Modifying sim_real to support XCFs and 16-pulse sequence

[egthomas]

As the title suggests, this pull request builds on #521 by adding XCF support to sim_real, which we believe was intended to take a FITACF-format file as input and generate simulated ACFs for each beam and range gate where the quality flag was set to 1 (ie for subsequent testing of different fitting algorithms). I've also added support for the extended 16-pulse sequence that was previously added to make_sim.

One outstanding issue (that hopefully @pasha-ponomarenko can help address) - I do not think the noise_lev parameter is being handled correctly. In AJ's original implementation, he used a plain-text file where some input value of noise_lev was used to scale the lag0 power to assign amp0 at each range gate, eg

fscanf(fitfp,"%d  %lf  %d  %lf  %d  %d  %lf  %lf  %lf  %d\n",&cpid,&freq,&prm->bmnum,&noise_lev,&nave,&lagfr,&dt,&smsep,&rngsep,&nrang);
...
amp0 = noise_lev*pow(10.,(fit->rng[i].p_0/10.));
amp0_arr[i] = amp0;

In AJ's original code he set the noise.search field in the output RAWACF file's parameter block to this input value, so I naively assumed that noise.search from the input FITACF file could be used as noise_lev for the above amplitude scaling. However this does not appear to be the case (at least not for the MSI-style radars like Christmas Valley and Iceland I've tested), causing the subsequent fits to the ACFs and XCFs to fail and the calculated SNR's to be wrong.

[pasha-ponomarenko]

@egthomas , I will try to help with this though I am going to be somewhat busy with other things in the next two to three weeks.

[egthomas]

@pasha-ponomarenko no problem, thank you!