MULTECH TEST PDF

INFORMATION THEORY, MATH AND PHYSICS DIGITAL Uses Binary digits Not continuous > discrete values (specific, distinct value) In the past > smoke signals, morse code > now ones and zeros ANALOG VS. DIGITAL Analog > continuous signal, infinite number of values > represent real-world information Bin...

INFORMATION THEORY, MATH AND PHYSICS DIGITAL Uses Binary digits Not continuous > discrete values (specific, distinct value) In the past > smoke signals, morse code > now ones and zeros ANALOG VS. DIGITAL Analog > continuous signal, infinite number of values > represent real-world information Binary code Gottfried W. Leibniz > first Binary concept George Boole > Boolean algebra, system based on 1/0, yes/no, on/off, logical operators AND, OR, XOR, NOT Binary – decimal conversion INFORMATION THEORY Claude E. Shannon > A Mathematical Theory of Communication Applications include lossless (removing redundancy) and lossy compression (approximation) Entropy > amount of uncertainty > how much chaos/disorder/uncertainty is there Entropy is 0 when the probability is 1/0 Entropy is maximal when all messages have the same probability > chaos Always a natural number The higher the amount of information, the lower the entropy Redundancy > predictability, repetition It is the part of message that can be removed without losing essential information People encode information using symbols, create data which are sent in messages Message > encapsulated data Data > made of symbols Symbol > used to encode information Information > what ppl understand 1 Bit and bitrate Bit = binary digit > 1/0 > 8b = 1B = 1 ASCII character (coding table) > 1kB = 1000b In bits > channel throughput (internet, bitrate, bit depth) In bytes > size of files, memory, storage COMPRESSION Processing of data in order to make them smaller > saves storage, time > saves money Compressed data require more computing power Lossless > entropy is the limit Huffman coding > optimal; most common symbol saved as the shortest series of bits (morse code > E) Run-length encoding > compresses the same symbols in a row Lossy > human senses or transparency is the limit Transparency > point at which ppl are not able to differentiate between encoded copy and original Coding format > algorithm how to decode the data (JPEG, H.264, MP3) Standard > standardised coding format (ISO/IEC 10918-1 > JPEG) Codec > SW or HW tool that applies the coding format to the data (JPEG codec, H.264 codec, VP9 codec) Container > a specific file format; an “envelope” that includes video+audio+subtitles+metadata in one file (MP4, MKV, AVI, MOV) PHYSICS Period T[s] = time of one cycle; in seconds Frequency f[Hz] = number of cycles per second; in hertz Wavelength λ[m] = distance between two closest points of the same phase; in meters 2 Voltage U[V] = difference in electric potential between two points; in volts (pressure of the water in the hose) Resistance R[Ω] = opposition to the flow of electric current; in ohms (like obstacles, sand in the hose) Current I[A] = rate of flow of electric charge past a point; in amperes (like diameter of the hose) Power P[W] = work per time; in watts (like how much water flows through the hose) WEBER-FECHNER LAW People do not feel the changes in stimulus linearly DECIBELS A unit of measurement used to express the ratio of two values (power or intensity) > used to measure how loud a sound is/strong a signal is Elementary unit is bel B > decibel dB is 1/10 of bel Nyquist-Shannon sampling theorem When the signal has maximal frequency we need to sample it at least with frequency two times higher than the maximal frequency When the sampling f is lower > aliasing (wheels backwards, audio distortion, moire) Anti-aliasing > gradients, smoother graphics > comb filter > two audio monitors are playing the same source but are in different distance from listener (like the local public announcement) COMPUTER HARDWARE, SOFTWARE AND LINUX SOFTWARE = programs and data on the computer Firmware = low-level SW, built into HW devices, helps them work properly (in cameras, drones…) > BIOS, UEFI OS = operating system > Windows x Unix-like (Linux, BSD (macOS)…) 3  1985 Microsoft BSD = Berkeley Software Distribution > family of open-source OSs > used for servers, macOS LICENCES AND SOFTWARE Copyright > protects the creator, restricts the usage of the work without permission Copyleft > protects the user, public access to work, changes released under the same licence GNU > Richard Stallman > creators and maintainers of copyleft GPL licences GNU General Public Licence (GPL) Free/Libre software = FLOSS > source code available to public, free as freedom Freeware > free as price, source code typically not available, often proprietary Permissive licence > free to use, modify, distribute, minimal restrictions Proprietary licence > restricted access, fees Public domain > free for anyone to use, modify, distribute LINUX GNU/LINUX > Linus Torvalds, most commonly used kernel in the world > Ubuntu, Manjaro… Hollywood, servers, cars, buses, VŠE, Android, Chrome OS… Window manager > defines the graphical user interface GUI > graphical user interface > KDE Plasma, GNOME Shell COMPUTER HARDWARE PC > IBM, Microsoft, Intel CPU = processor > the brain of the computer > Intel, AMD Can be integrated into it > GPU, mini memory Cache, PCIe (interfaces that allows communication with f.e. storage), Network Interface (Ethernet/Wi-Fi), Audio Codec (Sound Card) SMT = simultaneous multithreading > executes multiple threads simultaneously > more efficient 4 Motherboard > main board, everything else is connected to it CPU socket, RAM slots, PCIe slots (graphic, sound, network cards…), SATA connectors (HDD, SSD), M.2 slots (NVMe SSD, Wi-Fi/Bluetooth cards), power connectors, connectors for keyboard, monitor… 5 INTERNAL DATA INTERFACES AND MEDIA Serial ATA (SATA) > interface for storage media > HDD, SSD, DVD-ROM Serial interface > transmits binary data as a series of voltage pulses PCI Express (PCIe) > interface for GPU, NIC (network interface card), SSD, sound and capture cards >> M.2, U.2, USB-C NVME (Non-Volatile Memory Express) > logical interface for SSD, fastest SSD drives >> M.2, U.2, PCIe RAM (Random Access Memory) > memory for running processes; volatile memory (without power loses data); nowadays DDR5 RAM Storage > non-volatile memory (stores data without power) >> SATA HDD, SATA SSD, NVME SSD (M.2, U.2) HDD (hard disk drive) > mechanical drive with platters, moving parts > cheap, unreliable, noisier > for data archiving SSD (solid state drive) > flash memory (no mechanical moving parts), has own CPU and RAM > speed depends on interface (SATA < PCIe) GPU (graphics processing unit) > the graphical processor on graphic cards > nVidia, AMD, Intel ASIC (Application-specific integrated circuits) > en/decoding of video Filesystems > NTFS (proprietary file system by Microsoft), EXT4, APFS (Apple proprietary) Networking > WAN (wide area network), LAN (local area network), PAN (personal area network) Speed in megabits, gigabits IP (internet protocol) EXTERNAL DATA INTERFACES AND MEDIA USB (universal serial bus) > half-duplex (transmission of signal in both directions but not simultaneously) Thunderbolt > allows connection of monitors, GPU, NIC, power delivery Storage cards CFAST > uses serial interface, predecessor CompactFlash > fast, expensive XQD, CFExpress (physically same) SD Card (secure digital) > now 4th gen. > microSD, miniSD, SD UFS (universal flash storage) > in modern smartphones 6 Wires > fibers x metallic (cheap, limited in speed and length) Metallic > UTP (unshielded twisted pair), DAC (direct attached copper) Fibre > Optical fibre (uses light to send data, long distances, almost unlimited bandwidth RJ-45 (with UTP) SFP Wireless network > Bluetooth (latest 5.2, WPAN = wireless personal area network), Wi-Fi DIGITAL IMAGE HISTORY First digital photography > 1957, Russell A Kirsh > drum scanner First digital camera > 1975, Steven Sasson, Kodak Electronic still cameras > captures single frame of analog video First DSLR > 1987, Steven Sasson + Robert Hills, Kodak > commercial (91), Nikon body LIGHT AND COLOURS Sight and light perception is subjective Electromagnetic spectrum infrared (< 400 nm, temperature) visible spectrum (400-700 nm wavelength) > imaginery colours (magenta, rose) ultraviolet (> 700 nm, forensics) PRINCIPLE OF IMAGE CAPTURING Light goes through the lens to the sensor/film Lens focal length (mm) > basically how wide