Room acoustics and building acoustics measurements with chirp signals
Traditionally, the reverberation time is determined with swtiched noise signals. The room is stimulated with a broadband noise signal via loudspeakers and the signal is abruptly switched off. The reverberation time is determined from the decay curve of the sound level. In this measurement, the exciting signal must be well above the background sound level. The required difference should be at least 35dB-50dB. In noisy environments, rooms with highly absorbent walls and generally at lower frequencies, this required high power speakers and amplifiers. These have to be dimensioned accordingly and become very heavy, especially at lower frequencies, and are therefore bulky for mobile measurements. Furthermore, measurements with high sound levels are problematic, as they can lead to severe annoyance of the neighbors.
By using modern methods with logarithmic chirp signals, reliable measurements can also be taken at much lower sound levels.
• The loudspeakers and amplifiers can be made smaller and light-weight.
• Less annoyance due to lower sound levels.
• Less demands on the hearing protection of the measuring personnel
Practically, you can measure large rooms with small speakers. Another aspect is that with low signal levels, you can perform measurements during the working hours of the employees in an office. Already 5dB above the background level is sufficient for a measurement. This would be impossible with classic measurement technology.
If a small 10W amplifier is sufficient for the chirp measurement, you need at least 1600W for the classic measurement method with the noise switched off! for the same measurement accuracy. This applies to relatively short measurement times of just under 30s. With a measurement time of 3 minutes, 0.3W is sufficient. This corresponds to a headphone amplifier compared to a high-performance power amplifier.
Chirp Measurement Applications
• Measurement of the reverberation time in occupied offices during working hours
• Measurement of large rooms, halls, etc.
• Measurement with strong background noise (traffic)
• Measurement where you have to comply with quiet/noise regulations
• You no longer want to carry heavy loudspeakers and amplifiers
A chirp signal is comparable to a siren whose speed is increased. The signal starts at a certain frequency that increases logarithmically. When the upper limit frequency is reached, the signal repeats itself periodically. Chirp sequences always have a length that can be represented as a power of 2. So 2 4 8 16 32 64 256 1024 2048 4096 8192 16384 32768 65536 (64K) 131072 (128K) 262144 (256K) 512K 1024K etc. Long sequences from 64K upwards are used typically for room acoustics. Short sequences below half a second sound similar to birdsong and have given the measuring method its name. In the long sequences, you can clearly hear the pitch increase.
Chirp signals fall off at 3dB per octave in the spectral domain. This power distribution is more like pink noise, although the two signals sound completely different. The majority of the signal energy is therefore in the low-frequency range and is therefore ideally suited to the typical load capacity of the loudspeakers.
Furthermore, an upper and lower limit frequency can be defined for chirp measurements, so the entire signal energy is bundled in this frequency range. Therefore, an excitation signal is generated from the outset only in the frequency range where it is required and does not have to be filtered first.
Measurement method with chirps
The measurement of room acoustics with chirp signals belongs to the class of correlation methods.
With the two older methods (impulse and noise), the signal-to-noise ratio can only be improved by higher transmission levels. A higher transmission level means appropriately dimensioned and heavy loudspeakers, especially when measuring with the noise method.
The trick to measuring with correlation techniques is not to increase the level, but rather the “transmission time” of the speaker. Mathematical methods can be used to convert this longer measurement period into a higher level. A long measurement time is not a problem for loudspeakers. However, short and high levels can only be achieved with a great deal of technical effort.
The process can be clearly explained with an analogy from photography. Noise appears in the image when there is little light. This can be reduced by increasing the exposure time. In this way, you can achieve images that are as bright as day even in almost complete darkness. Here, of course, there is the restriction that this only works with stationary objects. But this is exactly the case in room acoustics/building acoustics. Nothing is moving.
With the noise method, you cannot improve the result with a long measurement time. The room is stimulated by the loudspeaker and then switched off. For the measurement itself, only the switch-off process is important.
In the noise measurement, acoustic interference has a direct effect and changes the measurement results. The same applies to the direct impulse measurement. A longer measurement time does not help here.
With both methods, the result can be improved by averaging individual measurements. But this is not very effective.
With these methods we have no possibility to increase the "exposure time". This is only possible with the correlation methods (MLS and Chirp).
Correlation methods can suppress interference. With classic noise methods, it is hardly possible to measure a room on a busy street with an open window. This is where the correlation methods can demonstrate their full strength.
These modern methods are integrated in our measuring software Akulap