1
0
mirror of https://github.com/f4exb/sdrangel.git synced 2024-11-26 09:48:45 -05:00

DATV demod: CNR adjustments

This commit is contained in:
f4exb 2023-04-01 13:41:49 +02:00
parent 9f71f4d402
commit 5d81a42564
2 changed files with 3 additions and 3 deletions

View File

@ -1895,8 +1895,8 @@ struct cnr_fft : runnable
{
fprintf(stderr, "cnr_fft::cnr_fft: bw: %f FFT: %d\n", bandwidth, fft.size());
if (bandwidth > 0.25) {
fail("cnr_fft::cnr_fft: CNR estimator requires Fsampling > 4x Fsignal");
if (bandwidth != 0.5) {
fail("cnr_fft::cnr_fft: CNR estimator requires Fsampling = 2x Fsymbol");
}
}

View File

@ -48,7 +48,7 @@ This gauge gives the MER estimation. The averaged value appears on the right.
<h4>B.4: CNR estimation</h4>
This gauge gives the CNR estimation. The averaged value appears on the right. Estimation is made comparing spectrum power in the center of the passband compared to the sides. The passband is the one presented to the demodulator and is always twice the symbol rate. the "center" is estimated at 60% of the full passband. Note that the RF filter comes before that and may change the aspect of the spectrum thus the best estimation is obtained when the RF filter width is close to twice the symbol rate and the surroundings are quiet.
This gauge gives the CNR estimation. The averaged value appears on the right. Estimation is made comparing spectrum power in the center of the passband compared to the sides. The passband is the one presented to the demodulator and is always twice the symbol rate. the "center" is estimated at 60% of the full passband. Note that the RF filter comes before that and may change the aspect of the spectrum thus the best estimation is obtained when the RF filter width is close to twice the symbol rate and the surroundings are quiet. This also means that you have to apply the correct symbol rate.
<h4>B.5: Output transport stream to UDP</h4>