presenter notes Image credit: Mary Kidd Before the rise of digital technologies, both professionals and everyday consumers relied on analog recording and playback devices. These worked by capturing information using what we call analog signals. What do we mean by "analog signal"? When you hear "analog", imagine the waves and ripples made by a pebble or a flower dropped into a lake. The resulting waves rise up and down in a smooth, continuous pattern. The water's undulations are a the physical result of the force of the rock hitting the water, and displacing the water by forcing it outward. The water is therefore reacting in an "analogous" fashion. Meaning, if we were to throw a larger rock into the water, you'd get larger waves, and if you were to throw smaller rocks you would get smaller waves.

presenter notes Now let's pretend we've taken a rock, and instead, throw it at the ground instead of in the water. If you do this, the rock will make a sound. Similar to waves on a pond, the sound the rock makes is a result of waves traveling invisibly through the air. Your inner ear anatomy pick up these waves in the form of vibrations, and those vibrations are translated by your brain into discernible sound.

presenter notes Similar to sound, light also travels in the form of waves. Different colors have different wave forms. For example, red light waves have long, infrequent waveforms, where violet light is short and frequent. Your eyes pick up on and react to these waves, and translate them into your vision.

Presenter notes Different sounds and images form different shaped waves. The wave's characteristics dictate what you hear. On screen is a diagram of a waveform, showing various characteristics like peaks and troughs, and cycles through time. The combination of characteristics is what produces a specific sound type, quality and tone, like high- or low-pitch, loud or quiet, a certain or combination of notes

presenter notes The tinfoil phonograph was one of the first devices ever made that allowed for both recording AND playback. You may have noticed that this used no electricity: just hand-cranking. Fast-forward nearly 100 years later, and you get more advanced recording and playback devices, including the vinyl record. Here, we can see a microscopic image of the surface of a vinyl LP. Here, you can see the sound waves literally etched into the plastic. The physical imprint made by Edison's needle onto tinfoil, or onto plastic, are both analagous to the waves produced by the things making the sound, whether you're shouting into a mouthpiece or a band playng music in a recording studio. This is why we call recording devices "analog": it is because the way they record information about the world, whether that's sound or light or color, is analagous to the waves occurring (or occurred) in the world.

presenter notes Record grooves being analagous to "real life" sound waves make tools like IRENE work. IRENE is a high-powered image scanner than can photograph the surface of a disc, and translate the image directly to sound. IRENE is used for records too fragile or broken to be digitized using normal means.

Analog signals "look" different from digital signals. Compared to an analog signal, which is a continuous wave, digital information looks more like buildings on a city block, or a staircase. Note, this is an abstraction of how signals actually look! This is because when you record something using a digital recording device, it takes samples at intervals, and each sample is represented by a binary digit (1 or 0). The higher the sample rate, the more binary data. Keep this in mind as we go through how analog recording works. We are going to use the example of a 1940s analog video camera. Image from: https://ez.analog.com/ez-blogs/b/engineerzone-spotlight/posts/accurately-measuring-our-analog-world?ADICID=ARWDEMEAP328165eBookAD7134-Blog

presenter notes Analog cameras use an analog recording technology that, unlike a record player, feels a bit more "hidden" since it all takes place inside of a dark metal or plastic case. Here, our black-and-white analog camera is capturing a flickering candle. The camera contains a contained tube composed of, at the front, a lens. When capturing a scene, the camera lens focused the image onto a metallic plate positioned near the front of the tube. This plate was coated with a photosensitive substance like selenium. Photosensitivity would cause the surface of the plate to react to the light. You can think of this plate as similar to a Polaroid photograph, but constantly reacting instead of producing a static image. The photosensitive plate would be scanned by a device called an electron gun that reads the varying photosensitivity imprinted on the plate. This beam moved meticulously from left to right and top to bottom, covering the entire surface of the plate. You can think of the electron gun working like a flatbed scanner moving beneath a flat image, except at a much higher speed. Each scan from left to right was incredibly swift, taking only about 63 millionths of a second. After completing a horizontal scan, the beam swiftly snapped back to the left to begin scanning the next line. As the electron beam traversed the plate, it translated the varying brightness and darkness of the live scene into a continuously fluctuating voltage. The brighter a specific area of the scene, the higher the voltage produced by the camera tube. This voltage variation resulted in the creation of an analog signal, encoded within an electromagnetic wave. Essentially, the peaks and troughs of this wave corresponded to the voltage fluctuations, which in turn reflected the brightness and darkness of the original live scene. So in many ways it is similar to how grooves mimic the shape of sound waves, except for cameras, it's analogous to levels of light and dark.

