Downloadable MPX files
| Low frequency performance |
 |
High Pass Filters and Squarewave TiltSound cards, with vanishingly few exceptions, have high pass filters, to block DC. It's normally not an issue - the roll-off starts well below 20 Hz, so the cards pass 20-20KHz audio, and reproduce the audible spectrum perfectly. However, in broadcasting, we care about peak control. Tightly clipped audio (such as the output of an FM audio processor) contains many waveforms resembling squarewaves. Squarewaves thus make excellent test signals when testing low-frequency performance of a signal path.
A typical pro-sumer sound card, such as M Audio Delta Audiophile 192, usually have relatively small capacitors on both the inputs and outputs, due to design constraints and budget reasons. This capacitor may equal a High Pass Filter with 5 Hz cutoff. It passes 20-20KHz audio with no problems - but take a look at what such a filter will do to a 30 Hz squarewave
This is a large amount of tilt. On a modulation monitor, the 100% squarewave would register as 130%! Tilt can be corrected for by a subsonic bass boost, to straighten the signal out again.
This looks pretty good, and if it was this simple, all would be well. However, music is not a constant squarewave - not even today! If we suddenly change the frequency on the fly, which is commonly done in actual music (the bassist playing a different note), other things will happen.
After high-pass filtering, things still look reasonable. However, if you look closely, right at the boundary point (time 0.31), you'll see that the lower edge has de-centered - it has moved upwards. If we apply the same subsonic boost we used to straighen out the squarewave the last time, things get worse.
Even though the individual waves aren't tilting, the whole waveform jumps. This would still be a 10% overshoot. If a modulation monitor registers spurious 10% overshoots when there actually weren't any, it would not be a usable modulation monitor. Luckily, MpxTool contains an (optional) Peak-sensing DC servo. If any part of the signal exceeds 98% modulation, the DC servo looks at positive versus negative peaks, and calculates a subsonic signal and adds it to the audio to cancel the DC offset in real-time. Because it's not a peak limiter, it does not hide any actual modulation overshoots, but rather corrects only for DC errors commonly occuring in receiving / capturing equipment. Here's an example of an Optimod 8500 (Impact Preset) composite output being recorded with a LynxTWO sound card.
This phenomenon is not a fault of the 8500. The Optimod 8500 has DC straight output, and professional FM exciters have DC straight inputs. The 8500 exploits this fact to essentially center assymetrical bass waveforms for an enormous amount of on-air bass. However, if any part of the chain is not DC straight (receivers and sound cards typically are not), the phenomenon seen above will occur.
Here's a different view of the problem - same section viewed on the MpxTool oscilloscope:
|
