Cool Technical Terms
By, Amir Majidimehr
I thought I create a thread with industry terms that you could throw around to impress your friends and people less than fortunate who don’t read these articles
.
1. Coding noise. In signal processing circles, we use the term “noise” to describe distortion. For example when we compress audio in a lossy manner such as MP3 the artifacts are called “noise.” The more formal term is coding noise.
2. Quantization noise. Building on above we now know that we are talking about some type of distortion. Quantization is the process of taking a continuous (analog) signal and describing it in discreet steps such as 16 bits for CD audio or 8 bits per component for video data. Since we are using steps, then anything in between is not represented in the digital value and hence, can become a form of distortion we call quantization noise. The less bits you use, the higher your quantization noise. If you ever see graduated colors such as the sky in video having contouring lines in them, that is quantization noise. There are not enough steps to properly show the changing levels.
3. Motion Comp. This is short for Motion Compensation and describes how video compressions in use today work. They divide the screen into squares or rectangles and track their movements. Instead of retransmitting the block again, we instruct the receiver to move the block it already has and hence save a lot of bandwidth or storage. This is called Motion Estimation. The receiver then uses the opposite – motion compensation – to move the blocks to where they need to go. MPEG-2, MPEG-4 AVC and VC-1 (video codecs in Blu-ray) are all motion compensated codecs. Motion-JPEG (M-JPEG) used in some security cameras is not.
4. Perf. This is a computer science shorthand for performance. “What kind of perf do you get with that computer” means how fast does it run.
5. Repro. Next time you talk to a computer support guy use this term as shorthand for “reproduce” and he will take you far more seriously! For example, if you can’t recreate a problem that happened a while ago but all is well now say, “I can’t repro it!” We use this term in software development circles. So the automatic assumption is that you are a computer programmer so don’t take it too far or you will get yourself in trouble!
6. Perceptual model. This means a system for defining how the ear or the eye perceive quality. For example, an MP3 codec uses a perceptual model of the human hearing system to determine what it can throw out what you might not miss. Perceptual models are also used for things like surround sound using headphones.
7. Gain. In engineering circles we talk about amplifier gain as a measure of how much the device is increasing (or shrinking) the input value. Think of it as “volume control” but used in a more precise way.
8. Buffer. In audio/video world, this is an amplifier that has unity gain. Remember per above, gain is amplification so unity gain means no amplification. Buffer’s job is usually to change other characteristics of the signal (e.g. its impedance).
In computer science, buffer is just a piece of memory that holds data temporarily. When a program reads a file from disc, it “buffers” it to memory. When you stream video and it takes a bit to start displaying it, the player is buffering or fetching some of it first so that it has less chance of running out of view to play.
9. Transfer function. This is a term that is used to describe the relationship of the output of a device related to its input. The transfer function of an ideal power amplifier looks like a straight line if you graphed it on an X/Y chart. An amplifier buffer has a line that is at 45 degrees.
10. Cache. Computers store their programs and data in memory. To get anything done then, the CPU which is the engine in your computer, has to keep accessing the memory. Alas, CPUs are much faster than regular memory. A cache (pronounced the same as cash) is higher speed memory that sits between the CPU and main memory. It is usually a lot smaller but holds the most frequently accessed data and thereby speeds up most operations that need access to data. We could have multiple layers of cache each become faster and smaller to save costs in having one large but expensive cache. Today's CPUs for example include three levels of caching before data has to be accessed in main memory.
As another example, the computer memory in turn is used as a “cache” for the hard disk (or even SSD). When you run a program for the first time, you see your computer disk light blink a lot. Quit out of the program and start it again and you see much less disk activity. That is because the computer memory acts as a cache and contains that portion of hard disk that held your program in memory temporarily, giving access to it at a much faster rate as computer memory is orders of magnitude faster than hard disk. This is one of the reasons why having more memory makes your computer run faster even though the only thing that does any work in your computer is the CPU, not the memory.
11. SSD, SLC and MLC. SSD stands for Solid State Disk. It uses flash memory to replace a (smaller) hard disk. Because it has no moving parts, it can have much faster performance than had disks. There are two types of flash memory: SLC which stands for single-level cell. And MLC which as you might have guessed stands for multi-level cell.
The SLC type of flash stores one bit per flash memory cell. MLC can hold 1 to 3 bits. The latter means 8 different values stored in one location. As a result, MLC flash is denser and far cheaper than SLC. Unfortunately it is also slower and much less reliable internally. Error correction is used so that as a user, you don’t notice this aspect other than shorter lifetime. And multiple units can be strung together to improve their speed.
One of the best ways to speed your system is to switch from hard disk to a high-performance SSD.
12. “IP control.” IP is the lowest level communication language used on the Internet and in your home network. When we use the term though, as in “does your lighting system have IP control” we mean whether the device can be controlled using your home network as opposed to needing a physical connection. Today, most devices are controlled using the latter. Hope is that in the future all devices are “IP controlled.”
