Graves on SOHO Technology

For the purposes of this review I thought I’d conduct a little test to mimic that kind of situations that I face when I travel. I often have a small group of people collected around a small table, perhaps at a hotel.

To mimic this I set the PSG-01s in the middle of my dining room table. It’s a rectangular table to the far ends are a little more distant than the sides. The PSG-01s is facing one of the far end, where my laptop can likely be heard.

Starting with the PSG-01s in its normal conference mode I made a recording as I moved from position to position around the table. All the while I blathered on about the PSG-01s, recording my voice to Cool Edit Pro 2.1 in a high quality mono wave format.

http://www.mgraves.org/wp-content/uploads/2010/01/YAMAHA-PSG-01S-CONFERENCE-MODE-WAVEFORM.jpg" rel="lightbox[7074]" title="YAMAHA-PSG-01S-CONFERENCE-MODE-WAVEFORM-500">

For this initial recording the image above is the waveform display and  the image below is the spectral energy plot.

If you listen to the recording and look at the waveform at the same time I think you’ll get a clear sense that the PSG-01s is not truly omni-directional in its conference mode. My voice when in front of the unit is loud and clear, it’s nearly as good at each of the positions 90 degrees to the side, but when I’m at the opposite end of the table I definitely sound distant.

This observation tends to indicate a cardioid pickup pattern. The following illustration is a cardioid polar pattern overlaid on my test setup.

Let me be the first to admit that this little test setup is less than ideal. I would need a much more controlled environment, with calibrated instrumentation to conduct a definitive test.

Further, my test setup is not an accurate reflection of the situation most commonly faced in the field. In hotel rooms and small offices it’s more common to have square or round tables, not the large rectangular table in our dining room. Thus the PSG-01s may actually do better than my little test would imply.

Even if the PSG-01s is not truly omnidirectional in conference mode, it’s pickup pattern is still much broader than the ClearOne Chat 50 that I’ve been using. This makes the PSG-01s a better choice for use when there’s more than one person participating in a conference call from this location.

The spectral energy distribution display may be a little hard to read when scaled down to fit the web page. If you click on the image above you’ll get a full resolution image which makes it much easier to read the vertical axis labels. Examining these labels you’ll find that the vertical axis measures frequency from 0 – 14 KHz.

Look closely and you’ll also see that the audio output of the microphone array has been rolled off after 7 KHz. You might expect this if I had passed the audio stream through a G.722 encoded call channel. I didn’t. I recorded directly into Cool Edit Pro on my laptop. So the PSG-01s is itself performing a low pass filter on the audio stream.

Normally I would not even feel that this was noteworthy since much of wideband telephony currently happens using codecs employing a 16 KHz sample rate, so rolling off above around 7 KHz is a perfectly rational thing to do.

However, I do find it a bit odd for such filtering to be implemented in a “Skype Certified” audio device. Skype often makes use of their own SILK codec which is capable of sample rates well beyond 16 KHz. The PSG-01s doesn’t really take optimal advantage of the available audio bandwidth.

Truthfully, in making this observation I am really picking nits. The PSG-01s sounds great. It delivers a genuine wideband call experience, and that’s nothing to sneeze at. It’s merely the inconsistency between the engineering and the marketing realms I often find amusing.