Hello

I am using raw amplitude and frequency (env~ and fiddle~ respectively) to make real time adjustments to images. My problem is that the changes being made are either to drastic in a very short amount of time or hardly any change at all. So what I end up with is almost a strobe light effect or the changes are so minimal it is almost not noticeable.

The image sources are in jpeg

The audio sources are a microphone and a wav file

I am feeding the data into these objects to manipulate the pictures:

pix_lumaoffset
pix_convolve 3 3
pix_contrast
pix_bitmask

and possibly others...

Is there any way to take a value range of 0.1 to 3.0 and exaggerate the top half of the value range. So that lets say 1.5 to 3.0 seem to be far greater then they really are after being scaled? Finally I would like everything below the threshold of around 1.5 to be literally 0.1 to 1.5.

Thank you!

http://www.pdpatchrepo.info/hurleur/Audio_Visual_installation_2.pd

Hello,

I am trying to work out a neural network in pd. I am a bit overwhelmed by some of the math and how you would enter it in the [expr] object. I have presumably figured out how to get the error of a network segment using the mean squared error (http://en.wikipedia.org/wiki/Mean_squared_error) which is shown in the patch.

My struggles come when I try to get the output from the network using the sigmoid function (

If anyone can help please do.

Thanks guys!

http://www.pdpatchrepo.info/hurleur/Neural_network.pd

Hello, this is my first post. I am pretty new to pure data.

What I would like to do is have a microphone input for my saxophone and a second one for either cassette or reel to reel. I want to let the tape which is pre-recorded with saturation, hiss and noise, go for the duration of a live tune. The kicker is that I want the tape to only be heard by the audience while a certain volume level is reached by my horn (like a gate). In this way the tape acts like an effect pedal to the saxophone because it is dependent on input from the horn.

I understand (somewhat) that a patch would have to include something like this:

adc~ adc~

*~

env~

dac~

Please help if anyone knows what I'm talking about.

AudioLemon,

Thank you so very much! I was able to achieve the exact effect I am looking for.

You just helped me get around a huge roadblock to my project.

I was able to get an exponential curve going to the input data however the changes were still not drastic enough. I resolved this by creating a second

[mapping/exponential_curve] object to control the first one's

double-exponential seat control parameter using the same input data, from the RMS Amplitude of an audio source.

Thank you again!,

joeltrane

Thank you Caucasian and lead i will try this!

Forgive me for being impatient but can someone please point me in the right direction?