Strength – direct sound as reference

No calibration is needed for this option. h10(t) is estimated by windowing the actual impulse response we want to compute Strength from. And Lref is calculated subsequently. In this case it is possible to set the Direct sound window length and Temperature shown below.

 

 

Set the Direct sound window length so the first reflection is not included. If you are unsure, you can use 5 ms.

 

This option is incorporated to provide a simple to use reference level calculation, without the use of separate reference response measurement or time-consuming reference level measurements. The reference free-field impulse response is estimated from the measured impulse response as:

                                         (7)

where r is the estimated source to receiver distance and w(t) is a half-Hamming window starting on the arrival of the direct sound in the impulse response. After windowing, the reference response is filtered, and the reference levels are calculated.

 

The window length can be specified in the dialog, in milliseconds. A recommended value is 5 ms, which provides a useful trade-off between incorporating the impulse response of the loudspeaker while windowing out reflections from the surroundings. In normal concert hall measurements especially the stage floor reflection can be bothersome in this respect, it will in most cases fall inside the 5 ms window.

 

The source-to-receiver distance is estimated based on the detected initial time delay of the direct sound, which requires that the measurement must have been performed in a way that the initial time delay is correct. The temperature on the measurement site can also be specified, in degrees Celsius, as this influences the speed of sound and thereby the estimated distance. No great accuracy is required, and in most room measurements 20°C, normal room temperature, can be used. It should also be noted that the initial time delay detected from the impulse response will provide the distance to the nearest face of the loudspeaker, not necessarily to the acoustic center of the source.

 

Comparison with properly calibrated measurements have demonstrated that this method will normally provide too high G values at low frequencies due to insufficient window length, fairly good G values at mid frequencies (500-1000Hz), and too low G values at high frequencies due to the influence of the immediate surroundings of the transducers. Because of this G cannot be said to be calculated according to the ISO 3382 requirements.