HPR4608: Simple Podcasting - Episode 1 - Preparation and Recording
This show has been flagged as Clean by the host.
Simple-Podcasting
01 Introduction
This is the first episode in a four part series on a simple way to create your own HPR podcast episode.
02
If it sounds contradictory to have four episodes on a simple subject, you only actually need the first episode to see how to create podcasts.
The remaining episodes are on steadily more complex subjects, with the later ones being more in the realm of gratuitous hackery for the fun of it.
03
I am fairly new to podcasting. I have done an HPR episode on Oathtool, another on the UCSD P-System, and an 8 part series on nuclear power.
Prior to this I have never done a podcast before.
Despite that, a number of people wrote into HPR to say that they really liked what I did.
This means that you too can make a first podcast and have other people find it very interesting.
04
Since I am fairly new to this, I thought I would document how I went about it for the benefit of anyone who wants to do the same.
This describes things from the perspective of someone who is very new to this sort of thing.
Later on I will get into some more advanced topics and then finish off with some blatant gratuitous hackery like how to use Libre Office Calc or GNU Octave in place of an audio editor for some things.
05 Initial Hurdles
There were several hurdles to get over before I could record an episode though.
The most obvious one to me was that I'm not the sort of person who can simply babble into a microphone.
That meant that I needed to have a way of recording things that would let me exclude pauses and repeat sentences that I had messed up.
06
However, since I was new to podcasting, I didn't know how to use an audio editor such as Audacity.
After a bit of thinking though I came up with a very simple solution to that which I will get into a bit later in this episode.
07 Picking a Topic
With the recording process solved, the next thing to do was to find something to talk about.
The key to this is to have some place to keep notes.
I use a note taking program for this, called Zim.
08
There are other programs which do something similar, but Zim is the one that I use.
Whenever I came up with an idea of a topic, I would add a note for it.
Whenever I came across any information relevant to one of the topics, I added it to the note.
09
You might think that you don't know of anything interesting, but the fact is that a lot of the rest of us are fairly sad individuals who are just as boring as you are and so find things like verbal tours through obsolete and obscure operating systems to be quite fascinating.
10
I am sure that you too know something obscure but equally interesting.
11 Writing a Script
Once you have a topic, the next thing is to write a script.
If you are good at talking off the cuff, then all you may need is an outline.
If you are like me however, then you will need to write down exactly what you are going to say in a way which you can read back later.
12
In this case, start with an outline and fill in the detail after the outline is written.
Again, I use Zim for writing my scripts.
It provides a simple way of organizing my scripts as I am putting them together.
It also provides character and word counts so I can estimate how many minutes of material that I have.
When I started I decided that I should target about 10 to 20 minutes for the length of an episode.
That's a personal decision and not something you need to follow for yourself, but it gives me a guideline to work to.
13
As a rule of thumb I find that if I multiply the character count by 0.0011, that gives me the approximate number of minutes of audio when recorded.
Your own number may differ from this, but it's a good starting point to work from.
If you think the episode is getting too long, don't worry. You can split it up into multiple episodes.
14
Once you have the script written and have, if necessary, split it into separate episodes, start numbering the paragraphs.
This is related to the recording method, which I will go into more detail later.
15
Each paragraph or section should be equivalent to 30 seconds to a minute of audio.
If you are just starting out in podcasting, this may be roughly how much you are comfortable with recording without pausing to collect your thoughts or stumbling over what you are saying.
We will knit these sections together with a very simple bit of software later.
16 Recording Equipment
You will need some sort of recording equipment.
While some people may talk about using a phone or an MP3 player with record function, or something like that, I'll stick with recording onto a PC.
17
My recording equipment consists of a Maxwell headset with headphones, boom mic, and USB connection.
There is no part number on it and can't identify it further than that.
The cost was probably around $20.
Similar ones sell for $5 to $35, depending on where you buy it.
I already had this, so I didn't have to go out and buy it when I decided to make a podcast.
18
A boom mic, that is a microphone that is on an arm attached to the headset, is good because it keeps the microphone at a consistent distance from your mouth without any effort.
