ZSuite
Version: 2.0
- 10-11-2023 - Initial Release
- 11-11-2023 - Added Seed Modifier
- 12-11-2023 - Modified Seed node to accept expressions
- 13-11-2023 - Added Latent Noise Provider (Transforming Ambient RF Signals using
RTL_TCPas noise)
ZSuite - Prompter
This node uses files to randomize pre-defined sorted subjects of random things.
Example line:
__preable__ painting by __artist__
This prompt will be processed with random line form the file called preamble.txt in the folder:
.\comfyui\custom_nodes\Zephys\nodes\blocks\.*txt
The node uses a trigger as input from any integer value type, to enforce a new prompt output.
ZSuite - RF Node
!!You do need an actual RTL-SDR device and a shell that can host the service for this node to work!!
The data capture in the RF Node is 4096 bytes / cycle governed by duration in seconds.
If the data length is smaller then the Latent shape it will reset to index 0.
Then the data is processed using numpy to normalize the data in workable format shown here:
# Ensure noise is a NumPy array
noise = np.array(noise)
# Obtain min/max values from noise data
data_min = np.min(noise)
data_max = np.max(noise)
# Scale and inject noise into latent samples
print(f"[ZSuite] Injecting noise...")
for i in range(s["samples"].numel()):
raw = float(strength * float(noise[i % noise.size]))
data_scaled = (raw - data_min) / (data_max - data_min)
data_scaled = 2 * data_scaled - 1
s["samples"].view(-1)[i] = data_scaled
(Prob poorly executed, my bad!)
If you desire to change the default device parameters locate the file:
.\comfyui\custom_nodes\Zephys\nodes\ZS_Rtlsdr.py
And open with a (code/text) editor and look for this constant definition header:
RTL Device Setup
For debian:
https://manpages.debian.org/testing/rtl-sdr/rtl_tcp.1.en.html
Commandline for listening:
user@server:~$ rtl_tcp -a <ip> -p <port> -d 0
After this you setup the RF Noise node to use this ip and port.