For performing correct simulation in MATLAB, which signal will represent a more correct version of acquired data:
Do I need to add noise before quantization of the signal?
Do I need to add noise after quantization?
Or quantized noise must be added?
Or there are other,correct options, which I do not consider?
Thank you in advance!
Well, the noise was present before you ever quantized the signal, hence, the noise should be added before quantization.
Note that, the process of quantatization adds the quantization noise. Hence, if you perform a quantization you do not need to addition any quantization noise, as it is added inherently in this process.
For the other question, if you want to simulate the measurement of a noisy signal, the above answer is the case considered.
The first, but it depends. As it was written above by colleagues, if modeled signal in reality is mixed with some noise (with known statistical properties) you have to add this noise before nonlinear transformation (quantization).
Think about at which stages your signal might become corrupted. For instance, if you will use lossy compression of a quantized signal (like MP3 music), then decompressed signal will become corrupted. To simulate that case you can add noise after quantization.
Without additional information about your concrete application, I would add noise before quantization.
Dear sir
the noise should be added before quantization, and you perform the quantization on the received data or signal.
BR
Thank you all for your responses.
I want to clarify my question.
Signals, which I use are the images, captured by the CCD/CMOS sensors. In addition, sensors are the source of the noise (it is possible that I accidentally do not take into account other sources of noise).
The "bible" for CCDs is "Scientific Charge-Coupled Devices" by J. Janesick. That book is more than you probably want to know but is a good reference to have and is enjoyable to read in sections.
Also many of the instruments on the Hubble Space Telescope use CCDs and there is a wealth of data from user's perspective in the archives of technical reports called Instrument Science Reports (ISRs). For example, yesterday I wanted to know about cross talk between electronic chains on the WFC3 instrument, so I read the section on crosstalk here: http://www.stsci.edu/hst/wfc3/ins_performance/anomalies
Again, that may be more than you wanted to know.
http://www.stsci.edu/hst/wfc3/ins_performance/anomalies
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing