hpr-knowledge-base/hpr_transcripts/hpr3673.txt

Episode: 3673
Title: HPR3673: Recording for Hacker Public Radio
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr3673/hpr3673.mp3
Transcribed: 2025-10-25 03:21:44

---

This is Hacker Public Radio Episode 3673 for Wednesday, the 31st of August 2022.
Today's show is entitled, Recording for Hacker Public Radio.
It is part of the series podcasting how to.
It is hosted by DNT, and is about 18 minutes long.
It carries a clean flag.
The summary is, My Experiences Recording Episodes of HPR.
Welcome to another exciting episode of Hacker Public Radio.
This is your host, DNT.
This one is prompted by an email conversation with Carl.
He wrote to a few of us sending some samples of some recordings that he had made of himself,
and he wanted to get some feedback on it.
Carl, you may remember, he was the one that put together that collaborative episode with
some shell snippets, and I sent a clip to that show.
In Carl's email, he mentioned that he thought my clip sounded pretty good, and so he thought
I might have something to say about recording audio.
So that was pretty funny because of how I actually recorded that clip.
So it was different than any previous episode of Hacker Public Radio that I recorded, because
I was at work, and there were only two days in which to submit the clip, because Carl
had helpfully set a deadline for this.
So I had only two days at work, and I didn't have anything to record it with.
So I went to the Play Store looking for an app I could install, and for simplicity, I
just looked for something I recognized.
I was hoping there would be an app from like Zoom or Road or something like that, one
of the audio equipment manufacturers.
And as it turned out, I found one that was called Dolby on.
So that was how I recorded that episode, and I agree with Carl, it sounded quite good.
It sounded a lot better than any of the prior episodes that I had edited myself.
So clearly the people at Dolby are cooking some fancy filters there without telling us
about it, and it's doing a pretty good job, it seems.
So before that, I had recorded HPR episodes in a few different ways.
The first one I think I recorded it in an iPhone 7 with the earbuds that come with the
iPhone 7, then later I think there was at least one that I recorded using a Zoom H1,
which is a small handheld recorder, one of those that have like an X, Y pair of microphones.
It's like a stereo.
It's really nice for recording ambient sound, for recording voice.
It picks up a lot of background noise, anyway.
Then I used the application that was shared by Norrist here on Hacker Public Radio, it's
called SoloCast, it's a Python application.
On the command line you prepare the script, and then the application will show you the script
one paragraph at a time, you record the paragraph, and then it'll ask you if you like it or
if you want to re-record it.
Once you have recorded all of your segments, it'll splice it all together for you and
do some normalization in a couple other things, I guess.
So I used that with this headset that I use at work for calls, that's called it's the
Sennheiser MbPro2.
It's decent, it's pretty convenient, but it doesn't sound very good, I guess.
Not too good.
And then finally this Dolby on app, and then here I am recording the second episode using
this app.
It's not free, of course, so leave your comments on that if you want or record your own show
about this use of non-free software and what it means, I want to hear your opinion on that.
So then going back in time a little, so I went to college for film, and during that there
was, I had one class that was about sound, and I also paid a lot of attention to sound
recording and sound editing as well for my own interest, but during this class we learned
a couple of things about editing sound.
First mainly we learned about the compressor and EQ, so let's talk about the compressor first.
The compressor is, you usually visualize it with this graph, so you're usually going
to have a setting that's called the threshold, it's going to be in decibels, like for example
I set the threshold to 16 decibels, so what that means is up to 16 decibels, the filter's
not going to do anything to it, and any sound that is above 16 decibels it's going to
attenuate.
It's going to attenuate following another setting that's called the ratio, that's going
to be something like 4, 1, for example.
So the ratio is how much you're going to attenuate that sound, that's above the threshold,
so if the ratio is very, very high, that means you're practically going to flatten anything
that goes above 16, and you're going to take it right down to 16.
What you're doing with that is if you're someone who's volume, who's speaking volume varies
a lot, it's going to kind of even that out a little.
You can imagine, for example, if you're recording someone playing the piano and they tend
to hit the keys really hard sometimes, which in person it can sound very good, but to
a microphone it can be very challenging, so the compressor is one of the things that
can help with that.
The compressor usually also has a setting that's called gain, that's basically just gain,
and it's just there because typically you're going to attenuate the top of the peaks there,
and you're typically going to want to lift up everything after that to bring it up closer
to the peaks you want.
You shouldn't overdo it with the compressor because I think it can make things a little
monotonous if the volume is kind of always the same, so it can make it sound less natural
or something like that, so you don't want to overdo it, but certainly I think with voice
over with podcasts, with you can get away with quite a bit of compressor I think, if
you disagree just let me know in the comments.
So at this class in college, honestly the professor never really talked about normalization
if I can remember it right, really I think what we would do is we would just try to get
just with the compressor alone, we would try to get as close to the peaks that we are
targeting as we can, and then we would just do the normalize would be at the very end.
So the getting to the peaks that you're targeting, what that means is like I think nowadays
with online stuff, you can easily go to minus 3, minus 1, and that's just fine.
You may hear stuff about peaks having to be at minus 12 or minus 16.
I think that's really kind of a technical standard to do with broadcast video and broadcast
radio, I don't think it's really relevant anymore, so you can go a lot higher than that
and most people I think do.
So you would try to get as close to that as you can with the compressor and then use
the normalize to get right there, then EQ.
So the professor taught us this interesting technique, I think he called it the sweep.
So the EQ, let's imagine a graphical equalizer, and you have the spectrum right there in front
of you on the left are the highest, the lowest frequencies, that is the low pitch stuff,
and on the right are the high frequencies, the high pitch stuff.
So then typically you would have a line going from left to right, and with the graphical
EQ you're going to mark some points there at different frequencies, and then you're going
to attenuate or boost the signal at those frequencies.
So what that's going to do is, for example, if you put some points here on the left and
then attenuate them a lot, your voice is going to sound less low pitched.
Here is also a setting that I don't remember where it's called, but it kind of makes
the curve more closed or more open so that if you boost the signal at one point and you
make the curve more open, you're going to be boosting a lot of the neighboring frequencies
right there.
So the sweep that this professor taught us, what it was, is you would start on the left
of the spectrum, you would mark a point, make the curve pretty closed, and then you would
boost it quite a lot, and then loop through a piece of your recording, maybe pick some
part that you think is kind of representative for whatever reason, just play it in a loop,
and then move the frequency of your point from the left all the way to the highest frequencies,
and listen as you do it.
So as you listen, the sound of course is going to change because you're going to be boosting
all these different frequencies and you may notice some things.
Keep your ear out for parts that sound really kind of nasal, just kind of annoying or something.
Just whenever it sounds really terrible, when the signal is boosted, stop and take a note
of what frequency that was, okay, and then keep going, then when you get to the end of
the spectrum, you will have written down maybe like three frequencies or something like
that, so then drop points at those three frequencies, and then attenuate them a little,
and this time use a much more open curve, and then see how that goes, and then now it's
going to be trial and error, just try a few different things here and there and see how
it sounds.
I think you would expect for it to sound more pleasant to the ear if you just do that.
Now there's also boosting the signal at different points.
I don't know much about that to be honest.
I know there are areas of the spectrum that you can boost that can help your voice somehow,
but I don't really know how exactly, so I don't think I'm going to be very helpful there.
What you can do also is just look for some presets in your editing application that are
named for things like voice over or vocal, whatever it is, and see what they do, and then
just try those things out with your own EQ for your voice, right?
So basically just do this once for your own voice, for your environment that you're going
to be recording in the typical environment, you're going to want to keep consistent,
and then just save it as a preset for your voice, so then you can just apply it to your
recordings going forward.
Now a kind of a thing I would say too is EQ is pretty subjective as is, I think, sound
editing largely, I think, so I think the only thing you can do is kind of lean into it
and just see what things sound like to you and see if it feels good or not, and just
go by that, you know, that's all you can do.
So then it's easy also to get for your ear to kind of get used to sounds.
So a lot of times it's worth doing this and then stopping, putting it away, and then
come back the next day, listen to it again, see if you still think it sounds good, a lot
of times you won't anymore for whatever reason, so it's worth doing that.
Now I talked about you're setting the EQ for you're saving a preset for your environment
that you're going to want to keep consistent.