By, Amir Majidimehr
I thought I create a thread with industry terms that you could throw around to impress your friends and people less than fortunate who don’t read these articles
.1. Coding noise. In signal processing circles, we use the term “noise” to describe distortion. For example when we compress audio in a lossy manner such as MP3 the artifacts are called “noise.” The more formal term is coding noise.
2. Quantization noise. Building on above we now know that we are talking about some type of distortion. Quantization is the process of taking a continuous (analog) signal and describing it in discreet steps such as 16 bits for CD audio or 8 bits per component for video data. Since we are using steps, then anything in between is not represented in the digital value and hence, can become a form of distortion we call quantization noise. The less bits you use, the higher your quantization noise. If you ever see graduated colors such as the sky in video having contouring lines in them, that is quantization noise. There are not enough steps to properly show the changing levels.
3. Motion Comp. This is short for Motion Compensation and describes how video compressions in use today work. They divide the screen into squares or rectangles and track their movements. Instead of retransmitting the block again, we instruct the receiver to move the block it already has and hence save a lot of bandwidth or storage. This is called Motion Estimation. The receiver then uses the opposite – motion compensation – to move the blocks to where they need to go. MPEG-2, MPEG-4 AVC and VC-1 (video codecs in Blu-ray) are all motion compensated codecs. Motion-JPEG (M-JPEG) used in some security cameras is not.
4. Perf. This is a computer science shorthand for performance. “What kind of perf do you get with that computer” means how fast does it run.
5. Repro. Next time you talk to a computer support guy use this term as shorthand for “reproduce” and he will take you far more seriously! For example, if you can’t recreate a problem that happened a while ago but all is well now say, “I can’t repro it!” We use this term in software development circles. So the automatic assumption is that you are a computer programmer so don’t take it too far or you will get yourself in trouble!
6. Perceptual model. This means a system for defining how the ear or the eye perceive quality. For example, an MP3 codec uses a perceptual model of the human hearing system to determine what it can throw out what you might not miss. Perceptual models are also used for things like surround sound using headphones.
7. Gain. In engineering circles we talk about amplifier gain as a measure of how much the device is increasing (or shrinking) the input value. Think of it as “volume control” but used in a more precise way.
8. Buffer. In audio/video world, this is an amplifier that has unity gain. Remember per above, gain is amplification so unity gain means no amplification. Buffer’s job is usually to change other characteristics of the signal (e.g. its impedance).
In computer science, buffer is just a piece of memory that holds data temporarily. When a program reads a file from disc, it “buffers” it to memory. When you stream video and it takes a bit to start displaying it, the player is buffering or fetching some of it first so that it has less chance of running out of view to play.
9. Transfer function. This is a term that is used to describe the relationship of the output of a device related to its input. The transfer function of an ideal power amplifier looks like a straight line if you graphed it on an X/Y chart. An amplifier buffer has a line that is at 45 degrees.
10. Cache. Computers store their programs and data in memory. To get anything done then, the CPU which is the engine in your computer, has to keep accessing the memory. Alas, CPUs are much faster than regular memory. A cache (pronounced the same as cash) is higher speed memory that sits between the CPU and main memory. It is usually a lot smaller but holds the most frequently accessed data and thereby speeds up most operations that need access to data. We could have multiple layers of cache each become faster and smaller to save costs in having one large but expensive cache. Today's CPUs for example include three levels of caching before data has to be accessed in main memory.
As another example, the computer memory in turn is used as a “cache” for the hard disk (or even SSD). When you run a program for the first time, you see your computer disk light blink a lot. Quit out of the program and start it again and you see much less disk activity. That is because the computer memory acts as a cache and contains that portion of hard disk that held your program in memory temporarily, giving access to it at a much faster rate as computer memory is orders of magnitude faster than hard disk. This is one of the reasons why having more memory makes your computer run faster even though the only thing that does any work in your computer is the CPU, not the memory.
11. SSD, SLC and MLC. SSD stands for Solid State Disk. It uses flash memory to replace a (smaller) hard disk. Because it has no moving parts, it can have much faster performance than had disks. There are two types of flash memory: SLC which stands for single-level cell. And MLC which as you might have guessed stands for multi-level cell.
The SLC type of flash stores one bit per flash memory cell. MLC can hold 1 to 3 bits. The latter means 8 different values stored in one location. As a result, MLC flash is denser and far cheaper than SLC. Unfortunately it is also slower and much less reliable internally. Error correction is used so that as a user, you don’t notice this aspect other than shorter lifetime. And multiple units can be strung together to improve their speed.
One of the best ways to speed your system is to switch from hard disk to a high-performance SSD.
12. “IP control.” IP is the lowest level communication language used on the Internet and in your home network. When we use the term though, as in “does your lighting system have IP control” we mean whether the device can be controlled using your home network as opposed to needing a physical connection. Today, most devices are controlled using the latter. Hope is that in the future all devices are “IP controlled.”
