Is Ipod Audio Good Enough?
Ipods and their work-a-likes are the rage among consumers, but there are many professional opinions about whether they are good enough to use as a playback source in sound reinforcement systems. I decided to investigate by testing some personal media players.
Ipods and their work-a-likes are the rage among consumers, but there are many professional opinions about whether they are good enough to use as a playback source in sound reinforcement systems. I decided to investigate by testing some personal media players.Tricky to Test
Hardware and file format were the biggest issues in the test. The hardware had to perform well for all major performance criteria — including output level, DC offset, frequency response, phase response, crosstalk, distortion and signal-to-noise ratio — to be taken seriously as a program source for a sound system.
Personal media players are not easy to test because they are playback-only devices. One cannot execute a typical two-port measurement. Fortunately, test equipment manufacturer Audio Precision developed a personal media player test for use with its top-of-the-line analyzers. The routine involves transferring some special multitone audio files to the unit, then playing them back into the Audio Precision analyzer, which determines each of the major performance criteria.
I augmented the Audio Precision tests with a few of my own that would be particular to audio technicians. Signal polarity is important, since the audio tech may want to use the personal media player in conjunction with a polarity test receive device (such as Ivie IE35, Galaxy Cricket, Audio Toolbox). This is done by playing a special pulsed waveform into the sound system, and sampling it with the receive device that shows whether the polarity is normal or inverted. One of the tested players was reverse polarity.
The output (source) impedance also is of interest; proper interface practice is to connect an analog audio source (a low impedance) to a much higher impedance input. I tested this using the voltage divider method, which involves progressively loading the media player with a variable resistance box until its output voltage is half of its open-circuit voltage. This creates a matched interface, which means that the value of the load resistance is equal to the source impedance of the media player. Any analog audio device should never see less that 10 times its source impedance when interfaced into a sound system.No Y-Cables, Please
Many audio people are given to inappropriately using Y-cables to sum the left and right channels for mono playback of stereo program material. So I again checked their specifications under this condition.
The tests confirmed what I have been saying for years: Don't do it. Every device failed the frequency response, distortion and signal-to-noise ratio tests. The very-low output impedance of these devices (less than 6 ohms) keeps them from working well into a matched interface, which is what a Y-cable produces. If you need mono, use one channel only or get a mixer.
Handheld speech intelligibility meters are hitting the audio marketplace. These need a special test waveform to be played over the sound system, which is re-acquired and scored by the handheld analyzer. The test is called STI-PA, or Speech Transmission Index for Public Address systems. The test requires an accurate playback source, and many have felt that the MP3 file format, with its lossy compression algorithm, is unsuitable for this application. Some of the personal media players failed this test, but it wasn't because of the MP3 file format.
My final test dealt with using personal media players as a sweep track source for acquiring the room's impulse response. A 14-second, logarithmic sine wave sweep is played over the sound system. A recording of this sweep is made at the desired listener position. Comparing the “dry” sweep with the recorded sweep derives the room's impulse response. (I have written much about this process in Syn-Aud-Con newsletters over the past few years.) This was the most difficult test for any media player to pass. Many of them proved unsuitable for this use, producing aliases or false frequencies in their output as the sweep approached 20kHz. I had thought that alias issues were in the past for digital audio, but this proved me wrong.Some Surprises
One would expect that a personal media player would need to have WAV file support to pass the tests. WAV files are sample-accurate with no lossy compression schemes to save disk space, and I used this format on media players that support it. I used MP3 format on those that did not. The first device I tested using MP3, the Samsung YP-U2J, failed miserably, which made me highly suspicious of the MP3 format. This turned out to be unfounded, as the Samsung Z5 passed all tests using MP3. One can theorize that lossy compression schemes don't do much to sine waves. It even proved acceptable for the complex STI-PA noise waveform, with its rigid time and frequency characteristics.