Now this is much more important than the EQ and the compressor, to be honest, if you do
no EQ or compressor, but your environment is great, your recording is going to be great.
If you do, if you spend 24 hours on EQ and all the other stuff, noise removal, whatever,
but the environment in which you recorded was crap, it's not going to sound any good,
no matter what you do.
So you want to look for a place that it's as little noise as possible and as little reverberation
as possible.
It comes from really from flat surfaces, so if your room has a lot of flat surfaces
in like the walls, for example, and no furniture, sound is going to bounce back and forth
a lot, there's going to be a lot of reverberation.
This is the thing that's going to sound the worst in a recording, so it should be the
number one thing you consider, or maybe not the number one thing, but one of the top things
you consider.
So a carpeted room, a room with more furniture, stuff on the walls will eliminate the reverberation
really quickly, actually.
Of course there's also equipment you can buy for that if you want.
So one thing that a lot of people will do is they'll just go in a closet and record
in a closet.
Guess it's pretty warm, but it works very well.
There's lots of clothes in there, so it's going to sound pretty much like you're in
a sound studio.
Another thing is the car, like the clip that I recorded for Carl's show I recorded in
the car.
Cars, car manufacturers have made cars a pretty hospitable environment for sound, so if you
have a car, it's worth trying that out too.
And then finally the time of day, pretty obvious, you know, for example, don't record it
when your sprinklers come on and then you can hear the water flowing through under the
house.
And breathing and warming up really important things too.
This is something I've noticed in myself for sure.
Whenever I record an episode, I feel like when I start I'm kind of nervous, I'm not
breathing right and it doesn't sound good at all, it's like embarrassing.
And then after a while I start to relax into it and I start sounding a lot better, I'm
talking a little more slowly, I'm breathing naturally, things are going great.
So I think the reason for that is because I have warmed up by then.
So what I've been trying to do is just warm up ahead of time.
In this case, for example, I prepared a bunch of three by five cards with a sort of a script
and then I just read through it on my own without the recorder even running.
I just read through the whole thing trying to get to this state before I actually start
recording.
So try that out warm up before you record.
Then I mentioned the three by five cards.
So when we were talking to, when I was talking to Carl, I mentioned how back in the day,
there used to be apparently a phone number you could call and you could just start talking
and it would record it and it would make it into an episode of Hacker Public Radio later.
So people would do that.
They would just call a phone number and in one take, they would record an episode of
Hacker Public Radio.
I think to me, that's the standard we need to strive for here because in this world, it's
pretty hard to make time for things like this.
You work full time and you have your family to take care of a house, whatever else is going
on.
There's not a whole lot of time left for making art or making stuff for recording episode
of Hacker Public Radio.
So if you can do it in one take, if that's possible, at all, I want to investigate that
and I want to figure out how I can do that, you know?
So what I'm trying here today is these three by five cards.
So I made them and then I read through them through one time without recording and then
I went back to the beginning and started recording it.
And it's pretty good.
I think I'm still going to have to take it to audacity, but not nearly for as much as
I usually do.
So I think this episode will be finished much sooner than most episodes have.
Most of my episodes anyway.
Yeah, I listened to an episode by this Mr. Gadgets earlier today.
And yeah, he just called in and started talking and 15 minutes later he had a show.
Anyway, that was pretty much all that I have.
In closing, I just want to say that none of this that I've talked about matters really
as much as just having something good to say.
Now that may seem hard to have something good to say, but here's the thing you do have
something good to say.
And the easiest way to do it is to just press record with your app or your recorder or
the Dolby on app, if that's what you want to do.
Press record and then say, hi, I'm a new HPR host and here's how I got here.
All right.
Thanks for tuning in and come back tomorrow for another episode of Hacker Public Radio.
Bye bye.
You have been listening to Hacker Public Radio at Hacker Public Radio does work.
Today's show was contributed by a HPR listener like yourself.
If you ever thought of recording a podcast, click on our contribute link to find out how
easy it needs.
HPR has been kindly provided by an honesthost.com, the internet archive and our
things.net.
On the Saldois status, today's show is released under Creative Commons Attribution 4.0 International
License.