presenter notes In a color camera, the fundamental recording mechanisms remain largely similar to those of a black-and-white camera. We still have the lens focusing the live scene onto a photosensitive plate and the electron beam scanning this plate to convert light into voltage. However, in a color system, we introduce additional components to capture the full spectrum of colors. In addition to one photosensitive plate capturing darkness/brightness, color cameras utilize multiple photosensitive plates, each coated with a different filter to capture specific colors. These filters typically correspond to the primary colors: red, green, and blue (RGB). As the live scene passes through the lens, it is split into its constituent colors by a prism or a set of filters. Each color component is then directed onto its respective photosensitive plate. For example, red light is captured by the plate with the red filter, blue light by the plate with the blue filter, and green light by the plate with the green filter. Similarly to the black-and-white camera, the electron beams scan each photosensitive plate, converting the varying levels of light intensity into corresponding voltages. However, in this case, we have multiple sets of voltage signals, one for each color channel, and one for light/darkness. The amplitude of the voltage signals generated by each photosensitive plate corresponds to the saturation of the respective color in the scene. Higher saturation levels result in higher voltage outputs, while lower saturation levels yield lower voltage outputs. By combining the voltage signals from all the color channels, we obtain a complete representation of the color image. This combined analog signal, encoded within an electromagnetic wave, carries information to recreate the full-color scene during playback or broadcast.

presenter notes The voltage fluctuations representing both the colors and brightness levels captured by the camera are then transmitted, either through wired connections or broadcast networks, to a display device known as a cathode ray tube, or CRT monitor. Once received by the CRT monitor, the transmitted voltage signals were interpreted and utilized to reconstruct the original image. The CRT monitor reads this information and employs electron beams within the tube to illuminate phosphor-coated pixels on the screen. These illuminated pixels combine to form a composite moving image, faithfully recreating the scene that was initially captured by the camera. It's a remarkable process where the intricate interplay of voltage fluctuations translates into a coherent and dynamic visual display, allowing viewers to experience the captured images and scenes in real-time.

presenter notes Inside a CRT or television, there's a crucial component known as the picture tube, housing devices called electron guns. These electron guns play a pivotal role in translating the voltage signals received from the camera into a moving image on the screen. Each electron gun corresponds to a specific aspect of the signal outputs transmitted to the CRT. For instance, there's an electron gun for the red information, another for the blue information, and yet another for the brightness information, and so forth. As the camera input signals are received, they emit a constant stream of electrons. The density of these electrons at various points corresponds to the peaks and troughs of voltage, containing information about the brightness and hues represented in a scene. The electron guns within the CRT push and accelerate these electron streams toward the front of the screen in a scanning pattern. This scanning pattern aligns with the format of the transmitted signal, ensuring that each electron beam accurately represents its corresponding aspect of the image. Simultaneously, rapid scanning occurs from left to right and top to bottom of the screen. This rapid scanning, synchronized with the signal inputs, is how TV screens 'trick' us into perceiving a moving image through time on the screen. So how does the scanner know where to start and stop a line of information on a screen? Gif of electron stream: https://commons.wikimedia.org/wiki/File:Electronbeamshiftingeffect.gif

presenter notes The video signal is encoded with information about the start and stop points of each of the horizontal lines being scanned, as well as information about the first scanline and the last scanline. This encoding allows for the precise synchronization of the scanning process, ensuring that the image is displayed correctly and seamlessly. In this way, not only color and brightness information but also information about frames is encoded and displayed on the screen, creating a cohesive moving image for viewers to observe. So how exactly does a TV know where to “place” a bit of color or brightness information on a screen? Let's break down how each part of the signal is read and rendered, akin to reading sentences in a book: Active Video Signal: The orange squiggle line is the active video signal. This is akin to the words on a page. It appears as a series of waves occurring over time on the horizontal axis, with varying heights (voltage) on the Y-axis indicating brightness levels. Front Porch: As we approach the end of one line of video, the signal flattens and dips, resembling a precipice. This flat part, known as the front porch, serves as a signal that a horizontal blanking period is beginning. It's like a stop sign on the road, instructing the electron beam to briefly pause. Horizontal Sync Pulse: Following the front porch, we encounter the horizontal sync pulse. This pulse signals the repositioning of the electron beam from the end of one line to the beginning of the next. This repositioning is necessary to ensure the electron beam returns to the left side of the screen before scanning the next line. Blanking during this period prevents unwanted visual artifacts or distortions. Back Porch: After the sync pulse, we encounter the back porch, indicating that the electron beam is now in position to begin scanning the next line. Color Burst Signal: Lastly, we have the color burst signal. This signal prepares the new line to synchronize the video signal with the color signals, ensuring accurate color reproduction. This pattern of video signal, front porch, sync pulse, back porch, and color burst repeats for each and every line of every frame of a video. It's through the careful interpretation of these signals that a TV or display device renders the analog signal into a coherent and visually accurate moving image for viewers to observe.

