|
 Measuring the Frequency Response
It is important when doing any kind of frequency analysis that the microphone stays in the same position throughout the test. Luckily for us, the size and shape of the CD case format of the Sound Check Analyser lends itself to an easy way of temporarily mounting it in the vehicle at almost driver head height. Photos showing the recommended mounting can be found HERE.
You will generally want to check the response as measured at the driving position, but if you are entering autosound competitions - where sound quality is usually judged from both the driver and front passenger seats - you should take measurements from both positions. Speaker positioning, mounting angles, reflected output (from the windscreen) etc. can create quite differing results according to the listening position.
While it is less work and it’s generally fine to take one measurement with all the speakers working, it is best to also measure the results for individual component speaker pairs. For example, have only the front midrange drivers working and plot their response. Then just have the tweeters working and make another plot (on the same graph but in a different ink colour), then the subwoofer and so on. What you will then have is a plot showing the crossover slopes as well as the frequency curve of each speaker pair. It can help to focus in on problems in the tonal balance (maybe you just need better mids!). Sometimes tweaking the crossover points or slopes is better than trying to EQ. And by comparing the individual plots with the overall plot (the one you took with all the speakers working) it might show up some phasing and summing errors.
When setting equalisers, avoid sharp dips and peaks, unless a steep dip is required to reduce a resonant peak - but to do that successfully you’ll need a parametric equaliser because graphic equalisers have an action that spreads too widely across the frequencies (too wide a bandwidth, to give it its correct term). Smooth curves are more musical. A series of sharp changes may give a flatter response, but at the expense of phase linearity and imaging.
The human ear is more forgiving than a microphone. Smooth corrections that give a final response of, for example, 40Hz - 16kHz (+ 3dB) are usually preferable to trying to achieve a totally flat response from 20Hz - 20kHz where the speakers and amplifiers may run out of headroom.
Plotting a Response Curve
Start by setting the playback volume so that the 1 kHz reference tone on Track 1 reads 0dB on the Sound Check Analyser’s meter. We advise the use of ear plugs for a more comfortable measuring experience! As the next 31 bands are played through from 20Hz to 20kHz, readings can be taken from the meter. The graph paper, which has identical calibrations, can be marked to show the level for each spot frequency.
When all the frequency points have been measured and plotted onto the graph, remove yourself to a more comfortable place and carefully form a response curve by joining up all the individual plot points. Make sure the line you draw has a smooth, gradual transition from one point to the next.
Now study the response curve, looking for any obvious probelm areas (where there is a large peak or dip between two or three of the plotted points). It’s now a case of altering crossover points or adjusting your equaliser settings to try to achieve a smoother transition from one band to the next. Use the "track repeat" (looping) function on your CD player to repeat any track (frequency band) requiring special attention.
 Optional Features
In most cases the built-in microphone will be perfectly adequate for the kind of measurements you’ll need to make, but the Sound Check Analyser also has an external microphone input (external mic is not supplied). Details of the external microphone input are included in the CD sleeve notes.
The self-contained battery pack will usually be the most convenient way to power the device. However, it does also have an external DC power input facility on the PCB. This requires a 6 Volt regulated supply (@100mA).
|
