For this experiment, my group (Harrison and Alec) ran sound through a Roland Jazz Chorus amp, with flattened EQ. I must say that this amp is not meant to reproduce sounds with the same accuracy as a studio monitor or a loudspeaker, and the sound was probably very colored. We figured that we should use the amp instead of the Mackie monitors that are in the tech closet because they (the monitors) cannot reproduce frequencies less than 150 or so Hz with accuracy.
We placed the amp about five feet from the front wall of Studio F next to a table from which we ran max using cycle~ noise~ and pink~ to produce the desired tones for the experiment. Placing the amp close to the walls would create a flutter effect, where high frequencies reflect off the wall and distort your perception of the sound. Alec recorded data and timed the experiment, Harrison took SPL measurements using the Decibel 10th iPhone app and I used Max to generate the sine waves and noise. We moved a chair with Harrison’s phone sat on it around the amp in order to record data.
POLAR GRAPH FOR WHITE & PINK NOISE:
POLAR GRAPH FOR SINE WAVES @ 9600Hz and 100Hz
When we recorded data, it was clear that SPL changed depending on the angle at which the SPL meter was placed in relation to the amp. It is clear that each tone or noise generated by the amp has a higher SPL on axis or off 45o off axis from the amp. This particular amplifier is designed to project sound forward - therefore it makes sense that sounds would have a higher SPL in front of the amp.
Off-axis sounds have lower SPL than on-axis sounds, because of the directionality of the amp. However low frequencies, with long wavelength are omnidirectional and tend to bend around smaller objects such as an amplifier. In this case the 100Hz tone had an SPL from 50-55dBA at 1.5 ft and the 250Hz tone had a range of 53-65dBA. This 250Hz tone has a shorter wavelength than the 100Hz tone and therefore is less omnidirectional.
High frequencies are very directional because of their short wavelength. A high frequency noise will reflect off a surface that it comes into contact with instead of bending around it. This explains why the 9.6kHz tone had higher SPL readings in the front and back than on the sides.
Hold up…. In the back? Why is this? I thought that high frequency noises were very directional. Off-axis high frequency sounds tend to cancel because of their similar wavelength. This explains why they have lower SPL than in the front or in the back. This still does not explain why the SPL is as high in the back as it is in the front. Because the amp was placed fairly close to the wall the increase in SPL could be attributed to rear-reflected sounds escaping from the back of the amp that are reflected off the wall.
When we measured SPL for pink noise and white noise there were several distinct differences in SPL. Firstly it would make sense to create a distinction between white noise and pink noise. White noise has energy per interval of frequency and pink noise has equal energy per octave. Therefore, white noise has more energy in the high frequency range and pink noise has more energy in the low frequency range. As shown by the polar chart and the data gathered, pink noise is more omnidirectional than white noise, although it has a lower average SPL than white noise does. White noise is more directional because it has more energy in higher frequencies, and is more directional. The noise signals differ from the sine waves because they are not simple sounds. The sine waves only vibrate at one frequency, whereas the noise signals are combinations of many sine waves. Therefore sine waves and noise will have different polar patterns.
If this experiment were conducted in an anechoic chamber, there would be no noise floor, unlike the noise floor of more than 30dBA in Studio F. The noises that the experimenters would make, such as rustling of clothes, footsteps or clicking of keyboard keys would be factored into the SPL reading. Furthermore, there would be no reflections from surrounding walls, providing a more accurate reading of the directionality of the speaker used in the experiment.
The Roland Jazz Chorus amp is designed to project on axis sound so that it may be heard far from the amp. This amp succeeds at its purpose, but does so with fairly large amounts of rear and side-radiated sound, which is to be expected from any amplifier or speaker. This experiment has also provided tangible proof of the directionality of high frequencies and the omnidirectionality of low frequencies, and shows off-axis sound cancelation in action.
When mixing music it makes sense to place monitors at head level and make sure that you are from 0 to 45 degrees on axis of the monitors. If you fail to do this, it could result in uneven frequency response and significant loss of high frequencies, which will negatively impact your mix. Furthermore, it is ideal to mix in an environment with a very low noise threshold. When we performed this experiment, it was nearly impossible to hear the 100Hz at low volume due to the background noise. This could cause you to boost LF and HF and negatively impact your mix.