presenter notes Once we reach the end of all the lines on a single frames-worth of video, we have another set of synchronization signals that work to signify when we’ve reached the end, that we need to blank the signal briefly, and return our electron beam to the top of the screen to start over again.

presenter notes Let's break down how video signals are recorded to tape, using VHS as an example: Just like with live broadcasting, a handheld VHS recorder captures a live scene using a lens. The lens focuses the image onto an image sensor, typically a charge-coupled device (CCD) containing tiny light-sensitive diodes called photosites. These photosites measure light intensity and color, translating this information into an electrical signal. Video Processing Circuit: The electrical signal from the image sensor passes through a video processing circuit. This circuit adjusts signal levels, enhances quality, and removes unwanted noise or distortion, ensuring a clean and polished video signal. Recording Head and Magnetic Tape: The polished video signal is then sent to a recording head. This recording head responds to variations in voltage by producing an electromagnetic current. A magnetic-sensitive tape, housed within a VHS cassette, is passed under the surface of the recording head. Imprinting Voltage Fluctuations onto Tape: As the tape moves past the recording head, the electromagnetic current from the head imprints variations in voltage onto the tape. This process repositions the magnetic particles bound to the tape's surface, creating a magnetic pattern that represents the encoded video signal. Overall, the process involves translating the electrical signal representing the live scene into magnetic fluctuations on the tape's surface. This recorded signal can then be played back or broadcasted by passing the tape through a playback device, where the magnetic fluctuations are converted back into electrical signals and ultimately displayed on a screen for viewers to watch. VHS cassette image from: https://publicdomainvectors.org/en/free-clipart/Video-cassette/40440.html

presenter notes The result of writing the video signal to tape was actually a series of tracks containing different “streams” of information. On a typical VHS tape, the audio is recorded in its own track, while video would be recorded on another. Video information would be recorded in a diagonal fashion, using a technique called "helical scan", which involved moving the tape past the recording head in a diagonal direction. This allowed for longer recording times and the ability to fit more information onto the tape. In addition to a separate audio and video track, was a third track, known as the timecode track. This track consisted of information that contains timecode signals that represented hours, minutes, seconds, and frames. This timecode track is used by playback devices to know when the beginning of a diagonal line of picture information begins or ends, or when a frame begins or ends.

Audio/visual preservation labs or transfer stations as they are sometimes called come in all configurations, shapes and sizes. This drawing I did in 2017 of a transfer rack gives you a general sense of the minimum viable requirements for a transfer station. A. Playback deck does exactly what its name implies: it plays back tapes. B. Monitor and scopes: Used to look at and analyze the analog video and audio signals coming from the playback deck. C. Time-base corrector: A device that ensures the lines that comprise the video signal are in order. D. Analog-to-digital (A to D) converter: Converts the analog signal from the playback deck into a digital signal. E. Metal frame that holds everything together

presenter notes We will look specifically at how the DC Public Memory Lab is set up to transfer VHS tapes. This diagram comes directly from their public documentation, and gives us a sense not only of the devices and software used, but also how things are wired to relay the original video signal to a computer in digital format. Starting from the left, we see that the VHS deck is wired to three components, a CRT monitor, and a time-base corrector, or TBC for short, and an analog-to-digital or A to D converter. We learned earlier that CRT monitor stands for “Cathode Ray Tube” monitor (it’s also fine to just call it “the monitor” or “the television” because it looks like an old TV). CRT monitors in transfer lab setups are used to view the original analog image BEFORE the signal goes into other parts of the transfer workflow. The monitor can help you understand whether or not a signal is even coming out of the VHS deck, and is kind of the first “test” for your transfer to pass. If it doesn’t show up on the CRT monitor, that may indicate something is wrong with your playback deck (or your monitor).

