While reviewing the source code provided on your GitHub repository, specifically in SubspaceNet-main/src/signal_creation.py, I encountered a point of confusion regarding the signal scaling.
I noticed the following line of code:
amplitude = 10 ** (self.params.snr / 10)
In standard signal processing definitions, SNR is typically defined as a power ratio. Therefore, when converting to amplitude, the factor should usually be 20 rather than 10, which would lead to the formula:
amplitude = 10 ** (self.params.snr / 20)
I am concerned that this might affect the consistency of the results during reproduction. Could you please clarify if the use of / 10 is intentional for this specific implementation, or if it might be a minor discrepancy?
Thank you for your time and for sharing your valuable research with the community. I look forward to hearing from you.
While reviewing the source code provided on your GitHub repository, specifically in SubspaceNet-main/src/signal_creation.py, I encountered a point of confusion regarding the signal scaling.
I noticed the following line of code:
amplitude = 10 ** (self.params.snr / 10)
In standard signal processing definitions, SNR is typically defined as a power ratio. Therefore, when converting to amplitude, the factor should usually be 20 rather than 10, which would lead to the formula:
amplitude = 10 ** (self.params.snr / 20)
I am concerned that this might affect the consistency of the results during reproduction. Could you please clarify if the use of / 10 is intentional for this specific implementation, or if it might be a minor discrepancy?
Thank you for your time and for sharing your valuable research with the community. I look forward to hearing from you.