[geoff]

On 25/04/2019 1:24 AM, JBI wrote:
On 4/24/19 9:18 AM, JBI wrote:
On 4/23/19 4:35 PM, Mike Rivers wrote:
On 4/23/2019 2:55 PM, JBI wrote:
The background hiss varies throughout the file since the radio
carrier wave became stronger and also faded at times.

"Adaptive" noise reduction seems appropriate. The trick is to apply
it gently as you have, and then repeat until you've reduced the hiss
to a tolerable level or distorted the program material to an
intolerable level. You can't have one without the other.


2)Â* There is a five second beat that occurs within every minute
throughout the entire file (every 50 seconds).

By "beat" do you mean a tone? If you can find a sharp enough notch
filter, you might just be able to filter it out, or get it down to a
tolerable level. If you need to paste over it, be assured that just
about every DAW program does an automatic crossfade at the beginning
and end of a spliced-in section. You can usually adjust the crossfade
time if you need to, and you might, to get the patch to be less
noticeable. I can't tell you how to do that in Audition, but RTRM.
Usually it's a matter of zooming in to the splice point and dragging
a little "handle" at the top or bottom corner of the crossfade area
to change its length and slope.

3)Â* Finally, there is a beat every second that occurs throughout the
entire song.Â* In between the beats, there is one second of
background silence/ noise.

Hmmm . . . . this is beginning to sound like a time signal.



LOL yeah, you got it.Â* What happened was that I got a really good one
hour capture, but you know how the time is announced at the end of
every minute.Â* Well, I wanted two hours of clear capture.Â* The first
hour was actually remarkable.Â* None of the SDR stations I used were
actually close enough for consistent signals, so I've had to depend on
propagation conditions and then got lucky.Â* Their SDR's are user
limited, so I just couldn't start recording and let it go or I would
be cut off unexpectedly.

So that meant, unless I wanted to spend days getting another lucky
day, I wondered how difficult it would be to take a sample of the next
hour and then combine it with the minutes of the good sample I have.
I think this will be possible; it's just to take a lot of time for 60
changes. I had hoped to take the overdub, which would be the same for
every minute.... "At the tone, 16 hours...", and then put the same
spacing between each sample over the hour.Â* I guess I'll just
duplicate it and then carefully move each sample into position.Â* The
good news is that the time stations already have their ticks, making
alignment easy.Â* I just have to do it 60 times.

Just for fun, I've come across another time station out of China
called BPM.Â* It's a lot tougher finding any close SDR's, but again got
lucky. The good news is that there are no announcements per minute,
only on the hour, but I still had to string together a minute of
perfect 125 Hz sub tones, and I had to obtain these from several
captured samples.Â* I did manage to do that all in Audacity and saved
frequently just in case the program froze.

Thanks for your help.Â* I guess I'll fire up Audition and get this done
before my trial expires.

Also, just wanted to add, that I've never been all that impressed with
software noise reduction.Â* I've only ever used it sparingly over the
years, but, more often that not, I've taken near noise free harmonics
from one part of a song, and then mixed them over another part that was
more noisy.Â* That can sometimes be tough and impossible; then I'll turn
to NR but prefer not to use it.Â* One thing that has helped is that the
NR algorithms have improved over the years.Â* I see that Audition's NR is
a lot better than Audacity's, for example, so the latter is what I'll be
using this time.Â* Of course, that may not be fair as I have never tried
the complete plug-ins available for Audacity so there may be better NR
out there somewhere.


Dunno about the others, but is a steady uniform background noise, in
SoundForge with it's NR2 (and prior) modules, you take a 'noise-print'
from an otherwise blank area, and it applies that in an inverse way to
the wanted signal (gross over-simplification there of course).

geoff