FAQs
Quite simple: have you thought about your listening environment ? The listening room influences the playback quality a lot. The distribution of sound waves has to follow the rules of physics regarding reflection, absorption, diffusion and resonances. It makes sense to measure all these influences and to get them under control.
Acourate is a Windows program to calculate audio filters. The generated filters are so-called FIR filters with a finite pulse response.
There are many kinds of filters, Acourate is dedicated to generate
- correction filters, these are filters for improving the sound. Keyword: How does room correction work ?
- frequency crossovers, these are filters to split the frequency range into subranges (e.g. active multiway speakers)
Other filters are used for special functions, e.g. different RIAA filters for pickup systems
Acourate supports many features to create optimal filters. These comprise the measurement of an audio system including analysis and calculation of reverberation time, frequency response smoothing, inter-aural channel coherence, harmonic distortions and further more.
All calculations are based on Pulse responses. They fully describe the properties of filters as well as loudspeakers and listening environment. Pulse responses can be manipulated by many functions, e.g. addition, subtraction, convolution.
Acourate immediately displays the pulse responses and all calculation results in a frequency response chart, in a phase chart and in a time domain chart. By the visualization the users get an idea about the result of calculations and the importance of different influences.
The main purpose of a correction filter is easy to understand. It shall correct given faults of an audio playback. E.g. if the frequency response of a loudspeaker is boosting certain frequencies in a listening room too much the correction will reduce this undesired effect and provide smooth reproduction over the full frequency range. As the left and right speakers produce more identical sounds the sound stage is definitely improved. In addition to the frequency response Acourate also corrects the time smear by phase correction. Thus the playback benefits from an improved step response.
The filters generated by Acourate can be applied in different ways in combination with the music reproduction, see How to apply Acourate filters ?
Music reproduction is not only influenced by the speakers but also a lot by the room. The radiated sound is changed by reflections, diffusion and absorption of many possible surfaces like floor, ceiling, walls but also table, couch, chairs, carpets and windows. Also the setup geometry, e.g. a non-symmetric installation, is contributing.
To improve it is necessary to identify the given actual situation. A measurement is applied to retrieve a pulse response. The pulse response precisely describes the complete behaviour of a linear time invariant system. The Acourate measurement uses the playback and recording of a logarithmic sine sweep test signal. The pulse response is deduced form the recording. It describes the system's behaviour for a certain time period. Now to prepare a correction it is necessary to analyze and interpret the pulse response. The applied strategy and calculations are important factors for the achieved result.
Another necessity for the desired improvement is the definition of a target. It may vary depending on personal taste. A common way is to describe the desired frequency response by a target curve. Acourate uses sophisticated functions to support an easy target curve design.
The resulting correction is calculated by Acourate with the given measurement and target. The correction is a pulse response now interpreted as a correction filter. The music information will be "corrected" by the filter in a process called convolution. The music is modified in such way that in combination with the playback system (speaker/room) it will reach the ear correctly. Another example to demonstrate the principle: a car which is dragging to the left side is steered to the right direction to achieve a straight forward movement.
Because a correction is mainly effective at the position of the microphone during the measurement it is necessary to take care about a smooth correction. A theoretical overall correction will generate a wrong result in case of even small positional changes. Thus a sophisticated correction should mainly influence the direct sound wave. Later effects caused by indirect sound (e.g. echo effects) are windowed out by a frequency dependent windowing.
A convolution is a mathematical expression which describes how a signal is combined with another signal or filter. In the discrete domain (working with digital and sampled data) the convolution is a calculation based on multiplications and additions. Here is an example:
A signal of length n=5 has the values 1, 1, -2, 0, 1
A filter of length k=3 has the values 1, 2, 1
Each value of the filter will be multiplied by each value of the signal, this results to
1, 1,-2, 0, 1 [= (1, 1,-2, 0, 1) * 1]
2, 2,-4, 0, 2 [= (1, 1,-2, 0, 1) * 2]
1, 1,-2, 0, 1 [= (1, 1,-2, 0, 1) * 1]
-------------------
1, 3, 1,-3,-1, 2, 1
The result is the sum of each column, the result signal has a length of n+k-1=7 and the values are now 1, 3, 1, -3, -1, 2, 1. for the given example 15 multiplications and 7 additions have been carried out. It should be clear that with long signals and filters the number of operations is very much bigger. By applying mathematical tricks like data transformation it is possible to reduce the calculation load and to achieve even close to real time behaviour. A nice transformation method is the Fast Fourier transform (FFT) and its inverse transform.
The convolution function is straightforward. The result has always to be the same with a given signal and filter, independent of the applied software . There may be algorithm differences regarding the efficiency and the required computation time. Depending on the underlying accuracy the results also can vary a bit. Acourate is doing all calculations with 64 bit double precision floating point.