This posting is in response to a question raised by Ogre, in the "Alec where are you... Thread"
Firstly, as I'm sure you know, vinyl, tape and wire recordings are based on Analog technology and as such, are subject to continuous degradation with time and use. Also, copying from analog sources, invariably results in further degradation of the sound. With the advent of Digital audio, CDs Digital Tape (DAT), mini disc etc., the disadvantages of Analog audio all but disappeared. However, the 'Front End' and 'Rear End' of Digital audio systems still require a bit of Analog technology - the microphone and associated amplifier to pick up the original sound and the power amplifier and Loudspeaker to reproduce the sound. Copying digitally does not degrade the recording and a copy of a copy of a copy......... should be just as good as the original - assuming no errors in the copying process.
As far as the direction of flow of an electric current is concerned, if you regard current flow as a movement of electrons, which are Negatively charged particles, they will always flow toward a more Positively charged area - Unlike charges attract -Like charges repel. As you say, in the Thermionic Valve the cathode is negative with respect to the anode and the electron current always flows from cathode to anode.
Water flowing Uphill ?? I suppose you could make it do so by exerting force on it (hose pipe) but remove the force and ....... gravity seems to take over again.
To answer your question about how the KN7000 is able to reproduce the musical sounds, requires a basic knowledge of the difference between Analog and Digital recording or storage. It's a bit long winded but may help you to get to grips with the principle. I have tried to keep the explanation as simple as possible.
First, take a simple example of controlling a light bulb. You can do this two ways
1. By a simple ON/Off switch
2. By a Dimmer controller.
Method 2. the Dimmer, is an example of Analogue control, where you can smoothly and continuously vary the light intensity of the bulb, giving you a virtually infinite number of output levels.
Method 1. the Switch, is an example of Digital control - the bulb is either ON or it's OFF.
We can extend the Digital control by arranging another switch to produce say a bulb output of half intensity, so operating one switch produces half output and then operating the other, produces full output. Taking this principle a bit further, we could add further switches to produce say quarter output, eighth output, sixteenth output and so on..... So, with our array of switches (and a bit of simple circuitry in between) we can produce quite a large variation in light output from our bulb. Not only can we operate just one switch at a time, but using combinations of switches, we can achieve intermediate light output levels - for example, if the Half and Quarter switches are both ON, then the light output will be three quarters of maximum intensity. If the 'Digital' control system is extended to just eight switches, then by using all possible combinations of the switches, we can achieve 255 different light levels and OFF. The switches would be arranged in a '1,2,4,8,16,32,64,128' sequence so that operating switch ,1 would produce 1/256 of the maximum light output, operating switch 2 would produce 2/256 of maximum output, switch 5 would produce 16/256 of maximum output and so on. Combining the switches produces all values in 1/256 steps, between 0 and full output - well, not quite, since all switches ON only produces 255/256 of maximum.
Digital devices like computers, our KN keyboards and even CDs, store data in a similar way to a switch. Their memory cells are just a two state device - the cell is either ON or OFF or a voltage is Present or Not present, There are no intermediate values. Their state is generally represented by a '1' or a '0' - '1' = ON and '0' = OFF. This is the basis of the Binary numbering system, on which all Digital equipment operates. So, if a set of 8 memory cells are configured '1', '0', '1', '1', '1', '0', '0', '1' this represents a Binary number '10111001' If we convert this into a number which we understand more easily ie Decimal, we would arrive at a value of 185. This is achieved by adding together the 'Weighted' values of each Binary Digit (the '0's and '1's) The convention is that the Left Binary digit has the highest value (128) and the values reduce as they move to the right. In the case above, if we apply the appropriate values to '10111001' we have :
(1x128)+(0x64)+(1x32)+(1x16)+(1x8)+(0x4)+(0x2)+(1x1). Adding up these values 128+32+16+8+1 = 185. So, with all the combinations of '0's and '1's, we can achieve all decimal values between 0 and 255. Any number can be represented in a binary form and each '1' or '0' is called a 'BIT' (Binary digIT) A group of 8 BITs is called a BYTE. So one Byte can represent any decimal number between 0 and 255. Larger numbers are represented by groups of BYTEs..A number represented by One Byte is often referred to as an '8 Bit Value' If Two Bytes are used, then this would be a '16 Bit Value'. A 16 Bit value can represent all numbers between 0 and 65535
Just to complicate things
a commonly used Value is '12 Bit' which uses One and a half BYTEs to represent the value. This is really a 16 bit value with the 4 left BITs always set to '0,' so it can represent numbers between 0 and 4095.
Representation of Fractional and Negative numbers is also possible but is more complex and is outside the scope of this article.
