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

367
hpr_transcripts/hpr3461.txt Normal file
View File

@@ -0,0 +1,367 @@
Episode: 3461
Title: HPR3461: Changes to HPR Branding
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr3461/hpr3461.mp3
Transcribed: 2025-10-24 23:56:10
---
This is Hacker Public Radio Episode 3461-4 Monday, 8 November 2021.
Today's show is entitled, Changes to HPR Branding, and is part of the series HPR Community
News. It is hosted by HPR Volunteers, and is about 41 minutes long, and carries a clean
flag. The summary is run, Dave and Ken read the entire email thread related to changing
the HPR theme.
Hi everybody, and welcome to another edition of Hacker Public Radio. This is Community News,
I reckon 0.1, because for 2021, 10, and this is 0.1, because this is the follow-up show,
to do those a lot of comments on one of the threads, so we're just coming back a week later to
record those. We're short of shows as it happens, folks, so if you are in a position to be able
to send in some shows, we'd really appreciate it. That would be great.
Yeah, so joining me tonight is first in the list is me Dave Morris, me Ken Fallon, and me Rowan
Horning. So we're just going to, there's a link in the show notes, so we're going through the
archive list on possible causes and solution to subscriber attrition, trying again without
encryption, and gone through that thread, because it actually turns out to be about the intro
and outro for the HPR branding. So without further ado, who's going to start Dave, everyone?
I can start with the first one, and it comes from, I might mess up people's handles along
the way, but excuse me if I do, and a core I think it is, who says hi, there was a community
news show about a month ago during which the producers, wow, producer, remarked that there was a
plateau in subscribers. I have a suggestion that may improve this attrition rate, and I'm sure
a lot of others listen to podcasts to wind down while going to sleep. The intro song is loud,
low bit rate sounding trumpet chip, June fanfare, which is fine. The thing is, this show is often
recorded at a low volume level, so the volume must be adjusted upwards to hear them.
And after being sued by a soft spoken hacker voice, the blaring loud chip tunes are
blasted in again at the end of the show. This has worked me up for my sleep many times.
I can only imagine the consternation from headphone users. I'm sure this has led to many rage
unsubscriptions by those less level headed than I. There is a much more mellow sounding intro,
that is same tune at a lower volume. Let's use sometimes. If there must be an outro song,
this should be it, and the loud blaring fanfare tune should only be used at the beginning.
Okay, and BK Navarit responded, oops, and I think it's good to sub to Stan Brian in Ohio.
That's his handle. Sorry, that is his.
Yes, he operates under those two. Because in many cases, people are using their real
email addresses, and then they're signing with the handle, so either if I see it,
or I think just go by what the world would say. Because if somebody was reading it,
themselves, they would do the same thing. They wouldn't read anything.
Okay, sorry about that. Well, should we just keep going, or do you want me to re-tape that?
Sorry, Tim. No, no. No, we'll keep going. So, I replied, so I'll do my own.
Hi, Anacore. Thanks for getting in touch. I replied to the list, but you can feel free to
reply to me directly. See my comments online. So, I'm commenting about his putting him to sleep,
and then I say, out. The community news is putting you to sleep. That hurts.
Joking aside, I doubt that most people use HPR to go to sleep. The variety of shows on HPR is so
different you would want to be on your guard when listening. For example, I will only listen to
a Dave Morris or Tukatoruto show, if I'm wide awake, focused and at a computer with access to
show notes. I'll only listen to that too, show if I have a cup of coffee before me. The purpose of
HPR is not to put people to sleep. It's to get them up, thinking, hacking, and recording shows with
the goal of sharing knowledge. Then in response to the intro as loud and chiptune blaster at the end
of the show, fair point, you're not the only one who mentioned this. We have several volunteers
who will help us clean up the audio and help to improve their setup. We will unoccasion farm
the poor audio out to services like audiophonic to see if they can fix it. However, you don't
mention which show in particular you have any problem with, or do you feel that this is a problem
with all of them. Not knowing which shows are having an issue, we can't tell if it's an issue with
the source audio, the processing, or the transcoding. If it's a community you show, then we can do
something about it as it's directly under my control as I'm the one producing it. I will ensure that
the intro and outro of next month, that was last month, one is added level and leveled manually.
I would like to give your feedback on the efforts afterwards.
If it's another show, then you can get in touch with me directly so that we can see if we can
address the issue. If you do feel the audio is poor, there is something that you can do to help
the project out and that is to report it to us before it hits the main feed. There's no need for
any audio files to suffer in silence, just subscribe to the future feed and provide Admin
at HPR with feedback on the show's audio. If there are issues, then we can try and fix them
before it hits the main feed. The link to hackpublicrelia.org fordslashcentication.php has got
information on how to subscribe to that by the way. Then in response to the sleeping and
rage on subscriptions, I reply, I doubt that the audio on HPR has been consistently variable
since its inception. We should be seeing a consistent rage of rage on subscriptions as well.
The dip beg, I have no idea what I was saying there. I imagine most people skip the outro entirely and
there's a more mellow intro and outro song. Yes, there must be as we need to provide feedback on how
you can contribute a show to the project. What? HPR public radio theme, I have no idea what I was
in there. A link to the theme page which shows the various different options that there are.
There are several intros and outros available which one are you referring to and I'll put it
in. Link to the theme music page. As your use case, I feel at least it's pretty unique. I'd
love to record a show telling me about it. Don't forget to introduce yourself and tell us about
your journey to HPR. Feel under absolutely zero pressure to send in the show. I mean, it will be great
if you could as we're running no one shows in the queue and have a call for shows open. But as I say,
let the fact that every single HPR show you're used and go to sleep has been contributed by someone
very much like yourself rather than just consuming concepts who took the brave step to themselves
become let me say a hero of sorts. If you do decide to contribute, it's easy just nip over to
the calendar for more information. Hope to hear from you soon. And the next one is from
Satu, I think. Do you want to take that day or do you want me? The silence which everybody will
now know about is because I cannot click on things on browsers and I hit the wrong, but the wrong
the one that said go back to the previous one or go somewhere completely different. Anyway,
yes, this is a message from Klaatu who says interestingly, my own podcast,
new world order has gotten a few emails from people over the years who say they fall asleep to
podcasts as well. Many people identify strongly with the brand of a podcast by its intro and outro
music. However, other people find the almost subliminal chatter about tech soothing and relaxing.
We can consider releasing a special feed without music, but that would not guarantee a relaxing
lesson. An HPR host might add loud sound effects or music or shout or whatever to their show,
independent of the intro and outro music. I've been looking into ways of doing a no music feed
my own show, but so far everything requires a bunch of extra steps that I'm not willing to do
myself. The reason I'm able to release a show so often is that most of the posting process is
automated, so I can afford no extra steps. I believe the same is true for HPR. Point is,
this is something to consider, and if I come up with any brilliant ideas for separating
the music from voice in an easy way, I'll let the admins know. Okay, and in response to that,
we have one from Andrew Conway. I'm a nocturnal podcast listener mainly because it's the only time
I have to listen properly. I cannot work, code, write, or even exercise and listen to speech
properly. Also, my daytime podcast bandwidth went down drastically when my daily commute ended
some years ago. I imagine that's become much more common for people in the last year or so.
This means that music that might wake me up puts me off listening to particular podcasts,
and so the HPR theme is a problem for me, but not so much that I would ever unsubscribe.
To be honest, unscraping from HPR sounds like a usefulism, like shuffling off this mortal coil.
If a musicalist's feet existed, a class to suggest, or a more gentle version of the theme was the
entrance default, I would definitely listen to more HPR shows. That said, I wouldn't want to create
more work for folk if I'm only part of a small minority, Andrew. Interesting. Jason Dodd replies,
lol to be honest, unsubscribing from HPR sounds like a usefulism, like shuffling off the mortal coil.
That's kind of true for me too. I still firmly underside of not caring about production quality,
not again looking for an easy way to eliminate the jarring change in volume, but it's not something
that would make me unsubscribe. That being said, HPR wouldn't make my top 100
list of things to go to sleep to. I think it's interesting that there are TV channels and various
professional streaming applications that have the same issue. Maybe this problem is harder to solve
than it looks on the surface. I typically fall asleep to TV, and it does irritate me when commercials
are so much louder than the shows. Wow, I'm so glad Anacore started this thread. I had no idea
that people were falling asleep to HPR shows. We do now keep the original unbredded audio,
and that's available on the internet archive. It would be possible to create an unbredded
feed from that. It may fix problems for some shows, but there are still a lot of shows that's
in a lot of hosts that send in shows with the intro and outro already included. On the main feed
can switch the default audio, on the main feed we can switch the default audio to a quieter one.