the picture will be > fish-eye >> super-telephoto >> the smaller the number, the wider the lens aperture (f-number) > lens speed, how much light goes in, what is in focus >> the smaller the f-number, the bigger the amount of light, the smaller the depth of field 7 Shutter Speed in seconds (1/) > mechanical x electronic (global x rolling) shutter Rolling shutter can cause distortion >> the faster the time, the lower the amount of light, the more “frozen” the moving objects Image sensor > converts photons to current CCD, APS (CMOS), QIS CMOS > lower power consumption, faster, cheaper, less bleeding The larger, the higher pixel count, the lower the depth of field, the higher dynamic range, the lower the noise Crop factor > medium and large formats, full frame (size as 35 mm negative), APS-C, MTF, sensors in smartphones (noisy, pixel wars) B&W sensor > Bayer filter, Foveon X3, 3CCD Bayer filter > Kodak, GRGB filter (1/2 G, ¼ R, ¼ B), each pixel only info about its own colour >> debayering (algorithms, creating full colour pixels) 3CCD/3CMOS > sensor without Bayer filter, higher sensitivity, light divided into 3 rays by prism (hranol), each ray whole colour spectrum Image scanner > flatbed (most common, cheapest), film, virtual drum, drum (best, complicated, hard to get) DIGITAL IMAGE STORAGE Image > bitmap graphics (matrix of points) x vector graphics (lines and points described by math) Pixel > relative size, unit of resolution, PPI (pixel, per inch) > density of pixels Colour bit-depth > how many bits are used to define colour of one pixel/colour channel Bits per channel > video, raw, photoshop, display > 6, 8, 10, 12, 14, 16, 32-bit Bits per pixel > image, OS, display > 18, 24, 30, 36, 48, 64, 96-bit Dithering > colours “blending” together to create a missing one, used for low bit-depths Colour banding/posterization > issue of gradients, not enough bits to create smooth gradient 8 Additive colour mixing elementary colours > RGB model > mixing of two colours creates complementary colour anything that produces/captures light > monitor, projector, camera Subtractive colour mixing Elementary colours > CMY (cyan, magenta, yellow) > complementary colours are RGB Whenever something reflects light > print (CMY + Key = black > saves money) Colour models > differentiate in colour mixing > RGB, CMYK, HSV/HSL, Y’CBCR HSL/HSV/HSB > hue, saturation, lightness/value/brightness > Photoshop HSV Y’CBCR > Y’ (luma, brightness), CB and CR (chroma, colour) Enables chroma subsampling (f.e. 4:2:0) removing of colour info, saves data ppl better at perceiving structure than colour Gamma = gamma correction/function Camera captures light linearly, sight more sensitive in low-light Tries to simulate eye sight Colour spaces (not models!) standardised space of colour spectrum area > for monitors, TVs, printers sRGB > Microsoft, for internet and CRT monitors Adobe RGB > 52,1 % visible spectrum, more green ProPhoto RGB > Kodak, 90 %, requires 16-bit/channel Colour temperature (in Kelvin) > the smaller the number, the redder the light is 9 IMAGE FORMATS RAW > digital negative, not a format, specific for each company and their SW (Canon, Sony, Blackmagic…), includes metadata DNG > Adobe, standardised RAW, based on TIFF TIFF (Tagged image file format) > lossless, both CMYK and RGB, supports layers and alpha channel (kind of) for transparency, for photo archiving GIF (graphic interchange format) > lossless, only 8-bit, do not use PNG (portable network graphics) > lossless, alpha, animated extension APNG JPEG (joint photographic experts group) correctly JFIF > file format for images encoded by JPEG compression or EXIF (exchangeable image file format) > adding metadata NO for text and vector graphics JPEG compression Conversion to Y’CBCR Chroma subsampling (usually 4:2:0) Value shift (0-255 > -127-128) Split into blocks 8x8 DCT II (discrete cosine transform) > lossless One channel at a time > luma or single chroma Combining cosine functions Quantization > lossy Divide and round up/down according to selected quantization table Determines the final quality Zig-zag ordering > to separate 0s RLE (run length encoding) > to remove duplicate values (0s) EOB (end of block) > fill in 0s Huffman or Arithmetic coding Ultra