ABX testing is a way to compare two similar versions of something, in this case sound. The user is presented with an audio stream called the A stream, and another slightly different audio stream called the B stream (or vice versa). The user can freely toggle between the A and B streams listening for differences. Once the inferior stream has been picked out, the user can then audition another stream, called the X stream, and then decide if the X stream is the same as the A or B stream.
Texas Instruments has been in the Class D power amp business for a while. And it shows. Like most tech, Class D first arrived as a solution in search of a problem and early efforts kicked around for decades waiting for the underlying silicon needs to catch up. But once it did, the days of other Class-whatever were numbered as nothing could compete with the cost and efficiency of Class D. For the last two decades, Class D amps have dominated in portable audio and PC sound. Not exactly hifi. And for more than a decade, Class D has been winning sockets in home audio 5.1 and 7.1 amplifiers thanks to its compact size, high efficiency and low cost.
About 2 years ago we did a post on an experimental mode that we’d introduced in the QA400 that used some interesting processing to reduce the noise in measurements. The full post is located here, but here’s a quick recap: AKM had published an app note (no longer on their site, it seems) that covered a technique for using N ADC converters to capture the same signal. By adding the outputs of all the converters together, the noise would tend towards zero while the signal would remain.
The use of notch filters is a well-known technique for improving the THD measurement performance of audio and RF equipment. THD measurements are a challenge because you usually have a massive fundamental that you aren’t interested in, and several very small harmonics that you are interested in. But with a notch filter, you knock the fundamental down before it has a chance to interact with the non-linearity of the measuring equipment input stages.
Note: The QA401 is now shipping. See the product page link here
Last February 2015 we wrote about the QA405, which was to be the first successor to the QA400. It is bigger/strong/faster than the QA400 in every way. We noted at the time that the product was complex and expensive. Those are both still true.
However, we also asked for feedback from users, and feedback we got! Of course, there were some that wanted $10,000 tester performance for $200. But most understood that the next 5 or 10 dB of performance comes at increasing cost. People appreciated the convenience and repeatability of the QA400, but really wanted differential IO and isolation. Over and over those two items kept coming up.
This post takes a look at a framework to enable you to write software that can perform custom analysis on captured waveforms, and display the results of that analysis, all within the application. This first release will focus on the QA400, and then we’ll add this same mechanism into an upcoming QA101/QA100 software release.
This will enable you to download visualizers from others to help with your data analysis needs. What could these do? These could be things to help analyze speaker parameters, room parameters, etc. But more importantly, the visualizers can display this information in a way that makes sense for the task at hand. Think of visualizers as domain-specific problem solvers.
We’ve had an ongoing background project which is best described as a QA400 on steroids. It features:
We had a need recently to verify the number of pulses in a data stream were correct. The pulses were driving a stepper motor, which added a bit of a twist because the stepper could jump from microstepping to full stepping once a certain speed had been reached. How could we be sure the precise number of pulses had been emitted by the FPGA? By using a custom interpreter, of course. It took about 15 minutes to modify an existing interpreter to do the trick.