For reasons not directly connected with this thread I was investigating sweep tones in Audacity - and used the Chirp function in Audacity to produce a sweep from 20 Hz to 20kHz. It's easy to do and might serve as a simple hearing check.
WARNING: If you try this, do not listen on headphones, or with speakers with a powerful amp with the volume turned up. At the low end you may hear nothing - similarly at the hign end.
I wanted also to see if I could put "dents" into the frequency response, and then check them with a spectrum analyser tool. Here at first it looks as though the Audacity spectrum tool doesn't work. However, I suspect it does work - but that there are limitations with using spectrum analysise tools, and also interpreting the results.
I also wondered if I could do the same using Logic, and I found this video which some might find either helpful or interesting - https://youtu.be/nRk1ZKorRR0
My point about spectrum analysis is that it might actually be hard to make a filter which gives perfect rejection or which gives very good approximation to a desired filter curve, and also it may be hard for analysis tools to make displays which fit the theory.
I suspect that theory and practice in this area are not an ideal match.
In the digital domain it is possible to make good filters using z-transforms for both FIR and IIR filters, but checking that they actually work may be hard unless one has good lab kit or good analysis tools.
WARNING: If you try this, do not listen on headphones, or with speakers with a powerful amp with the volume turned up. At the low end you may hear nothing - similarly at the hign end.
I wanted also to see if I could put "dents" into the frequency response, and then check them with a spectrum analyser tool. Here at first it looks as though the Audacity spectrum tool doesn't work. However, I suspect it does work - but that there are limitations with using spectrum analysise tools, and also interpreting the results.
I also wondered if I could do the same using Logic, and I found this video which some might find either helpful or interesting - https://youtu.be/nRk1ZKorRR0
My point about spectrum analysis is that it might actually be hard to make a filter which gives perfect rejection or which gives very good approximation to a desired filter curve, and also it may be hard for analysis tools to make displays which fit the theory.
I suspect that theory and practice in this area are not an ideal match.
In the digital domain it is possible to make good filters using z-transforms for both FIR and IIR filters, but checking that they actually work may be hard unless one has good lab kit or good analysis tools.
Comment