A Funny Thing Happened on the Way from the D/A Converter ...

Ultrasonic Noise from EMU 0202 USB `Soundcards'

Well here's a funny thing ... This is the output of an EMU 0202 soundcard D/A converter replaying what should be silence. Both channels do this. A second EMU 0202 does this. Probably they all do it.

What we are looking at here is a large amplitude, ultrasonic, pseudo-random signal that can reach up to 2V peak-to-peak. It has a spectrum that starts climbing from pretty much nothing starting at around 50kHz. It seems to extend up to around 300kHz peaking somewhere around 110-120kHz.

Now, of course, you can't hear this (unless you are a bat). But, even so, I find this really very surprising. I had a high opinion of this `soundcard' (and I still do), but it looks very much like they have omitted any kind of analog low pass filtering on the D/A outputs. Admittedly, that filtering might not be so easy given that the sampling rate is variable. Then again, a slope starting around 30kHz surely couldn't hurt. Given that humans are going to be listening to this signal eventually.

Although you cannot hear this signal, it isn't so certain that it can't have any effect. It could potentially cause intermodulation distortion in downstream equipment, or fool level measuring systems, or ... who knows what? I mean ... it isn't exactly small now is it? You could argue that downstream equipment should have low pass filters on its inputs. Well, maybe ... maybe not.

It makes you wonder how common this sort of thing is with other makes and models. Probably very common. Although if you search for things like "Audio D/A converter ultrasonic noise" on the Internet you won't find anything about this. Maybe I haven't found the magic search terms. Or maybe no one has made this sort of simple measurement. Hard to believe ... but if you do all your measurements with another soundcard or only ever look up to 20kHz, you won't see it.

Digital audio does keep throwing up surprises. Surprises for me, anyway.

Go home ...