Jazz Guitar

Folk/Rock Guitar & Vocal

A Bluetooth Blind Test

The test below will allow anyone to hear how Bluetooth affects the quality of music reproduction.  You can read through the technical details below or just scroll down to "How to Do the Test."

Why I Did This

In order to perform a legitimate evaluation of Bluetooth's sound quality, you must compare the original source material with material that has been data-compressed, transmitted, then decompressed through Bluetooth. The two must be compared through the exact same system at the exact same volume. Otherwise, you have no way of knowing if the differences you hear are due to Bluetooth, or to the differences between the two systems.

With Bluetooth headphones, especially, you are not listening under the same conditions when you compare the sound of the Bluetooth connection with the sound using a cabled connection. The wired connection usually relies on the digital-to-analog converter (DAC) and headphone amp built into the source device (i.e., your phone or tablet), while the Bluetooth connection uses the headphone's internal digital-to-analog converter and amp, and may even add equalization through a built-in digital signal processor. It may be a valid test of whether that particular headphone sounds better with a wired or wireless connection, but it's not a valid test of Bluetooth itself.

And if the test is not blind -- i.e., you know which sample is Bluetooth before you listen -- you have no way of knowing how your biases might affect your results.

How I Did This

For this test, I took four 1-minute clips of music and processed them in a variety of ways. First, I ran a 192 kbps MP3 file through SBC (the standard Bluetooth codec) and aptX (an optional Bluetooth codec that claims to deliver better sound). Although almost all files transmitted through Bluetooth originate as MP3s or another data-compressed format, I also tried transmitting an uncompressed WAV file through aptX just to see if it would make a big difference. (Due to an unreliable Bluetooth transmitter in my laptop, I was unable to send WAV files through SBC.)

I then recorded the resulting signals by receiving them on an Audioengine B1 Bluetooth receiver, connected via Toslink optical cable to an ART USB Phono Plus, which converted the Toslink signal to USB so I could then record it on a laptop using Sony SoundForge. I recorded these as WAV files, so they will give you the exact same signal you'd get out of the Bluetooth chip built into your headphone or wireless speaker or adapter. They remained in the digital domain the entire time, were never converted to analog at any stage, and received no additional compression.

The Test Clips

Jazz Saxophone #1: "Amazonas" by L.A. saxophonist Terry Landry, recorded by producer Joe Harley live in the studio and mastered by Bernie Grundman; recorded and mastered on DSD

Jazz Saxophone #2: "JD4" by N.Y. saxophonist David Aaron, from the CD Dismiss Television

Jazz Guitar: "Clinton Hill" by N.Y. guitarist Spencer Katzman, from the CD 5 is the New 3

Folk/Rock Guitar & Vocals: "Childish Things" by N.Y. singer/songwriter George Stass

How to Do the Test

For each music clip, there are four versions: original WAV, MP3 through SBC, MP3 through aptX, and WAV through aptX. The four versions of each clip are presented in random order. Clicking on any image will let you open the associated sound clip. (This works in Chrome, Firefox and Explorer, but not Safari.) These are roughly 10MB files, so it will take several seconds to open each one. Use high-quality wired headphones or connect your computer to a good stereo system. DO NOT use Bluetooth headphones for this, because that will add an extra layer of Bluetooth processing. Now note which clips sound best and worst to you. When you're finished listening to each version of the four clips, click on the "Which Is Which?" button below the clips to find out which formats you were hearing.


Note that there's a fifth selection on the jazz guitar clip. For this one, I've added a file of WAV through the new aptX HD format, which uses a higher bit rate than original aptX and thus promises better quality. I created this file using an Astell&Kern KANN portable music player and a new aptX HD version of the Audioengine B1. Because the ART Phono Plus wouldn't take signals from the new B1, I had to do this on a friend's computer that accepted the B1's Toslink signal. Unfortunately, it was a complicated process and I couldn't monitor the results; this was the only file that didn't have the first few seconds chopped off. 

Jazz Saxophone #2

Jazz Saxophone #1