lee/hpr-knowledge-base
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Initial commit: HPR Knowledge Base MCP Server

Browse Source 
- MCP server with stdio transport for local use
- Search episodes, transcripts, hosts, and series
- 4,511 episodes with metadata and transcripts
- Data loader with in-memory JSON storage

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
Lee Hanken
2025-10-26 10:54:13 +00:00
commit 7c8efd2228
4494 changed files with 1705541 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

170
hpr_transcripts/hpr3973.txt Normal file
View File

@@ -0,0 +1,170 @@
Episode: 3973
Title: HPR3973: Creating an equalizer preset for your episodes of HPR
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr3973/hpr3973.mp3
Transcribed: 2025-10-25 18:09:46
---
This is Hacker Public Radio Episode 3,973 for Wednesday, 25 October 2023.
Today's show is entitled, Creating an Equalizer Preset for Your Episodes of HPR.
It is part of the series podcasting Houto.
It is hosted by DNT and is about 16 minutes long. It carries a clean flag.
The summary is, a method of creating repeatable processing for your podcasts.
Hello and welcome to another exciting episode of Hacker Public Radio.
I am your host DNT.
I am here to talk a little bit more again about how to record, well not how to record, but how to process audio for an episode of Hacker Public Radio, for example.
This is because I think I came across a really kind of a neat way to do it.
So I would like to share it with you all.
The first thing is, this will work for the same person, the same microphone, recording at the same location, under roughly the same conditions.
This should always be the case anyway if you are going to record voice. That is why people have studios.
Choose your microphone, choose your location, and choose yourself, I guess.
So then what we are going to do is make a preset, I mean we are going to make three presets, one for EQ, one for compressor, and one for normalize I think it was.
And then we are going to save them, and we are going to set a macro that runs all three in sequence.
So then doing your episode of Hacker Public Radio is going to be sit here, record it in Audacity.
Don't bother editing things later, like if you have to cough just hit stop, delete it right away, and then continue recording.
So that you don't have to come back and edit the whole thing because that takes a long time.
So then once you are done you just run that macro, export it as a flat file and send it to the team.
So you are going to use the effect in Audacity that is called EQ.
It is under the EQ and filters section, sorry it is called graphic EQ.
And what that is is several faders, one for each frequency.
It looks kind of like it is like simulating an actual mixing board where you would have actually set frequencies as opposed to having that graph.
That is kind of a pretty common thing in software where you are going to have a graph and then you drop points and you pull them up and down.
And you can also set the kind of curved size.
And then you are supposed to get a very fine green control of the EQ in that way.
But I find it actually kind of unnecessary, this is actually just fine just using the faders.
And there is something kind of nice about the way it mirrors an actual mixing table from way back when.
So take a sample of what you recorded, just a short section and select it in Audacity, then go to the graphic EQ.
Then you are going to start with that flat view where every fader is right in the middle.
And then what you are going to do is just take the first fader, pull it all the way to the top and then hit Preview.
And you are going to listen to your voice.
So at this time it is good to have some decent headphones, like the type of headphones that you would probably use to listen to music.
Because you are going to look for some fine details that you won't be able to detect for example using the headset that I record with.
So then go through the entire set of faders one by one, turn it all the way up, listen to it.
Then go back, click flatten to go back to that flat view where every fader is right in the middle.
Then go on to the second fader, pull it all the way to the top, listen, then flatten, go back.
So you are basically going to listen to your sample with one fader, one frequency fader, pulled all the way to the top.
One at a time, then take a pen and paper and write down the frequencies where there are some really nasty artifacts going on.
So this is going to be things like it is going to kind of make your teeth vibrate, it is going to sound like you are in the bathroom.
That is sort of weird thing is what you are looking for.
Take note of which frequencies that happens in and also take note of how bad it is, how bad it sounds.
So these are the frequencies that we are going to attenuate.
And depending on how bad it sounds, you are going to attenuate it more or less.
Then we are not going to boost anything, frankly it is kind of hard to detect what are good things that are happening in the sound,
escape, so I don't boost anything, I just attenuate and then I boost everything else.
Because you kind of don't want also I think these sharp differences in different frequencies, you know, kind of like with sunlight,
you know, sunlight is all over the entire visible spectrum and artificial lights that are considered poor, one of the things that are common among them,
is that the light is not spread consistently throughout the spectrum.
It is focused on certain areas of the spectrum and that is why they look bad and that is why they feel bad to your eyes.
So I am going to place some samples here and I am going to tell you what frequencies they were of with my voice, with this microphone,
recording to my computer, to audacity.
What were the frequencies where I found nastiness going on?
And hopefully this will survive the compression and you will be able to detect the stuff.
I am sorry to expose you to it by the way because it is kind of uncomfortable.
So then I just took a note of all of them and then I made an EQ setting, a preset,
where I attenuated all the areas then, where this nastiness was happening.
And then I saved it as a preset and then I just applied to my voice.
When I did it the first time I actually got some positive feedback and I actually am going to thank Ken for writing the sample text that I read for these tests that I am going to play here for you.
And just to be clear these samples have the EQ already applied, taking the one frequency that I am going to name before the clip plays and boosting it all the way up.
All right, so let's see.
400 hertz.
The phone ring.
That was odd. I didn't have a phone.