19
The disadvantage of the particular model that I have is that there is noise in the signal, which you can hear in my first two podcast episodes.
Despite the noise, people still liked the episodes so don't get too hung up on audio quality.
I will talk later about how to fix noise issues like this by filtering.
However at this point I am just going to stick to the basics.
20 Recording Software
For recording software, I used Gnome Record on Ubuntu.
This is licensed under GPLv2 or later.
21
Is very basic
The only options are to select the file format, and select stereo or mono
The sample rate for flac is fixed at 44.10 kHz, which is what HPR wants.
22
If you are using different software, possibly on another operating system, the principles are the same.
There are probably equivalents which you can find if you look for them.
Perhaps you or other listeners could make an HPR episode recommending one.
23
When using Gnome Record, use the menu located in the upper right of the window bar, which has three small horizontal lines as an icon.
Set the preferred format to FLAC.
Set the audio channel to mono.
24 Recording
Get comfortable at your desk.
Get a cup of tea ready as your throat may get dry.
Set up the hardware.
25
If using a boom mic on a headset, adjust the mic so that it is roughly at chin level.
Avoid putting a boom microphone directly in front of your mouth. You should speak over the top of the boom microphone, not directly at it. This will prevent you from breathing on the microphone, causing noise problems.
26
If you have a different type of microphone, you may have to experiment a bit using short test recordings to find the optimal position.
27
Using your recording software, make a test recording and listen to it.
If it is too quiet and the input volume is already up all the way, we can adjust this later with software.
28
If the test recording sounds OK though, then you are ready to start.
29 Recording using Gnome Sound Recorder
I will now describe how to use Gnome Sound Recorder.
If you are using different software the details may be different, but the basic principles should be similar.
Using the mouse, click on the "Record" button.
It will start recording, showing the waveform of the recording as it goes.
30
To stop recording, click on the square "stop" icon that appeared at the bottom.
Give the recording a name, using a numbering system starting at 01.
To accept the recording, click on the check mark button on the right.
To save the recording, click on the down pointing arrow on the right.
31
The file name will default to the name of the recording which we just gave it.
The numbers should match the paragraph numbers in your script.
The recording will be saved as a flac file in your home directory. there is no option to save it anywhere else and you will need to move it to your preferred destination manually.
32
You can now delete the copy of the recording which Sound Recorder keeps by clicking on the garbage can on the left. This does not affect the copy on your disk. You will want to delete these extra copies as you go along, as there's no easy way to do this later and an extra copy of the recordings will accumulate in a dot directory somewhere and take up space.
33
If you make a mistake or are otherwise dissatisfied with that paragraph, just delete the file and record it again.
Keep the pauses at the start and end of each audio segment equivalent to normal pauses between words. This is actually fairly easy to do.
When you are done you may have anywhere between 2 and 4 dozen separate flac files.
34 Using the keyboard shortcuts with Gnome Sound Recorder
Here are the two most useful keyboard shortcuts for Gnome Sound Recorder.
Press Ctrl - R to start recording.
Press "S" to stop recording.
35
You still need to use the mouse to click on the check mark button to accept the recording.
There are supposedly keyboard shortcuts to save the recording to disk and to delete the recording, but these don't seem to work, at least not in version 43.beta on Ubuntu 24.04
Starting and stopping via keyboard shortcuts is still useful however.
36
You can use other software, and I will talk later in another episode about using command line software such as ffmpeg to record.
37 Tips on Recording
If you are new to podcasting or just are not good at making long speeches, keep each recording segment short, a minute or less being a good target.
If you stumble over what you are trying to say, don't worry, just repeat the recording for that section.
38
Talk clearly in even, measured tones at a reasonably constant volume.
Remember who your audience are.
They are people who are listening to your podcast while they are doing housework, or gardening, or driving a car, or walking down a street, or taking some exercise.
Very few will be sitting at a desk in a quiet room like you are when you are recording.
39
Try to make sure that what you are saying comes across clearly.
If the loudness of your voice varies too much, they won't be able to hear you in the quiet parts.
The worst thing to do is to trail off into an imperceptible mumble at the end of each sentence.
Listeners will not be able to follow you if you do that and may give up trying to listen to your episode.
40 HPR Audio File Requirements
HPR episodes are mono, not stereo.
If you send in a stereo file, they will convert to mono.
However, you may wish to convert to mono yourself for the purposes of better duplicating the final result when you review your own work.
41
The easiest way to create a mono recording is to record it as mono in the first place, if your recording software has this option.
If you are using Gnome Sound Recorder, there is a setting for this.
I described how to set Gnome Sound Recorder to mono just a few moments ago.
42
If your software doesn't have a mono option, or if you have already recorded it and now wish to convert to mono, you can use ffmpeg to do the conversion.
ffmpeg -i stereosample.flac -ac 1 monofile.flac
43 Alternatives to Gnome Sound Recorder
An alternative to Gnome Sound Recorder is KDE Recorder.
This is also available as a snap on Ubuntu.
The license is GPL-2.0-or-later.
44
However, I found it to be a bit more difficult to use than Gnome Sound Recorder.
Selecting the microphone source was difficult.
It shows several sources rather than just taking what the OS says is standard.
45
This may be because it is a KDE app running on Gnome. Perhaps this is easier if you are using KDE.
Every time I unplugged my headset and plugged it back in, it added more audio sources to its list.
None of them worked however until I selected the correct one, exited the program, and then started it back up.
46
All files are saved to the Music directory, there is no choice offered.
Audio format selection is more difficult, it being a two step process.
There was no option for mono recordings, only stereo.
Flac recordings were 48 kHz rather than HPR's preferred 44.1 kHz. Converting this would require more post-processing using audio software to change it.
47
It seems to offer no advantages over Gnome Sound Recorder on Ubuntu, while making selecting sound sources more difficult.
If you are using a Gnome desktop you are better off with Gnome Sound Recorder.
However, it is all a matter of personal preference, and if you find that you like KDE Recorder better, then go ahead and use it.
48 Other Alternatives
There are of course still other alternatives.
Many people recommend using Audacity to record.
However, I don't know how to do that, and the premise of this podcast episode is that you have something you would like to make an HPR episode but are put off by the difficulty of learning how to do so.
49
However, Audacity is a very capable audio program and I have nothing against it.
I will describe how to use a feature in Audacity to help overcome an audio problem that I encountered, but I will save that for another episode.
50
As well as GUI programs, there are also programs which allow you to record audio from the command line.
I will describe a couple of these in another episode.
If you are wondering why you may want to use a command line program for this purpose, one of the advantages of this is that it lets us write scripts which automate the recording process and eliminate some of the manual steps that I outlined above.
51 Combining the Segments into a Single Audio File
At this point you will have recorded your podcast episode as a series of several dozen flac files.
We will now stitch the separate flac files into a single file.
We do this using either ffmpeg or sox.
52 FFMPEG and Sox
FFMPEG is a set of command line programs for converting and manipulating audio and video files.
Various parts are licensed under the LGPL V 2.1 or later, and GPL v 2 or later.
53
Sox is also a set of command line programs, but for audio only.
Sox stands for Sound Exchange.
Sox is licensed under similar terms as FFMPEG.
In fact Sox actually uses FFMPEG for certain operations.
54
For our purposes here, with one exception you can do everything audio related with either FFMPEG or Sox, except for one thing which I will get to in another episode. That one is related to doing some gratuitous hackery when analyzing audio files, so it may be irrelevant to anything you need to do.
Since the two are more or less equivalent for our purposes, I will provide examples using both.
55 Combining Audio Segments
The best way to combine multiple audio segments is with a simple shell script.
A copy of this will be in the show notes.
56 Using FFMPEG
Doing this with FFMPEG requires just two lines.
# First create the list file.
printf "file '%s'\n" [0-9][0-9].flac > podseglist.txt
57
This first line creates a file called "podseglist.txt" which contains a list of all the two digit numbered flac files in the current directory, together with some other necessary text.
I have assumed you wish to give these files two digits. Add more digits out if you feel this is necessary.
