groundhog1
Electrical
- May 4, 2003
- 43
I am learning about HDTV communications right now and am stumped on something. I am trying to imagine what the signal spectrum looks like for the uncompressed serial data stream.
The High Definition data rate is apparently 1485 Mhz.
Then I read that it is NRZI (non-return to zero inverted).
Then I read that if you have all ones, you get the max signal rate since a one is a transition and a zero is when the signal stays the same. And, with NRZI, for every clock cycle you get one NRZI cycle because the NRZI output only transitions on the falling edge of the clock (data is read on rising clock edge). So you get NRZI high on the first fall edge then get NRZI low on the next clock falling edge, etc. So the NRZI signal is 1/2 the clock freq.
So if you assume a string of ones, you basically get a pulse train. Then imagine a sinc (sinx/x) shape of lines. The first line will be at DC. The second line will be at ??. THe null will be on the next 3rd spectral line (assuming pulse width is same as pulse space.
If the data stream is actually at 1485 and the clock at 2970 Mhz, then the second line will be at 1485 and the first sinc null will be at 2970 Mhz.
If the data stream is actually 742.5 Mhz and the clock at 1485 Mhz, then the second line will be at 742.5 and the null at 1485 Mhz.
So which is it? Or is it something else?
If you try to imagine the other extreme, a string of zeros (which I think is prevented), you get complete DC.
Isn't the max string of any one state like 20 ones or zeros in a row?
Thanks,
groundhog
The High Definition data rate is apparently 1485 Mhz.
Then I read that it is NRZI (non-return to zero inverted).
Then I read that if you have all ones, you get the max signal rate since a one is a transition and a zero is when the signal stays the same. And, with NRZI, for every clock cycle you get one NRZI cycle because the NRZI output only transitions on the falling edge of the clock (data is read on rising clock edge). So you get NRZI high on the first fall edge then get NRZI low on the next clock falling edge, etc. So the NRZI signal is 1/2 the clock freq.
So if you assume a string of ones, you basically get a pulse train. Then imagine a sinc (sinx/x) shape of lines. The first line will be at DC. The second line will be at ??. THe null will be on the next 3rd spectral line (assuming pulse width is same as pulse space.
If the data stream is actually at 1485 and the clock at 2970 Mhz, then the second line will be at 1485 and the first sinc null will be at 2970 Mhz.
If the data stream is actually 742.5 Mhz and the clock at 1485 Mhz, then the second line will be at 742.5 and the null at 1485 Mhz.
So which is it? Or is it something else?
If you try to imagine the other extreme, a string of zeros (which I think is prevented), you get complete DC.
Isn't the max string of any one state like 20 ones or zeros in a row?
Thanks,
groundhog
