# Topic: Sam, if your listening...

1. ## Sam, if your listening...

Any idea how the filter (highpass, lowpass, etc...) works. I can play with it, and get different results. But I have no idea what the different variables actually do. E.G. \"modwheel\"? How does it change the filter demension. \"Sensitivity\"? What does that do?

BTW, I\'m waiting for Bruce Richarson to do a review of the PPP piano. I don\'t want to fork out more money until I know what I\'m getting. It\'s the midrange achilles heel that I\'m worried about. As good as the samples are out there, none has conquered the midrange. All sound just a bit unnatural (except of course that amazing Swedish piano--which is the best sound I\'ve heard .... here\'s hoping).

2. ## Re: Sam, if your listening...

What \"amazing Swedish piano\"?

Lawrence

3. ## Re: Sam, if your listening...

Well for this to make sense you have to have a good understanding what a filter does. A low pass filter has a cutoff point, and lets say that cutoff point is set to 1 KHz. All frequencies in a signal below 1 KHz will be passed with their full intensity (low pass means low frequencies are passed unchanged but high frequencies are reduced, attenuated, filtered, whatever term you want to use). The amount of filtration on higher overtones in the signal is dependent upon the design of the filter and how far they are above the cutoff point. A common filter design reduces signals above the cutoff point by 12 decibels for each octave, so in this case a 2 KHz tone would be reduced to 1/4 the amount in the original signal (each 6 decibel reduction halves the amount of energy) while a 4 KHz tone would be reduced by 24 db.

If you remember Fourier analysis every signal can be reproduced as a series of sinewaves (pure tones) varying in amplitude over time. So a piano playing an A at 440Hz isn\'t just putting out a sinewave, but a signal rich in overtones, most of which are close to multiples of the fundamental frequency. Also, the initial \'attack\' of a piano note (like when the hammer hits the string on a forte strike) will be loaded wiht the most complex series of overtones and also the greatest amount of high frequencies. As the note dies away, the high frequency overtones die out first, and the tone gets simpler and purer.

In the days before gigasampler when we had to loop our piano samples, we would simulate this by looping the sound of the piano, like record the attack for 1 second, but then maybe have the next second repeat over and over (or maybe play back and forth). This would not sound piano like at all without a lowpass filter, where the filter cutoff point is modulated by an envelope, so that the cutoff point might start at 10KHz, but the cutoff point would be modulated down to 100 Hz after maybe 7 seconds. This would reduce the high frequencies in the sound over time and generally make the sample quieter, softer, smoother.

If you want to see what I\'m talking about, use the modified piano patch I described a while ago, record a low note with the mod wheel down, and while it sustains, move the mod wheel up. Now look at the recorded waveform in an editor. With the mod wheel down, the filter cutoff point is low, and the waveform won\'t be to jagged (few high frequencies to move the signal around quickly). With the mod wheel up and the cutoff point higher, there will be lots more sharp edges and fast wigglies (high frequencies) in the signal.

There\'s lots of things that set the filter cutoff point. With high sensitivity, high key velocity will give you a higher cutoff point.

Oops, got to run, I will finish this later tonight!

Hope it makes sense so far...
best wishes
-sam
ps I also enjoyed the latest revisions of your Bach tunes

4. ## Re: Sam, if your listening...

Ah, that\'s what \"sensitivity\" means. I\'m trying to take the boom out of the Steinway Bass below middle C. Initially TOO much was removed, and I couldn\'t seem to adjust the effect more finely. Now I realize why. I moved sensitivity to \"high\". I should try \"low\", initially.

Fourier analysis is a branch of mathematics about which (like almost all mathematics) I know precious little.