DMF - final report for digital music formats PDF

Preview

Summary

final report for digital music formats...

Description

SID: 1535115

Digital Music Formats

December 2016

In this report, I experiment firstly with a MIDI keyboard and what MIDI Monitor reads about its signals, and various results of different actions made on the keyboard. Secondly, I experiment with MATLAB, manipulation and creating waves and analysing their behaviour when different parameters are changed. Part 1 - Viewing MIDI Messages 1-4. To start with, I set MIDI Monitor up as instructed in the first 4 points of the brief, and played a few notes to see what happened in the program, and it came up with the following: Figure 1

5. As shown in the picture, I was using my own synthesiser/MIDI keyboard, A Roland JUNO DS-61, for this experiment rather than the Oxygen-49 MIDI keyboard. Otherwise, every column was showing the expected values. 6. I then tested out the transpose function on my synth and got these results: Figure 2

SID: 1535115

Digital Music Formats

December 2016

Figure 2 shows that I started on a F#3 (MIDI Note 42) and, after accidentally shifting the whole keyboard down an octave, transposed up a semitone 6 times until I reached C3 (MIDI Note 36), where the keyboard allowed no further transposition. I then tried the highest natural note on the keyboard, C6, (MIDI Note 72) and again transposed it up 6 semitones until I reached F#6 (MIDI Note 78). I did not notice anything abnormal about the sound of the higher transposition, nor the results shown in MIDI Monitor. Figure 3

Figure 4

7-8. As shown in figures 3 and 4, the maximum velocity is 127 and the minimum velocity is 1. What is also shown is the fact that my JUNO has old style Note-offs. 9. Below, Figure 5 shows what happened when I moved the pitch bend wheel, first moving down to its most negative position, then back to zero, and then up to its most positive position and finally back down to zero. The minimum value is -8192, the central value is 0 and the maximum value is 8191. Figure 5

Figure 6

10. The wheel bends the notes either 2 semitones or down. This means that the value for all 4 semitones is 16384, and therefore the value for a single semitone is 4096. Figure 6 shows that the resolution for the pitch wheel is 87. 87/4096 is the pitch step, which is 0.0212 semitones.

SID: 1535115

Digital Music Formats

December 2016

11. The next screenshot seen below in Figure 7 shows what happened when I activated various other controls on the synthesiser. Where it says 'Control' under the message column, this indicates that I moved any one of these: the modulation wheel, the vocoder selector, the volume sliders and of course the keyboard itself. Where it says 'Program,' this indicates that I changed the patch or selected the pattern sequencer. When the pattern sequencer was selected, the Channel changes from 1 to 16. Figure 7

12. I switched the keyboard off, then on again and reselected it as a source in MIDI Monitor, and Figure 8 below shows what happened during this process (Nothing). Figure 8

SID: 1535115

Digital Music Formats

December 2016

Part 2 - Software Synthesis in Matlab 1. Sine Wave Figure 9

the script above shows how I constructed the sine wave, and the two screenshots below show the graphical representation of the wave; Figure 10 is zoomed out and Figure 11 is zoomed in. Figure 10

Figure 11

2. Additive Synthesis Below, Figure 12 shows the script I used to generate, play and graph the original square wave, and Figures 13 and 14 show the wave created, zoomed out and in respectively. Figures 15 to 17 show the same results but with the adjusted parameters shown in the 4th line of the Matlab script. The same goes with Figures 18-20. In the first square wave, it shows in Fig.14 that because the waves being added are not similar, there is an obvious square effect on the resultant wave. Whereas looking at the second wave that was generated, where the waves that are combined are much more similar, the square effect on this resultant wave is much less obvious. Lastly, the final wave is different, due to the fact that there is subtractive synthesis as well as additive (the second and

SID: 1535115

Digital Music Formats

December 2016

fourth waves are in anti phase with the other two). Figure 20 looks more like a Sawtooth wave than a square wave as a result this combined synthesis. Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

SID: 1535115

Digital Music Formats

December 2016

Figure 18

Figure 19

Figure 20

3. Envelopes Below, as shown in figures 21-23, is the script I used in this part of the experiment, the resulting envelope and the resulting wave in respective order.

