Author:
Source
In this short article, I’d like return to a unique feature of my filter generation service. This is the differentiated management of direct and reverberated sound.
As explained in the “Technology” section of my website, it’s vital to have both direct and reverberant sound well balanced and consistent between them.
Reverberant sound is impacted to first order by the radiation pattern of loudspeakers, and many of them show defects, particularly in the transition zones between different loudspeakers.
This is particularly true of 2-way loudspeakers, as in this example, which shows a dip in the transition zone between the woofer and tweeter. Correction helps to reduce this gap and achieve a
better balance between direct and reverberant sound.
Another advantage of digital correction of reverberant sound is its temporal aspect, reducing the impact of early reflections. These reflections are important for the perception of space, but
they also need to be accompanied by good diffusion to improve the soundstage perception. It’s not always possible to install diffusion panels in a living room not dedicated to listening, and my
innovation makes it possible to simulate the effect of such panels.
To characterize the effect, it’s valuable to plot the auto-correlation of the impulse response. This indicator shows the degree of similarity between direct sound and reflections. The lower the
auto-correlation, the higher the temporal diffusion. On a theoretical impulse response composed of a perfectly linear direct sound and a reflection equal to half the direct sound (as if the wall
reflected the sound with just an attenuation but no diffusion), the auto-correlation indicator would be as shown in the figure below.
Let’s move to a real case using measurements from one of my customers.
With the HAF correction, the auto-correlation value decreases significantly with more random pattern, corresponding to a better diffusion.