HDR image format > by Google, uses JPEG compression JPEG 2000 > lossy and lossless, in cinematography WEBP > by Google, open standard, lossy, can have lossless alpha, supports animation, EXIF, XMP metadata 10  HEIF (high efficiency image file format) > container!!, by MPEG, non-destructive edits, animations, alpha, EXIF XMP, different codecs (HEIC, AVIF) HEIC, AVIF (open standard) Vector graphics Defined by points, lines, curves, polygons Lossless, for graphics Formats: Postscript > Adobe, can include both bitmap and vector Adobe Illustrator SVG (scalable vector graphics) > mainly web DISPLAYS CRT (cathode ray tube) > big, heavy, good colours LCD (liquid crystal display) > most common > backlight shines through dimmable colour filters TN (twisted nematic) > fast, bad viewing angles and colours > gaming IPS (in-plane switching) > slower, better angles and colours > content creation VA > similar to IPS, better black > TVs LED (light emitting diode) > not a type of panel, but light source!! > edge-lit x full-array OLED (organic light-emitting diode) > each subpixel emits light, expensive, high contrast, can be transparent Monitor characteristics Bit-depth > 6, 8, 10 FRC (frame rate control) > dithering by flickering Response (in ms) Viewing angles (170°+) Colour gamut (in % of colour space) > volume of colours which the monitor can display Contrast > range black-white in the same time Refresh rate > how often can screen change > basic 60 Hz, new TVs 120 Hz Resolution, PPI, diameter, matte/glossy, conectors!! 11 CONNECTORS VGA > analog, no sound, NO-NO DVI (digital visual interface) > digital and analog, no sound HDMI (high-definition multimedia interface) > proprietary, up to 16b colour depth DisplayPort > royalty-free > miniDP DIGITAL SOUND Sound = vibrations propagation acoustic wave Acoustics = branch of science studying sound Speed of sound > cca 340 m/s in air > the stiffer the medium/higher the temperature, the faster the propagation Frequency: 20 kHz ultrasound Binaural hearing = human hearing difference in time in which sound arrives to left/right ear we can guesstimate direction of source most sensitive perception > mid to high frequencies Wave propagation Oscillation orientation Longitudinal > oscillation on the plane of wave propagation direction Particles move back and forth in the same direction as the wave (like wheat in the wind) Transverse > oscillation perpendicular to the place of wave propagation direction Particles move up and down perpendicular (kolmě) to the direction of the wave (shaking dust off a doormat) Wave propagation > travelling x standing (wave vibrates, drumm) In phase > waves have the same amplitude at the time Opposite phase > waves cancel each other out In closed space, the sound is reflected > gets stronger/weaker 12 Huygens-Fresnel principle > sound propagates to all directions, with the same speed >> spherical wavefront Doppler effect > distortion of sound caused by movement Object moves towards me > waves compressed > higher frequency, higher pitch Object moves away from me > waves stretched > lower frequency, lower pitch TYPES OF SOUND WAVES Acoustic pressure > expresses intensity of sound power increases with square of pressure 0 dB SPL (sound pressure level) > threshold of hearing 94 (1 Pa, mic sensistivity), 120 (very loud, may pain), 130 (threshold of pain) 194 dB SPL = 1atm > loudest undistorted sound Types of dBs dB FS (full scale) > negative values, max 0 dB FS = maximum value that can be represented before clipping (distortion) occurs SOL (standard operating level) +4 dB u (professional), -10 dB V (consumer lever) Beats > “pulsing” sound when two sound waves of slightly different frequencies are played tgr Binaural beats > different sine wave in each ear 13 Audio signal = electronic representation of sound mechanical motion (kinetic) >> changes in voltage (electric energy) air movements > electricity > bits > and back electroacoustic transducer > microphone and speaker device that converts electrical energy into mechanical energy in the form of sound waves Digital sound > analog-to-digital and digital-to-analog converters ANALOG AUDIO End of 19th century > phonograph Sound recorded as a groove or electro-magnetic track > higher level of noise, noise from dirt Gramophone records, (tapes, discs) Fully mechanical, needs preamplifier (low signal) 33, 45 or 78 rpm speed > revolutions per minute Quality also depends on speed of medium (faster, better) DIGITAL AUDIO Analog signal must be digitized (converted to 1s and 0s) > AD and DA converters > minimum Continuous function (analog) to discrete values (digital) Clipping > signal too strong, AD converter cannot convert it > distorted signal, can damage dev. DAW = digital audio workstation > professional sound card and other HW > Avid Pro Tools Mono (single source), Stereo (two+ loudspeakers, channels), Quad (four channels, speakers) Phantom centre > imaginary source of sound between 2 loudspeakers Sound recorded as LPCM or PDM (audio encoding formats) LPCM Most common format of digital sound sampling Measuring the voltage X-times per sec (sampling frequency) and recording the value into Y-bits (bit-depth) > the higher the better > but not over 44,1 kHZ (2x 20kHz), yes for postproduction Bit-depth determines dynamic range and sample precision 14  PDM > Only 1-bit depth, sampling frequency 2.8+ MHz > in multiplies of CD sampling frequency, incompatible with LPCM CDDA (compact disc digital audio) > standard format for audio CDs Stereo LPCM, 44.1 kHz, 16b DVD-Audio > audio format > LPCM or MLP, Blu-Ray LOSSLESS FORMATS WAV (waveform audio file format) > Microsoft, IBM originally lossless, uncompressed formats for LPCM, limited by 4 GB AIFF (audio interchange file format) > Apple originally lossless, uncompressed formats for LPCM > AIFF-C compressed FLAC > royalty-free open audio coding format > reduces size to 50-70 %, uses RLE Monkey’s Audio > freeware, better compression than FLAC, more CPU power, X Apple ALAC > Apple LOSSY FORMATS Exploits imperfections of human hearing Psychoacoustic models > mathematical/computational models used to simulate how humans perceive sound > sensitivity, loudness… > applications > audio compression MP3 > by MPEG (moving picture experts group), open format MP3 player > 1998, 128 MB with Smart Media card AAC > mostly in H.264/265 video on YT, Netflix Profiles: HE-AAC V1/AAC+ (low bitrate), HE-AAC V2/EAAC+ (parametric stereo), EXTENDED HE-AAC VORBIS > open, BSD licence, X Apple OPUS > open, BSD licence, most efficient audio codec DTS x DOLBY Digital (surround sound, AC-3), Plus (E-AC-3, better compression), Atmos, AC-4 (7.1.4, DVB-T2) THX > certification of sound quality, by George Lucas 15 CONNECTORS AND CONNECTIONS XLR > three pin (2 wires + ground), for balanced audio (reduces noise), microphones and speakers, locking mechanism, male/female Audio Jack > several sizes, mono/stereo, for headphones, el. guitar output, speakers, can be balanced (but stereo connector becomes mono) RCA > red (right), white (left/mono), yellow (analog video), unbalanced, for S/PDIF (orange) TOSLINK > optical, black, for S/PDIF S/PDIF = Sony/Philips Digital Interface Format Digital audio interface standard for transmitting stereo audio signals between audio devices > high quality audio transfer (home cinema) RCA/TOSLINK, Dolby Digital/DTS AES3 > “professional” S/PDIF, interface standard with timecode, in broadcasting over XLR, BNC (connector) or IP network DANTE > interface standard, by Audinate, proprietary Uncompressed, multi-channel, low-latency > broadcasting, live events Over IP network, Ethernet for transmission (uses physical cables) MICROPHONES AND SPEAKERS Microphones Analog devices, output is analog signal Dynamic (moving coil mic., no power) x condenser (requires phantom power, better) Usually XLR or Audio Jack Polar patterns > in which direction is the mic the most sensitive Handheld, Lavalier, Shotgun, Studio Mic Speakers Usually membrane with electromagnet Subwoofers (low-frequency sounds) x tweeters (high-frequency sounds) Active (needs electricity, better) x passive (no power needed) Studio speakers = near field monitors (usually active) x home cinema (usually passive) 16 AUDIO MEDIA FM RADIO > FM = frequency modulation, analog DAB (digital audio broadcasting) > newer DAB+, uses HE-AAC v2 Podcasts and RSS (rich site summary) > regularly released programme on Internet, RSS written in XML (enables following of news websites, blogs, podcasts) Streaming > buffer (temporary data storage, better stability), problematic capacity of content providers DIGITAL VIDEO Film/video = sequence of images creating illusion of motion Film = motion picture Used horizontally, in IMAX vertically > grain, digitalization by scanning Video = electronic medium > YouTube, Netflix… TV = medium broadcasting video Framerate = number of frames per second > more, better Film 24, PAL TV 25/50 Frame = whole picture at the time Field = only either odd or even lines Recording methods, how frames are created and displayed Progressive (P) > nowadays, shows full image Interlaced (I) > 2 fields, shown sequentially to complete a full frame (in older broadcasting, flickering lines Progressive segmented frame (PsF) > compromise 3:2 pull down > to convert film shot at 24 fps into format suitable for TV/video displays (use 30 fps) > film slowed, 4 frames >> 5 HFR (high frame rate) > faster than 24 fps standard > usually 48 fps (Hobbit) Shutter speed = time of exposure of single frame Defines smoothness of motion Expressed as an angle > how much of rotary disc shutter is open > common 180° < 360° >> shutter partially closed, exposure time shorter than frame rate duration 17 Aspect ratio > movie industry > different; TV and digital video > 16:9 Bit-depth > before HDR 8b, 10b in professional environment > now 8-12b limited range > TV doesn’t use all 8b of values SDR (standard dynamic range) > resolution does not matter maximum brightness 100 cd/m2, 8b > 6 stops, 10b > 10 stops HDR (high dynamic range) > better contrast between the darkest and brightest colours Higher brightness than 100 cd/m2, >= 10b x HD (high definition) > refers to resolution, typically 1920x1080 x UHD (ultra high definition) Standards Rec. 709 > standard specifying SDR HDTV 1920x1080, 23.976-60 fps, 8b/10b colour depth, 35 % of all colours, no HDR Rec. 2020 > standard specifying SDR UHDTV with WCG (wide colour gamut) 3840x2160 and 7680x4320, up to 120 fps, 10b/12b, 76 % of colours, no HDR Rec. 2100 > standard adding HDR to Rec. 2020 Adds resolution 1920x1080, new proprietary colour space (from Dolby) EOTF and OETF = electro-optical transfer function How to transfer digital data to light > specifically in relation to brightness and colour How a display turns el. signals into visible images (for correctly looking HDR) 2 EOTF systems (standards) PQ = perceptual quantizer > provider higher precision for HDR content By Dolby > for standards HDR10(+), Dolby Vision > in 4K UHD Blu-Rays HLG = hybrid log-gamma > for broadcast and live transmission, open standard by BBC+NHK, compatible with both HDR and SDR displays, no metadata OETF the same but for cameras Standards compatibility > complicated, HDR10 most compatible 18 3D Video > stereoscopic, not really 3D In past anaglyph glasses, nowadays polarized or active shutter glasses Fist 3D movie > Ghosts of the Abyss (Cameron) Distribution for download > SbS (side by side, vertical) x TaB (top and bottom, horizontal) HDCP (high-bandwidth digital content protection) > against piracy, stupid COMPRESSION PRINCIPLES Lossy video compression Intra compression > compression within a single frame Prediction > transformation > quantisation > entropy coding Inter compression > across frames > only temporal changes recorded motion vectors > pieces which moved since last keyframe, recorded as vector > difference is encoded (prediction error) Other tools Slices > frame sliced into multiple independent slices > to improve encoding efficiency Tiles > frame is divided into independent horizontal rectangular regions Deblocking filter > removal of blocking artifacts (at hard edges or visible lines) Coding tree > frame is divided into blocks SAO (sample adaptive offset filter) > improves single colour areas and edges Types of frames I-frame (intra) > intra compression Keyframe (complete frame, serves as reference point) Anchor frame (reference frame used to encode subsequent frames) P-frame (predicted) > only changes in comparison to previous frame intra + inter compression, anchor frame B-frame (bidirectional) > only changes in comp. to previous/subsequent frame Larger buffer, increases latency GOP (group of pictures) > frames between two I-frames, defined