presenter notes Photo of a transfer station with multiple CRTs Image from <a href="https://parkslibrarypreservation.wordpress.com/2019/05/01/how-to-build-a-video-preservation-rack-for-in-house-digitization/" target="_blank">How to Build a Video Preservation Rack for In-House Digitization</a>

presenter notes One thing that is not a part of the DC Memory Lab setup that is worth talking about are scopes. As we run an analog signal through a transfer setup like this, one way to “see” the waveform is to run the signal through various scopes. Scopes can give you an actual visualization of varying voltages, coming through on a waveform, to give us clues whether or not we are capturing it properly. The scope you see here is known as a vectorscope, that allows us to see chrominance – or color – information. A vectorscope shows us a circular display. The center of the circle corresponds to 0 color, while the rest of the circle represents all possible colors. The lines and dots on the screen represent hues. Hue corresponds to each line’s angle, while saturation is represented by the distance of a dot from the center of the circle. The display also contains information about a target range. So visually if you see the lines or dots exceeding these targets, then that may indicate a color imbalance. The knobs to the right of the vectorscope can be used to make corrective adjustments to the signal. You can either do that during playback capture, or just note this information and correct it post-capture.

presenter notes Another common type of scope is known as a waveform monitor, which gives us information luminance, or brightness/darkness levels. It displays the video signal as a graph where the horizontal axis represents time, and the vertical axis represents the brightness levels. The darkest parts of the image are represented at the bottom of the graph, and the brightest parts are at the top. You can use the buttons or knobs to adjust the brightness or darkness by trying to “fit” the waveform in between the scale at the left, so it falls within a target range.

presenter notes The VHS deck is wired to a device known as a time-based corrector or TBC. A TBC is used to correct errors in the timing of analog video signals caused by variations in tape speed or aging of playback equipment. The TBC uses a stable reference signal to retime the incoming video signal and stabilizes the synchronization signals, ensuring the image is displayed correctly without distortion. It is a critical component for ensuring high-quality and accurate analog to digital video transfers.

presenter notes This is what one type of TBC looks like. Very unassuming... yet powerful! TBCs can be purchased as an external device (as they did for the DC Memory Lab) but sometimes can come pre-installed within certain playback decks.

presenter notes Image credit: Duke University blog post Digitization Details: Thunderbolts, Waveforms & Black Magic: https://blogs.library.duke.edu/bitstreams/2014/06/19/digitization-details-thunderbolts-waveforms-black-magic/ Notice how the VHS playback deck is wired via the time-base corrector to the Analog-to-Digital converter, while the VHS audio signal is routed directly to the A-to-D. We will talk about why this is in a second. But first, let’s talk about what the A-to-D is. The A-to-D does exactly what its name suggests: it converts the analog signal into a digital signal, which can ultimately be read and turned into a digital file on a computer (and as you can see, the next stop after the A to D is a computer). So why is the VHS playback deck wired to the A-D via the TBC, but the audio is routed separately? This is done to improve the quality of the video signal captured. TBC’s sometimes introduce a slight delay in the video signal while its doing its “correcting”, which can cause the audio and video to become out of sync if they are passed through the TBC together. To avoid this issue, the audio signal can be routed directly to the A to D device, bypassing the TBC. This ensures that the audio and video remain in sync, while still allowing the video signal to be corrected by the TBC.

presenter notes Lastly, let’s talk about the computer, which represents the end of the road for the video signal, but the beginning of everything else in terms of digital preservation workflows. In the DC Memory Lab setup, the computer comes with a number of different software programs installed. With any computer that is being used to capture a digital signal and turn it into a digital file, you need to use some sort of capture software. DC Memory Lab uses BlackMagic Media Express. There are many types of software that can be used for a setup. One tool I have used recently is known as vrecord. This is a screen capture of what you see when you open up record. Here, you establish settings to do with which folder you’re placing the final file, the file format, the codec, bit depth, whether to generate frame-level MD5 checksums, metadata XML outputs, and other settings. Once you click the “OK” button, you would then press play on your VHS or other playback deck, you would first run what’s known as a “pass through” to just test whether or not the signal coming out of the video is looking and sounding ok. Once you’re done checking, you would rewind the tape back to the beginning, press play on the deck, press ok in vrecord, and run the capture.

presenter notes The azimuth angle refers to the angle at which the tape head is oriented in relation to the tape itself. Proper alignment of the azimuth is essential to ensure optimal capture of a tape and to avoid degradation of the signal quality. Azimuth misalignment in a tape occurs when the angle of the tape head does not precisely match the angle of the recorded tracks on the tape. If the azimuth angle is not set correctly, the tape head may not fully capture the audio signal, resulting in increased noise, and other issues. Azimuth misalignment can occur due to a variety of factors, including wear and tear on the tape head, improper calibration of the playback equipment, or incorrect handling of the tape. When azimuth misalignment is suspected, the engineer can fix it by either taking a screwdriver and screw or unscrew the head to adjust its angle in relation to the tape. Interestingly enough, if a tape was recorded with a misaligned azimuth to begin with, the engineer may need to adjust the tape head azimuth “incorrectly” in order to capture as much of the magnetic-imprinted signal as possible.