Figure 21

Figure 22

Figure 23

SID: 1535115

Digital Music Formats

December 2016

Figures 24-26 show what happened when I altered the values for a and b, first both simultaneously and then individually. the first screenshot shows that when a and b were both doubled, the gradient was made shallower and the final amplitude was doubled. This meant that the amplitude decrease was more gradual, and final volume was higher than the original. The second screenshot shows the effect on the wave when a was halved and b was doubled, which was a much steeper gradient and a greatly reduced final amplitude. The volume was therefore reduced much more significantly and finished at a much smaller value than the original. The final screenshot, shown in Figure 26 shows what happens to the wave when a was doubled and b was halved. there is not much discernible difference between this and Figure 24.

Figure 24

Figure 25

SID: 1535115

Digital Music Formats

December 2016

Figure 26

4. Amplitude Modulation Figure 27 shows the original script I used, and figures 28 and 29 show the graphical representation of the wave, zoomed out and in respectively. Figure 27

Figure 28

Figure 29

SID: 1535115

Digital Music Formats

December 2016

Figure 30

Figures 30-34 show several examples of what happens when the values for carrier, modulator and index are changed in this amplitude modulation experiment. There are two waves being combined in the ampmod to create the result. Firstly, figure 30 has the original carrier and modulator but the index is multiplied by ten. the first wave is not altered by this, but the second is. There are two key differences between this wave and the original. The first is that there is an additional peak that is nearly as high as the first peak in the cycle. In the original, this is present but just much less pronounced. The second is that the amplitude of the whole wave is 10 times louder. The sound that is produced is much coarser, seeming more like a square wave than the original, and the volume is higher. Figure 31

Next, in figure 31, the Modulator is the variable that is changed. Once again, the second wave in the equation is the one that is affected by this change. the graphical representation of the result is quite different from the original, firstly because the wavelength is only 0.002 seconds compared to the original 0.01, so the wavelength is 5 times smaller, and because the shape of the waveform is not the same. In this part of the experiment, the only discernible change was the frequency, which

SID: 1535115

Digital Music Formats

December 2016

was higher, and can be explained by the equation Frequency = 1/Wavelength, showing that the relationship between Frequency and Wavelength is inversely and directly proportional. Figure 32

Figure 32 shows what happens when the carrier is multiplied by ten and the other variables stay at their original values. This is different to the previous two examples because it is the first wave in the equation that is altered, not the second. Here, the frequency is much higher and the waveform is massively different. Here, the wavelength is only ~0.0001 seconds long, and the waveform resembles a simple sine wave, albeit with amplitude that changes cyclically every 0.01 seconds. the sound was of a much much higher frequency though the amplitude was the same. Figures 33 and 34 show what happens when all values were simultaneously increased or decreased. Firstly they were all multiplied by 10, and secondly they were all halved. the differences that were met in the first instance we that the amplitude was x10, the wavelength was x0.1 and the waveform was slightly different, though had similar dimensions. The differences with figure 34 were that the amplitude was decreased by roughly 30%, the wavelength was doubled, and the waveform was also slightly different. The sound for the first of these two examples was much louder, abrasive and higher pitch, and the second was lower, quieter and smoother. Figure 33

SID: 1535115

Digital Music Formats

December 2016

Figure 34

5. Frequency Modulation Figure 35

Figure 35 shows the MATLAB script I used to make this frequency modulator, and the zoomed in graph of the resultant wave with the original parameters. Figure 36

SID: 1535115

Digital Music Formats

December 2016

above in figure 36, the modulator is changed and the other values remain the same. This equation is similar to the last in that there are two waves being combined, though this is additive rather than multiplied. Here the second wave is altered by the modulator, and it resulted in the pitch increasing, and the wave appearing to sound smoother. Figure 37 shows what happens when the index is multiplied by 10 and the remaining values are unchanged. Here the pitch was increased again but the wave still sounded rough, and there were no other changes. Figure 37

Figure 38 shows the result when the Carrier is the affected parameter. again, with this one the pitch is the only change, it increases, and there is no further change to the wave.

Figure 38

5. Anything For this part, I decided to form a sin wave and make it "noisy." I used the linspace function to create 400000 points along a vector between -2pi and 2pi, which I defined as t. I then plotted that sine = sin(t). I made a a small variable in order to keep the amplitude of the wave down. I then

SID: 1535115

Digital Music Formats

December 2016

defined noisysine as shown below in figure 39. The two graphs below show firstly the original sine wave and its noisy neighbour. Finally, I wrote the noisysine to a wave file using the audiowrite function.

Figure 39...