by 2 numbers M > distance between anchor frames N > size of GOP, distance between keyframes 19 Video coding formats and standards Profiles > set of predefined parameters/features within a video codec How the video should be encoded/decoded Help ensure compatibility across different devices Levels > set max resolution and bitrate > help ensure compatibility Professional (intermediate) coding formats Apple ProRes > proprietary lossy format, only I-frames,.mov AVID DNXHD/HR > only I-frames ProRes RAW > 12+ bit depth, metadata Blackmagic RAW > open, but not open-source, partial debayering in camera Delivery coding formats MPEG, VCEG MPEG-1 Part 2 (VHS quality) MPEG-2 Part 2 (DVB-T/S, DVD, Blu-ray) MPEG-4 Part 10 (DVB-T/T2/S2, Blu-ray, YT, Netflix, iVysílání) MPEG-H Part 2 (DVB-T2, UHD Blu-ray, Ntflx) > high efficiency video coding MPEG-I Part 3 > versatile video coding VPX family > VP8, VP9, AV1 VP8 > used by Wikipedia VP9 > HDR, WCG (wide colour gamut), transparency, used YT, Ntflx, Wiki AV1 > HDR, WCG, transparency, YT, Ntflx, Vimeo, twitch, Meta Encoding/decoding video SW en/decoder x Hybrid en/decoder x HW (fixed-function) en/decoder SW encoder Uses CPU, available almost immediately, better efficiency than HW encoder Uses instruction sets (and multi-threading) > specialized commands to speed up the process Not suitable for real-time encoding, not for devices with battery/slow CPUs 20  HW encoder (= fixed-function) Uses ASIC (application-specific integrated circuit) > for better efficiency Not available immediately, for better quality encodes new HW necessary Adobe encoder SW based encoder but can use HW acceleration > CPU and GPU Quick Sync > HW based (CPU >> GPU), fast encoding video, bad quality Interfaces SDI (serial digital interface) in broadcasting, up to 10b, 4:2:2, single direction SMPTE ST 2110 standards set of standards defining transmittion of audio/video over IP Requires PTP (precision time protocol) > precise time sync. NDI, NDI HX, HX2, HX3 (network device interface) lossy compression Video distribution TV DVD (digital versatile disc) 9 regions (continents + travelling, space) Blu-ray, UHD Blu-ray blue laser, 3 regions Streaming over HTTP (hypertext transfer protocol) to viewer > YT, Ntflx, FB live-streaming to VOD service (video on demand) > Ntflx, Disney+, HBO Digital cinema DCP-o-matic > open source software for creating DCPs (digital cinema packages) DCP > standard format for digital cinema projection 21 Television History Mechanical TV (2nd half of 19th c. >> 30s) Grayscale analog TV (end of 19th c. >> 70s) Colour analog TV (50s >> late 2010s) Colour analog HDTV (Japan, 70s >> 90s) Digital SDTV (80s >> now) Digital HDTV (90s >> now) Digital UHDTV-1 (00s >> now) Digital UHDTV-2 (2010s >> future) Broadcast TV Terrestrial 4 main digital TV standards in the world > DVB (almost all), ATSC (N. Am.), ISDB (Jap, S Am.), DTMB (China, Cuba…) DVB DVB project, doesn’t specify codec but physical layer (signal distribution) -T (terrestrial), -S (satellite), -C (cable) DVB-T2 (CZ, in 2017) Multiplexing > multiple channels in single stream HbbTV > hybrid TV > internet services in TV (iVysílání) Teletext, EPG (electronic program guide) Specification in D-Book by ČTÚ (český telekomunikační úřad) 22 LINUX, SHELL AND BASH & FFMPEG Shell = a user interface of OS > GUI (graphic user interface), TUI (text), CLI (command line) Linux > Bash (most common shell in Linux), Windows > PowerShell (Terminal) SSH > secure shell protocol, for remote connection to computer Commands Echo > print text Cat > print content of file Touch > create new empty file Ls > list Cd > change directory ~ (tilde) > home directory Nano > text editor Root > super user > sudo Bash script > start with shebang (#!/bin/bash) FFMPEG = open source SW for multimedia encoding/decoding/editing Compromise between file size, quality, speed Libraries > components with special functions for en(de)coding/format conversion Audio encoding: libopus, libvorbis, libfdk_aac Video encoding: libx264, libvpx-vp9, libaom-av1 3 binaries > ffmpeg (encoding etc), ffplay (player), ffprobe (file info) Presets Codec selection > -c:a/v + library Bitrate > -b:a/v Multiple passes > for video > improve quality and efficiency of encoding 23