40
The Equivalent Number of Bits in our previous example is no longer 8 but 6 !
Even with an ideal 8 bit Analog-to-Digital converter, the noise in the analog system reduces the
Equivalent Number of Bits (ENOBs) to 6.
8.3.8 How does the signal-to-noise ratio (in dB) relate to Equivalent Number of Bits
(ENOBs)?
A useful formula to convert SNR (in dB) to ENOBs and vice versa is: SNR = 6*N + 2
SNR is the signal-to-noise ratio (in dB) and N is the Equivalent Number Of Bits.
The above formula (SNR = 6*N + 2) can be derived as follows:
The dynamic range (or the "minimum" SNR) of a system with N "true" bits is 2
N
.
The dynamic range (in dB) for a system with N "true" bits is 20 * log
10
(2
N
)
=> SNR = N * 20 * log
10
10(2) = 6.02N; Since 20*log
10
10(2) = 6.02
This is approximately equal to 6*N. An additional 2dB are added due to quantization noise in the
Analog-to-Digital conversion.
=> the minimum SNR for an N bit system should be = 6*N + 2
If N=6, SNR = 6*6 + 2 = 38 dB
If N=7, SNR = 6*7 + 2 = 44 dB
If N=8, SNR = 6*8 + 2 = 50 dB
If N=9, SNR = 6*9 + 2 = 56 dB
If N=10, SNR = 6*10 + 2 = 62 dB
From the above examples, it can be inferred that in order to have a system that is "true" 8 bits,
the SNR of the video signal to be digitized must be > 50 dB. Note: this SNR relates to the signal
AT the Analog-to-Digital converter of the Image Processor (not at the output of the camera) !
The average "security and surveillance" analog camera has 50 dB at the camera output.
By the time the video is processed in the front end circuits of the Image Processor in a noisy
computer environment, it is likely to pick up a considerable amount of noise. A typical signal
at the Analog-to-Digital converter would have 42 dB of SNR ! This would mean that even with an
"ideal" Analog-to-digital converter, the system would have only 6 Effective bits !
It is pretty clear that the only way to "preserve" the ENOBs of the system and make it insensitive
to noise and to system effects that are outside the control of the designer is to digitize the video
within the camera. Once the video data is digitized, the ENOBs will be maintained even in a
noisy system.
8.3.9 Why would I need to control the Gain of the Camera in my application?
In some applications, it is important to "tune" the dynamic range of the camera to that of the
application.
Summary of Contents for DigitEyes Series
Page 14: ...9 Figure 4 3 TC 245 Gate Level Drawing Texas Instruments 1994 ...
Page 61: ...56 12 Appendix D Camera Mechanical Drawings Figure 12 1 Camera Mechanical Drawings ...
Page 63: ...58 Figure 13 2 Camera Noise Spectrum Min Gain Bandwidth 10kHz to 4 2MHz ...
Page 64: ...59 Figure 13 3 Camera Noise Spectrum Max Gain Bandwidth 100kHz to 4 2Mhz ...
Page 65: ...60 Figure 13 4 Camera Noise Spectrum Max Gain Bandwidth 10kHz to full ...