So here I hope you can detect that there is this kind of a sound that accompanies my voice.
That it sounds as if I were talking into a cup or something like that.
Here it is again.
The phone ring. That was odd. I didn't have a phone.
All right, now let's check out 1.6 kilohertz.
A tall blonde walked past my window.
I knew she was tall because I lived on the 6th floor.
So here I start to get something in the background that's getting boosted at this frequency.
There is some kind of a constant sound going on in the background that is getting boosted in this frequency here.
The voice itself is, I think, mostly unaffected, but we'll see that it will get worse as we go up the frequency.
There will be two other areas that have a lot of that.
And then some, here's 2.5 kilohertz.
Walked past my window. I knew she was tall.
So there it is. A lot of that background thing that we saw in 1.6 kilohertz.
And now it's also affecting some of the consonants and stuff.
Like that SH sounded really harsh there.
It almost hurts your ear when it comes at you.
Let me play it again just in case you missed it.
I knew she was tall.
You got that?
Okay, so now let's look at 5K.
5 kilohertz.
That was odd.
So I cut that up to make it a little short and safe sometime.
But yeah, again, you see that the S is really harsh there.
It comes in really uncomfortable.
And then that's why you write it down so that you can attenuate it one more time.
That was odd.
Yeah.
So hopefully you are agree with me that those sounded gross.
And so then I think when you do this, it turns out that you make your EQ attenuating these frequencies that you took note of.
And all of a sudden it sounds pretty good.
And maybe you can't even quite tell why, but it does.
Then you get some positive feedback about your sound.
And so you say, hey, maybe I should record an episode of Hacker Public Radio about how to do that.
So that's how that goes.
So next you will do your compressor.
The compressor will take some of the parts of your recording that are louder, let's say.
And it will kind of turn them down a little bit.
And it does that following a curve, right?
So that the louder you are compared to a certain threshold you give it, the more it will attenuate your voice.
The purpose of that is probably, I think primarily, situations where someone just varies a lot, the their volume.
So you can kind of smooth it out slightly by using the compressor.
So I've done it, the way I've done that recently is I've just kind of tried to do it visually, which has been kind of interesting to me.
So I just kind of zoom out of the whole recording and I take a look at the way you form.
And I just try to see how much variation there is here and there.
And I try to create a compressor preset that kind of puts it pretty even mostly.
It brings it a little bit closer to being even, just a little bit.
And it really doesn't matter too much to be honest.
I think you can skip the compressor if you don't feel like doing that.
But it's kind of a common operation pretty much.
It's one of the things that you will always do with the recording.
So yeah, that's how I've been doing it.
I just kind of do it visually.
Look at the waveform.
And I see, for example, here I can see that some parts are a little quieter than others.
When I apply the compressor and watch how it changes the waveform to see it make the parts that are louder,
bring them a little closer to what the other parts are like.
And so I do that.
I apply it to the entire recording and just watch the waveform.
It's pretty hard to detect what exactly the compressor does using just your ear.
So that's why I think I thought I should try just doing it based on the waveform.
And I think it works all right.
I'm putting a screenshot of the waveform for this recording before and after.
The three presets are applied after the whole macro is run.
So you can see how the waveform looks different.
I'm also going to put in the show notes screenshots of my EQ preset.
And then I will just type in the notes information about the compressor and the amplify presets with the caveat that, again,
you should not expect it to work for you because you have a different voice, likely a different microphone in a different room.
Then the final thing after all that is you would amplify or I guess you can normalize if you want your entire recording.
So normalizing you tell it what you want your peaks to be and then it will amplify your recording to make it so.
I kind of prefer to use amplify because I'm not so sure about that.
I don't know. It's kind of hard to tell.
Nowadays there's not really these guidelines about where your peaks really need to be.
So you can kind of just amplify and see if it sounds.
If you compress your recording so that everything is generally in the same kind of a vicinity,
then you know this is audio. This is like speech, I mean speech podcast.
So you know it's not like an orchestra playing where you would want good definition everywhere.
And you would be expected to be playing this on a really high end thing that would be able to reproduce those sounds at all the different volume levels.
You know this is none of that so it doesn't really matter too much.
That's why you can kind of compress it fairly aggressively I guess.
So yeah I've been just using amplify.
I don't I'm going to put it in the show notes won't by how much I amplify but it really won't matter because it just depends on your voice, the place you're recording.
And the microphone you're using right so you should just try it out on your own.
So yeah you make those three presets then you create a macro that applies them to your audio.
Then you export it and you upload it to hacker public radio.
So that's how I've been doing that's how I've been doing my recordings lately.
And I'm kind of liking liking it.
It's making it simpler for me so this was me describing that to you.
Hopefully it'll be useful to someone.
If not sorry you had to sit through all 15 minutes of this but hey.
Come back tomorrow then and try it again.
There will be another one tomorrow and if it's if you really don't like any of it then just record your own episode and send it into hacker public radio.
All right thanks for tuning in.
You have been listening to hacker public radio at hacker public radio doesn't work.
Today's show was contributed by a HBR listener like yourself.
If you ever thought of recording podcast, you click on our contribute link to find out how easy it means.
Hosting for HBR has been kindly provided by an onsthost.com, the internet archive and our sync.net.
On this advice status, today's show is released under Creative Commons Attribution 4.0 International License.