Impulse Responses of the room
Right source position:
Left source position:
Comparison at right source position
comparison left/right source
When we look closer to the impulse responses we can see several peaks after the actual impulse. These peaks are reflections from the surrounding walls, seats and objects.
For the different microphone positions we get, as expected, different room impulse responses. As they are at different distances to reflecting surfaces the peaks have different time spacings and amplitudes.
So if we want to check if the reflections can be considered as echoes it's enough to look at the time when the first reflection peak occurs. For the mesurements made with the source placed at the right side of the black board, all reflections are between 10 and 25 ms after the actual impulse. That's a too short time delay for our ears to perceive the difference between direct and reflected wave as an echo (at least 40 ms would be necessary). Furthermore we can say that for the positions where the first reflection comes within a short enough time delay the delayed sound will influence the intensity of the direct sound, i.e the original sound level will increase.
In order to see which peak comes from which reflecting wall we can make some calculations using the path distance of the reflected sound waves in order to find out at what time delay each reflected wave hits the microphone.
For instance for microphone at position 3:
- distance to left wall: 1.1m
- distance to front wall: 5.7 m
Right source:
- distance to left wall: 6.14m
- distance to front wall: 1.15m
So the reflected paths have the distance:
on the front wall: d=8.5om --> t=25,0 ms
on the left wall: d=8.55m --> t=25,2 ms
We can deduce that the reflections from both walls nearly arrive at the same time at the receiver position. That could explain the negative interference (big dip, opposite phase of incoming waves). The second peak probably then comes from the wave which is reflected on the front wall first and then on the left wall.
For every microphone position, looking at the distances to the reflecting surfaces (on the first page), we can deduce to which surface first and second reflection can be assigned.
Here is a link to more detail sketch with sources' position and microphones' position. The sketch will be opened in another page then one can have the well defined position while reading the conclusions.
For the right source position
Comparison microphone position 1 / position 2
For instance, if we compare the first and the second position, we can clearly see the time difference at which the first reflection appears. At position 1 the first big reflection appears at around 10 ms, while for the second position it appears at nearly 20 ms with a way smaller amplitude. This time shift and amplitude difference could be explained as follows:
At the first position (in the middle between left and right wall, a few meters away from the white/black board) the first reflecting surface would be the white board.
This could be the reason for the first blue peak. In case of the second position (on the right, next to the brick wall and the windows with curtains, farer away from the white board than position 1) the waves which should be reflected from the wall are, in fact, absorbed.
So the next big reflecting surface is the white board which is farer away.
We also have to take into account the absorption of the seats. The farer the microphone is away from a reflecting surface the more sound power is absorbed on its way to the wall and back.
Comparison position 2 / position 3
When we compare the second and the third position (next to the left wall which seems to be well reflecting)
we can see that, for the third position, there is a quite important reflection shortly after 10 ms which probably comes from the left wall which is much less absorbant than the brick wall on the right.
Comparison position 4 / position 5
Again, in this case, for position 4 and 5, the distance to the white board in the front of the room is equal (and far away). The only difference is that position 4 is closer to the brick wall and position 5 closer to the slick wall. This time both measurement points are close to seats.
We can notice that there is nearly no amplitude difference in the first peak of the 4th or the 5th position. As the front board is quite far away we can assume that for both positions, the first reflection comes from the slick, hard wall on the left of the room. This would explain why there's only a time shift between the two peaks.
Comparison position 1 / position 4
At last, if we compare the 1st to the 4th position which are both nearly at the same distance to the left wall, but with a great distance difference to the front wall, we can see that the 1st position has much quicker reflections than the other one (nearly 10 ms less). That would mean that the reflections due to the front wall are more important, i.e they influence more the impulse response than the left wall reflections do.
Comparison left source / right source
Another interesting point could be to investigate the difference between the two source positions when measuring at one defined point.
For instance when we look at the comparison in point 3, we see that the first reflection due to the right source is coming much faster than the first reflection due to the left source. At first sight this could seem weird, as the left source is much closer to the microphone. But, as a matter of fact, it might be quite logic. If we look at the time difference between the direct wave and the first reflected wave for the two sources, we can see how this order appears. For the right source the path difference of the direct and the reflected wave from the front wall is smaller than the path difference for the left source
When we make the calculations with the actual distances we had used for the measurement, we get a difference in time delay of about 10 ms. On the figure we get an approximate time delay of 7 ms. The difference of the values can be a result of the approximations we did for the correct microphone and source positions. Nevertheless the calculation shows us that our figure shoud be correct.