degrade~ bit range
Hello,
I was quite amazed that Max already had built-in beat crusher that sounds pretty decent. The "degrade~" object.
It works fine, but I'm just curious about its rate and bit depth relation.
So far, whatever sample I've played through, the bit depth doesn't do anything significant over the value of 4.00. Somehow in the help file, it says that the range should be 1 to 24. The funny thing is that it seems that the help file's own "Tone Sweep" sound file itself, does not do anything particular when bit depth over 4.
What is all about between 4 and 24?
What's the mystery in that range?
Thanks.
well a sine of 8 khz will still sound the same when the samplingrate is 44100 divided by 4.
Are we talking bit-reduction or sampling-rate reduction here? Sampling rate is all about Nyquist frequencies innit... Bit reduction however depends on the interpolation change from one sample to the next e.g. a square will sound no different but a sine or saw will sound remarkably different
I noticed that as well. It's cool but does not do much till about 6 or so. You could try Soundhacks +Decimate~ object which has a lot more crunch.
It is possible to design your own bit depth reducer by multiplying with a certain (high) value, rounding to whole numbers, and dividing again by the same value. The range and curvature of the control of this value is something that can be of your own design. Therefor it can be made more intuitive.
@jvkr I've thought of this too - before I discovered [degrade~] actually - and I mused that since it involves converting data from the MSP to the Max domain and then back again, that the numbers probably wouldn't tally up that well and you'd probably get more than you bargained for on the audio crunching front!!
As a Windows user I don't have Soundhack which is very upsetting as it does many wonderful things for me that the Cubase plugins can't. Trying to reinvent a few within Max but the results are less than acceptable. If anyone knows of a Soundhack equivelent for Windows I'll be all over it :)
@ mr_mapes
That link I posted is to free SoundHack externals for Win/Mac/Linux. There are 7 of them +binaural, +morphfilter, +spectralcompand, +spectralgate, +decimate, +chebyshev, +matrix, +compand, +delay, +pitchdelay and +bubbler - good fun.
A thousand thankyous mr grizzle!
> it involves converting data from the MSP to the Max domain
That's not correct. Anyway, it can be achieved simpler than what I said before, without multiplication and division.
I'd never heard of or used [round~] before. Nice.
Just to be pedantic...
The sampling rate determines the range of frequencies that can be represented (0 Hz to the Nyquist frequency). Reducing the sampling rate by half, down to 24,000Hz or 22,050Hz, won't have much audible effect in many cases because some sounds have relatively little energy in the uppermost octave of our hearing range. Reducing the sampling rate much lower will of course produce more noticeable aliasing.
The bit precision determines the signal-to-quantization-noise ratio. Reducing the bit precision to 16 will in most cases be imperceptible because it's still "CD quality". Reducing it to 8 will produce a noise level that is clearly audible, but it's just white noise and thus not very interesting as "distortion". It's only when you get down to those ridiculously low bit depths that the artifacts have more character.
See "How Digital Audio Works" in the MSP Documentation (search for 'Nyquist' and 'precision') for more.
One gets used to the facebook interface, and finds themself wishing there was a 'like' button. Excellent explanation, Mr Dobrian.
@Orkester
this forum seriously needs a 'like' or even 'prioritize' button; succinct, concise and vital explanations such as this deserve to be pushed to the front of the forum queue; to prevent n00bs (guilty) of asking questions that have already been answered. Dreaming...
Brendan