presenter notes Sticky shed syndrome happens when the binder that “glues” the magnetic coating of a tape to the plastic, which carries the signal, has a chemical reaction to the moisture in the air. This causes the coating to flake off the tape, leaving behind deposits on the mechanics of the playback machine. Sticky shed can be recognized by a telltale high-pitched “squealing” that occurs during playback. Tapes that have sticky shed can be treated by placing them into a special oven. By gently heating the tape at a controlled temperature, it can encourage the substrate to re-glue itself back to the tape. However, tape baking only works enough to get one more good pass, before the binder starts to react again with the moisture in the air. You can read more about magnetic tape degradation in Richard Hess’ paper “Tape Degradation Factors and Challenges in Predicting Tape Life” (https://www.richardhess.com/tape/history/HESSTapeDegradationARSCJournal39-2.pdf)

presenter notes Much of what I am about to talk about next is directly derived/remixed from Ashley Blewer’s audio/visual training website (https://training.ashleyblewer.com/). This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License which allows me to do this, contingent on me giving credit where credit istdt due. I highly recommend you check this site out and if you are interested/have the capacity, step through it in its entirety or through one or more of the independent study paths to do with digital or audio/visual preservation. Ashley is a brilliant teacher, and has made tutorials readily available for anyone, free of charge.

presenter notes What does a media file consist of? In Ashley’s tutorial, she explains a video file’s contents “made up of streams, wrapped in a container, and some of those streams are encoded with a codec.” Let’s talk about what she means by streams, containers, and codecs.

presenter notes Media files at first may just seem like any file. However, they're slightly complex, in the sense that they are multi-layered. A good way to think about a media file is a kind of bucket that holds things. The technical term for this is a "container" or a "wrapper". In order to tell what type of container a media file is, all you need to do is look at the file extension at the end of the file: .mp4 means we are looking at an MP4 container, for example. The media container contains what are known as "streams", which are the files that live within the container. Depending on the container, you can have multiple different streams. A simple way to look up what streams can live within a container, you can look at its Wiki. For example, for MP4s: https://en.wikipedia.org/wiki/MP4_file_format On the page there should be a "data streams" section: https://en.wikipedia.org/wiki/MP4_file_format#Data_streams Here you can browse the streams. See: https://en.wikipedia.org/wiki/Comparisonofvideocontainerformats

presenter notes Within the data streams section of a media container you'll notice it mentions "codecs". A codec refers to how a data stream may be encoded, or decoded (the word codec is a portmanteau of the words "encoded" and "decoded"). The purpose of a codec is to make files more manageable, in terms of the amount of storage space they use on servers, the amount of bandwidth required to transmit them over networks, all the while retaining a certain quality that can either be discernable or not, which depends in part on whether the codec applied uses a lossy or lossless compression algorithm.

presenter notes Lossy compression is a method of reducing the size of a digital file by discarding some of the original data. This type of compression works by removing information that is deemed less important or perceptually irrelevant, in order to achieve higher levels of compression. The bad part about lossy compression is its impossible to “unsmush” a lossy compressed file: once the information is lost, it is lost forever. Lossy compression could be appropriate for service copies, while lossless compression or uncompressed files are more appropriate for preservation masters and mezzanine files. Determining what approach to use is contingent on your institution’s specifications.

presenter notes MP3, which stands for “MPEG-1 Audio Layer 3” is a type of lossy codec.

presenter notes Why use MP3? MP3s arose as a way to transmit data over networks quickly.

presenter notes In digital preservation, certain containers and codecs are used in order to ensure that we are capturing signals optimally, balancing that of storage and budget constraints. The screen capture shows NYPL’s specifications for analog video cassettes. They use the Matroska or MKV container or wrapper. The MKV files can contain a video stream, which uses the FFV1 codec, and an audio stream, that uses the FLAC codec for the audio stream. Both codecs use lossless compression encoding technology.

presenter notes Lossless compression is a method of reducing the size of a digital file without losing any data or quality. This is achieved by identifying and eliminating redundancies and patterns within the data. Some techniques include: Dictionary-based compression: This technique involves creating a dictionary of frequently used data patterns in the data to be compressed. The dictionary is then used to replace the repeated patterns with shorter codes, resulting in a more efficient representation of the data. Run-length encoding: This technique is used for data that contains long runs of the same value or repeating patterns. It replaces these runs with a shorter code that represents the value and the length of the run. Huffman coding: This technique involves assigning shorter codes to more frequently occurring data values, and longer codes to less frequent ones. This results in a more efficient representation of the data.

Original Sentences

Predictive Coding Example