1

I'm working on a research project where we would like to apply convolutional neural networks to an image representation of a signal. However, it seems that if I would use a spectrogram, I would end up loosing the phase information. As a result, my question is:

Does an equivalent transformation of a signal to a spectrogram image exist in which the phase information is part of the resulting image?

Thanks a lot and have a great day, Maxime

From https://en.wikipedia.org/wiki/Spectrogram:

Limitations and resynthesis

From the formula above, it appears that a spectrogram contains no information about the exact, or even approximate, phase of the signal that it represents. For this reason, it is not possible to reverse the process and generate a copy of the original signal from a spectrogram.

Maxime Leclerc
  • 11
  • 1
  • Presumably your signal is not speech or music, where phase carries little or no information ? – Paul R Apr 19 '17 at 14:31
  • Hi Paul, you are right, I am working with hardware radio signal identification. – Maxime Leclerc Apr 19 '17 at 14:33
  • OK - interesting - I think you might get a better response at http://dsp.stackexchange.com though - SO is really just for programming questions, whereas DSP.SE is more about DSP theory, and there are a lot of very knowledgeable people there. – Paul R Apr 19 '17 at 14:36
  • Sounds good I'll try DSP.SE – Maxime Leclerc Apr 19 '17 at 14:46

0 Answers0