You may be thinking - why go to all this complication just to store a numeric value? It is quite possible to represent any Analog value by storing a voltage on a Capacitor - the number 185 could be represented by a voltage of say 1.85Volts and number 220 could be represented by a voltage of 2.20Volts. However, Capacitors are not perfect and the stored voltage will slowly drift away from it's original value, thus producing false information. Using the Digital approach, a memory cell is either ON or OFF - nothing in between.
There are several types of Digital memory - some are Volatile and some Non Volatile. Volatile memory retains its content as long as power is applied to the device. Non Volatile memory is capable of retaining it's content, even when power is removed. Some examples - RAM in your PC is Volatile - it only retains data as long as power is applied. Other memory on your PC, like the BIOS is in a ROM or Read Only Memory which is non Volatile. All the original Sampled voices and styles on the KN keyboards, are stored in non volatile memory, which cannot be overwritten. The SD card is also a non volatile memory but is of a different type, in that it will retain the data written to it, until it is overwritten - just like a disc, but using a different technology to store the '1's and '0's.
All the built in voices in the KNs are Sampled from real instruments. This process requires the original acoustic sounds to be converted first into an electrical signal, in Analog form and then to Digital form, as streams of '1's and '0's. The first conversion is of course carried out by a microphone. The small electrical signal is then amplified and fed to a device called an 'Analog to Digital Converter' or ADC. The digital output from the ADC can then be stored in memory. To play back the data from memory, the Digital data is fed to a 'Digital to Analog Converter' or DAC and then to a normal Analog amplifier and speaker system, which should faithfully reproduce the original sound - well that's the ideal situation.
To understand the sound Sampling process, imagine the following process.
Try to describe an egg or an apple, over the phone, to someone who is at a remote location, and has never seen either of these objects.
If you place a hard boiled egg in one of these devices which uses a series of wires, to slice it up, for a salad, you end up with a series of disc shaped pieces of egg. Alternatively, you can slice an apple with a knife, into a series of discs. Now, provided these pieces are reassembled in the correct order, you can reconstitute the original egg or apple. Any solid object can be described as a series of slices, assembled in a particular order.
Taking the case of the egg, it could be built up from a series of say 10 circular discs of different diameters. So, having 'dismantled' the egg, you could measure the mean diameter of each slice and if all slices were the same thickness, you could instruct your friend at the remote location, to make 10 wooden discs of the measured diameters and thickness. Each disc would have a number (1-10) related to it's diameter so that when the wooden ones were put together in the correct order, they would give a fairly good representation of the egg. However, since no information was passed about the natural slope on the edge of the original egg discs, and the wooden discs probably have straight edges, and the reconstituted wooden egg would have a stepped outline, instead of the smooth outline of the original egg.
This situation can be improved by taking more slices, say 100 instead of 10, then the slope on the edges of the original will be less significant and the surface of the reconstituted wooden egg will be much smoother, and more representative of the original egg. If we used about 1000 slices - the reconstituted egg, using say thin card discs instead of wood, would be almost a perfect replica of the original.
Instead of describing each of the diameters as a decimal value, they could be described in Digital form ( '1's and '0' )
The accuracy of the diameter of each disc would be a function of how many BYTEs are used to describe the particular diameter, If only one BYTE (an 8 Bit value) is used, then the accuracy would be 1 part in 256 which is about 0.4%. Improved accuracy of the measurement can be achieved by using a 12 Bit value (1 part in 4096) or better still, a 16 Bit value (1 part in 65536) about 0.0015%.
So, having grasped this basic idea, (I hope) if you now imagine the sound waveform of say a pure sine wave 'Flute-ish' sound, the ADC takes regular slices, or samples of the sound waveform, at fixed regular intervals, and then passes them to a series of memory locations. This process is a bit difficult to describe without diagrams. Think of a room with a flat even floor and a very wavy ceiling. Start measuring at one wall, the distance between floor and ceiling and move in a straight line toward the opposite wall, measuring the floor-ceiling distance at regular intervals. Note all the distances measured and you will then have a representation of the shape of the ceiling, which could then be drawn on a piece of graph paper.
In this analogy, the floor-ceiling distance represents the amplitude of the sampled waveform and its 'BIT resolution', while the number of samples taken represents the sampling frequency - the number of samples taken per second ( Hertz or Kilohertz)
For good quality audio, 16 Bit sampling would normally be used with a sampling frequency of 44 Kilohertz. I'm not sure what is used in the original samples in the KN keyboards - could be even better at 24 Bit.
The process holds good for Digital sampling and storing of any physical quantity, which can be converted to an electrical signal.
I hope this 'short' dissertation is of some help - it really only scratches the surface. If I can be of further help, just ask.
------------------
Willum
[This message has been edited by Bill Norrie (edited 01-17-2003).]
Previous Topic
Index