The file name "podseglist.txt" is purely arbitrary and you can use whatever name you wish.
58
Now we need to concatenate the files.
# Now concatenate them
ffmpeg -f concat -safe 0 -i podseglist.txt fullpod.flac
The second line calls ffmpeg, tells it to perform a concatenation operation, turning the multiple files listed in "podseglist.txt" into one and saving it in a file called "fullpod.flac".
59 Using Sox
The Sox version is simpler.
sox [0-9][0-9].flac fullpod.flac
60
This will concatenate all the two digit numbered flac files in the current directory into one file called "fullpod.flac".
61 Review the Combined Audio File
Next you need to review your combined audio file.
Listen to the resulting file.
If you are satisfied with it, you are done recording and are ready to upload.
If you are unhappy about some part of it, you can re-record just that section and run the combining script again. The numbers in your script will help you find the appropriate file.
62
The people running HPR will worry about adding the introductory and concluding music, converting it to mono if it is currently stereo, and adjusting the output level to make the volume consistent with other HPR episodes.
If there are noise problems that you want to try to correct I will cover that in another episode in this series.
63 Prepare the Show Notes
You will need to have a few things ready when you go to upload your episode.
These are the title, summary, tags, and show notes.
The title should be something short but descriptive.
If this episode is part of a series, you probably want to use a consistent title and include an episode number.
64
Next, you need a summary. This is a brief description of what the episode is about. Try to be clear about what it is you will be talking about.
However, there are limits on the length of the summary. The limit was 100 characters at the time that I was writing this.
65
The title and summary will be automatically added by HPR to the beginning of your episode, so put some thought into what you write here.
The title and summary are read out by a text to speech program, so avoid difficult abbreviations or words that the software may not know how to pronounce.
66
Next you will need to pick some tags. These are used for search purposes. I will let someone else recommend how you should pick tags.
67
Next, you need to have show notes.
If you have written a script, you can simply copy-paste the whole thing into the show notes.
68
At one time there was a limit on the size of the show notes, but that limit was removed recently.
69 Uploading the Episode
I will let someone else describe the process of uploading the audio file and associated title, summary, and show notes.
However, you will need to have an email address ready to use as part of that process, so if you have multiple email accounts you need to settle on which one will be used as your HPR contact address.
70 Conclusion
The preceding is how I created my first two podcast episodes, which were on Oathtool and the UCSD P-System operating system.
Plenty of people wrote in to say that they liked them.
Nobody complained about the quality of my narration or the technical quality of the audio.
You should be able to do the same.
71 Further Episodes in this Series
I have covered the basics, but there is more that we can do to improve the audio quality if you are so inclined or if you encounter an audio problem. In further episodes in this series I will cover the following:
72
Basic filtering with FFMPEG and Sox to cover general cases. It's a good idea to use this sort of basic filtering on your audio whether you notice any problems or not.
73
"De-essing" to improve perceived voice quality slightly in order to overcome sound artifacts inherent to using at least some microphones.
74
Normalizing audio to adjust the sound levels for easier reviewing.
75
Analyzing the audio signal with Audacity to discover the characteristics of any noise problems that you may hear.
76
Advanced filtering with with FFMPEG and Sox so solve specific problems such as I had with my headset or which you may have with environmental noise such as fans.
77
Command line recording and playing of audio using FFMPEG and Sox. This can help automate the process by automatically numbering the small audio files which are part of the recording process which I have described.
78
I promised some gratuitous hackery, and I will provide it in the form of describing how to do audio spectrum analysis using Libre Office Calc spreadsheets and GNU Octave mathematical software in place of Audacity when troubleshooting audio problems.
79
This concludes the first episode in a four part series on simple podcasting.
Scripts for this episode.
#!/bin/bash
# First create the list file.
printf "file '%s'\n" [0-9][0-9].flac > podseglist.txt
ffmpeg -f concat -safe 0 -i podseglist.txt fullpod.flac
#!/bin/bash
sox [0-9][0-9].flac fullpod.flac
Provide feedback on this episode.
