Notes from the Test Bench
By Bruce Hofer, Chairman & Co-Founder, Audio Precision
I just got back from some much needed “down time” with my wife enjoying the sun, trade winds, and amazing beauty of the big island of Hawaii. Among other things while we were there, we took a helicopter tour of the uninhabited northern coast with its breath-taking valleys and 2000-foot waterfalls—sights that can only be seen from the air. Several onboard video cameras recorded our adventure, resulting in a very nice DVD souvenir.
I mention this because a video can often provide many of the aspects of a live experience. AP has decided to implement this idea in a new series of videos to be released over the next year. The first in this series covers the user interface (UI) of our growing APx family of instruments. It provides a convenient and easy to understand introduction that simply can not be adequately conveyed in the plain text and static pictures of a brochure. The APx UI represents a significant departure from previous approaches, and it has been widely praised by both experienced users and new customers.
For a real hands-on demonstration, simply contact your local AP representative or distributor. We will also be at the upcoming AES convention in San Francisco (Nov 5-7) and the Electronica trade show in Munich Germany (Nov 9-12). I will be at both of these shows, and hope to see you there!
Output: Testing USB Headsets with APx
AP has released a new Technote on testing USB headsets. The full 20 page Technote 111: Testing USB Headsets, with accompanying APx projects, may be downloaded from ap.com.
Testing a USB headset presents a number of challenges. While a regular headset consists of a standard analog microphone and earphone, a USB headset adds an adapter module containing A/D and D/A converters and a USB plug.
USB adapter module and analog headset cables.
The analog I/O on the USB adapter is pretty straightforward—the device we use in our Technote, for example, has 3.5mm phone jacks. If the device you are testing is instead hard-wired to its USB module, you can just cut the wires and proceed from there.
Microphone and headset tests aren’t addressed here. Once the headset itself is detached from the USB module, it can be tested like any other acoustic device. APx500’s acoustic response measurement is well suited for this purpose. You’ll need acoustic coupling fixtures or a head (and torso) simulator to properly do acoustic tests on the headset.
Why is there no USB jack on the front of APx analyzers? Because USB is a computer bus that relies on an operating system and its drivers to operate. Without that, there’s no communication with the attached devices. That’s actually just fine for our purposes. You just plug the headset into the host PC running the APx500 measurement software. If you wish, you can use a separate PC for the headset and its drivers in order to maintain complete independence, but it isn’t necessary for getting valid results.
There are three basic types of tests that can be conducted on the USB adapter:
The connections for these first two are shown in the diagram below:
Connections for record-to-disk and play-from-disk tests.
A real-time connection to the device under test permits closed loop testing that is difficult or impossible to implement with the two-step process of recording a file and then analyzing it afterwards. Tests of maximum input or output level, gain and linearity, and delay or latency become much easier to perform. The connections needed are shown below:
Connections for real-time tests.
The full Technote gives more details, with all of the steps necessary to conduct the three types of tests. And, if you get stuck anywhere, AP’s Technical Support department is always available to help you.
Sound Advice: Measuring Internal Sample Rate Error with AP2700
Many devices contain internal clocks that are not accessible, but that need to be measured. The USB headset in the preceding article is an example of an inaccessible A/D converter. Inaccessible D/A converters are found in MP3 players and smart phones. In some cases, disassembly would make testing overly complex. In others, the device must remain assembled so that the internal temperature and environment is not altered.
The internal sample rate error of these types of devices can be indirectly determined by measuring the frequency error of a known tone. AP has developed a set of AP2700 macros that do just that.
Measuring A/D Error
To measure A/D clock error, record a tone of precise frequency to a file using the device under test. The Analog Generator with the waveform set to Sine (D/A) works well for this purpose, as it has a frequency accuracy of 2 PPM (±0.0002%).
Analog generator set to Sine (D/A) waveform.
Now, load the macro “A-to-D Sample Rate Error.apb” and start it running. At the prompt, browse for the wave file that was just recorded, enter the frequency of the recorded tone, and click OK. The analyzer will now perform an FFT, measuring the level of the tone in the central and adjacent bins, to determine its precise frequency. When finished, an FFT response graph will be shown, and the prompt will reappear with the frequency error displayed. Since the nominal sample rate of the recorded file is known, the actual sampling frequency can be calculated. In the example shown in the figures below, the device under test has shows a sample rate error of 1.0000%. Since the nominal sample rate is 48,000 Hz, the actual sample rate of the device calculates to 48,480 Hz.
A/D FFT spectrum graph.
A/D frequency error result.
Measuring D/A Error
The process for measuring D/A clock error is just the opposite. A prerecorded file of precise frequency is played on the device under test, and its analog output is connected to the analyzer input. You can make your own reference files using an audio editing program such as Adobe Audition that has a precise tone generation function built in. The macro download includes .wav and .mp3 files with a precise 1 kHz tone at both 44.1k and 48k sample rates, which should suffice for most situations. Run the macro “D-to-A Sample Rate Error.apb” and follow the prompts. The example in the figures below shows a device that plays slightly slow.
D/A FFT spectrum graph.
D/A frequency error result.
Test Results: AP News & Events
©2010 Audio Precision Inc.