering. Look forward to more of your explanations!
00:00 Intro
00:41 Raster Images
05:44 Color
22:28 Raster Image Storage
25:09 Popular Raster Image Formats
33:16 Image Compression
48:50 medical images use more than 8 bits per channel for what they call window and leveling in DICOM images used in CT scanners for example.
7:21 minor correction: frequency is actually (wave) speed over wavelength
Unimportant slide deck nitpick: At 18:00 the 16 bits should represent 0 to 65535.
Yes, indeed!
If we were to find the area under the curve for the graph at 11:48 (Sensitivity in our eyes for specific colors), what would that area under the curve represent?
The area under each curve would correspond to the expected sensitivity to a flat light spectrum (you can think of this as white light, though the definition of a white light spectrum varies and it is not always flat, see: CIE Standard Illuminant D65).
@@cem_yuksel Okay, got it! Thank you Cem :)
Is there an example of what is "beyond white" for HDR value storage? I'm having a hard time picturing what this translates to. I know my monitor has a brightness setting, so I can make all colors brighter or more intense. Can we represent these HDR only values on the color space cube at 15:12?
well it's a month late but they'll appear white, from what I understood, think in terms of intensity. Consider the situation where you're looking at the sky on a very bright sunny day. Think of the view you see as a 2d image (like you're looking at a photo) and the light that enters your eyes as colour data of this photo. Your eyes will try to compensate for excessive brightness by constricting the pupil to reduce the no. of photons that enter the eyes, you can approximate this to dividing the intensity of all colours by a fixed value. In this case you'll be able to distinguish between the sun and the sky. Now consider for a moment that the data is in ldr form. So all bright light will be capped at 1 (not preserving the data in this case). You won't be able to distinguish between the sky and sun. They'll look white to you. I found this very good image that reflects just that: vfxdoc.readthedocs.io/en/latest/shaders/img/LDR-HDR.png The image doesn't show if your pupils do constrict in the ldr case, this would be simply reducing the brightness of the overall image, you won't get back the details to distinguish between sky and the sun. I hope this makes some sense :) I edited the original image to what a similar brightness adjustment would look like: imgur.com/a/nsvWGkG
Would be great to add a slide showing the use of HDR keeping the values that are out of the visible spectrum when editing them. i.e. two images side by side HDR vs LDR before and after adjusting the gamma/b&c values where you can see how the LDR clips when it runs out of data while editing. Good example would be an indoor shot looking out a window where depending on the light level you either lose all the outside data or all the inside data.
Really great lecture, thanks so much Cem :) Been a programmer for past decade and only recently discovered a passion in programming graphics. Wish I found this sooner! Does the University of Utah offer other online graphics courses such as specific courses in path tracing/ray marching/ray tracing and so forth? Are there preliminary math courses for graphics at all at your university specifically or is Calculus, Analytical Geometry and Linear Algebra enough to get a good understanding of graphics?
Just to clarify, this is not an online course. This is a course for which I have recorded lectures and I've decided to share them publicly online. I plan to do the same for my other graphics courses, as time permits. Obviously, we have more courses at the University of Utah and not all of them have recorded lectures.
@@cem_yuksel Thanks for the clarification and again thanks for sharing the lectures! Look forward to the rest of your content/lectures as and when you release them :) Will take a look at the university's website. Take care :)
In the HDR we are splitting the color into 64K values. So we have more levels of each RGB colors. Doesn't that mean if we do color manipulation , we still do color value clipping in HDR? Because just like the LDR (256 levels), if we multiply the colours in HDR by 2, we need 128K levels to represent the result, but we only have 64K.
Someone explain this, please.
By any chance is it possible to access other computer graphics related courses taught by University of Utah?
Hi…thank you for sharing the videos of your lectures. Would it be possible to share the slides too? Would love to take notes on them. Love these high quality lectures.
ua-cam.com/video/zllIPDaiOyk/v-deo.html
Thank you ! Good video !
Thank you!! :)
شرح مو حلو بنو 😭