Mobile Electronics Magazine

Switch to desktop

September Issue Feature: Tech Today - DSP Tuning, Part 4

9-28-2017, Mobile Electronics -- In the last segment on this topic, I suggested a process to be used in the installation bay for tuning cars that’s effective and efficient. For some readers, this process may differ greatly from what you’ve been told by numerous enthusiasts, sound quality judges and other accomplished tuners. So, before we get started here, I’d like to explain why.
There’s a big difference between tuning cars as a profession and tuning cars as a hobby. For the hobbyist, the tuning is often the end rather than the means. The tuner likes to spend hours experimenting, listening and retuning. For the professional, these extra hours spent on listening and retuning eat into our profits and our ability to move on to the next car.

Finding Balance
First, my objective in all of the tech tips I write—whether those tips are on the Audiofrog web forum, on Facebook or in articles like this one—is to provide an appropriate balance between speed, predictability and performance. In the interest of speed and predictability, I favor objective processes that don’t require us to use our ears and make a thousand subjective analyses and an endless series of adjustments. There’s a place for subjective analysis, but that’s after all objective measures have been exhausted.
There’s a temptation among many of us to see the speedy and objective process as worse than the lengthy “artisan” process of tuning primarily by ear. This is a fallacious argument if the quality of the performance that the two processes provide is the same or even similar.
Here’s an example: Every subwoofer box built for a round subwoofer needs a hole in the baffle that fits the subwoofer. What’s the appropriate tool? Most of us would immediately say it’s a router with a circle template or a circle cutting jig. For some, the answer is a CNC router. We don’t bash these processes as lacking the necessary opportunity to include our “art” or our skill in the process, even though cutting the circle freehand and with no line to follow using a jigsaw would better demonstrate our circle cutting skills.
What are the chances that we’d cut a circle with a jigsaw and no guide to follow as accurately as our CNC machine? Not good. Eventually, with a series of files and sandpaper, we might get close and after a much longer process, we could demonstrate that the outcomes are similar. What’s the difference? Cost. If we’re charging the customer $30 to cut a round hole, then it behooves us to use the most efficient method that provides an appropriate outcome.
In our circle cutting example, the objective process is not only speedier, it’s more accurate, too. Determining if the circle is correct is a simple matter; either the speaker fits or it doesn’t fit.
“Hey, there’s a difference. Sound is subjective but a circle isn’t!”
Yes, that’s true. Sound is somewhat subjective. Some customers prefer more bass. Some prefer less high frequency content. That doesn’t change what stereo systems and a stereo recordings are designed to do. That design dictates what’s correct. The system is correct when the left and right frequency responses match and the signals arrive at the listener in phase. After it is correct, we can make some adjustments for personal preference.
When we are tuning cars, our objective tools and the information they display allow us to see how far from correct we are in each step of the process. Because of the way our brains work in processing what we hear, it’s helpful for us to use analysis methods that correlate well with what we hear. Some measurement processes are better than others.
We’re all probably familiar with the situation in which the RTA curve appears to be correct, but the car sounds terrible. In some cases, tuners use this as a reason to reject the tool, rather than look deeper into the reason that what appears to be correct is not. The first question to ask in that situation is, “Does this measurement make sense?” The second question is, “What am I really measuring?” The RTA doesn’t lie, but it also doesn’t completely characterize the performance of the system. It shows us one aspect of performance.
Do tools exist that allow us to completely characterize the performance of the system? Sure. Do we all know how to use them? No. Is it necessary to completely characterize and correct everything? No.
Our next consideration should be, “Which deviations from correct are inaudible?” We don’t need to focus on those. If we don’t need to focus on them, then we don’t need to spend money and time analyzing them during a production tuning process. As a skills-building exercise to be conducted on our own time, learning those processes and how to use those tools may make us better able to identify problems and solutions, but those activities should be extracurricular. When we’ve improved those skills to the point at which we can deploy them to increase accuracy, predictability or efficiency, we should introduce them into our process.
The objective of this article is to clear up a few misunderstandings about what we measure, what it means and what’s sufficient based on audibility. 

Read the rest of the story HERE.

Rate this item
(1 Vote)

Copyright - Mobile Electronics Association 2020

Top Desktop version