for what type of data would you like to add noise? having trouble finding a usecase where seismic vibrations and soundwaves combine together. Both depend strongly on theyr location, especially soundwaves, are you looking for background soundwaves like random animal and wind in a forest, or something in a city, or really just noise in type of noise on the sensor itself?
#Realistic Noise Generation
worthy sky
worthy sky
understood. So youre looking for something to create "new" Traindata based on the existing pattern and additional noise, correct? if so you could simply add gaussian or white noise to id, decide a signal noise ratio wich looks realistic for your data and create the new Data by simply adding the generated noise to the existing data.
If you want to go for a network to do so, you can use GANs, there you can simply use ur existing data as "label" and let the GAN learn how to mimic them, then you can generate tons of traindata. This sounds really simple, but strongly depends on your data, im not sure if theres a network already trained for "realistic noise on a vibration dataset".