Tell me which one you would like. So Clatu is next and he says we could make it a policy to request
shows without music, and just resolve to attach the intro and outro ourselves as we do already
as needed, or for our music feed. Then by default we would be getting shows that would go straight
to our music less feed. Speaking as one of the hosts who bakes in the intro and outro music,
I'd be happy to stop doing that and the interest of making separate feeds available.
Okay, and then on October 17th I was responding to Ken, and I guess the original feed, it says,
so it quotes, the intro song is loud and the bit rate,
sounding trumpet chip ping fanfare, which is fine. The thing is, and it sort of repeats that part.
Let's get down to however you didn't mention, which show in particular you are having a problem with,
or do you feel this is a problem with all of them? And I responded, I have noticed for my show
that often the intro outro is louder than my recordings, both when I record using my phone or my
computer. I have tried to boost the audio levels to match the intro and outro. My audio editing skills
are still minimal, and often my audio would start to sound too harsh when I try to amplify it too
much. Until today I didn't think about trying to download the intro to use it to help match the
level. See, I did download the theme music and it's the link to it. This seems to be the standard
intro clip used to attach to shows. One thing I did notice when I'm porting this clip into my
audacity project, the average volume DB levels showing on the meter fall between negative 12 DB and
zero DB, often being around negative 6 DB. When recording on my phone, the average levels are
between negative 18 DB and negative 12 DB. Given how hot the recording levels are for the intro clip,
along with the type of audio it is, it had to perceive loudness much greater than a typical
voice recording. Perhaps a simple solution to production process for the janitors is to use a
clip that has its DB levels decreased by negative 90B to negative 12 DB. This range seems to match
the loudness of my default recording levels. Granted that would make the situation much better
for me, smiley face. I have a Samsung Galaxy S9. I'm sure the recording levels vary to a degree
from phone to phone, but maybe a quick survey would give an average range to target. If nothing
else, having an intro outro clips be a little lower than the user's content would solve the sleepers
issue. They would turn it up or not to hear the intro, then turn it down to the desired level
when listening to the content. If they fall asleep, the outro will be lower and not wake them up.
On another point, I didn't notice the loudness normalization effect in Audacity. I'm going to try
using it from my next show. After adding the main tracks of the show, I will use this effect to
try and bring the level up to the level of the intro clips. Okay, so next is from Taj Taj Sarah,
and it's a reply to an earlier show, but I'm not sure that I should even mention that.
What he says here is, I'm going to be the pedantic audio nerd here. Using decibels to measure
this sort of thing can quickly lead to issues. The above example kind of proves that. That going
into a really long round about audio measurements and acoustics. The industry standard for measuring
finalized audio levels like this are L-U-F-S, loudness unit full scale. The level that is
standard for podcasts is minus 16 L-U-F-S. On tenacity stroke or dacity, this can be achieved
using the loudness normalization effect. I usually do some work on audio prior to using that plug
into even out the levels. This would ensure that the intro, body and outro of the show were the same
loudness. Yes, Ken, I know I owe you a show. I agree with Clare too. I usually bake in the intro
and outro, but would be more than willing to not if it makes another feed easier to produce.
And Taj says, I would subscribe to a feed that did not have the intro music. I like the branding,
but I know there have been several comments about improving the music. The best solution may
be to offer a feed with only the contributor's voice. The length of the intro annoys me. TTS
followed by ad for A-H followed by Horrible and Rockest music. There's much softer version of the
music which in my humble opinion should be used. And I'll do my moment. Too long didn't read
episode 3, 4, 5, 4 where I uploaded a show as a demo and it only had a very short six second
intro clip, the acoustic version. So regarding hosts, uploading intro and outro. From a show
processing point of view, we have never been able to automate the posting process due to the fact
that some people do and some people don't add the intro or under outro. Even if they are flagged
correctly in the opera form, I sometimes forget and end up having two reprocess files because
there are two intros or none. And in addition to that, quite often people don't flag it correctly as
well. For this reason, I would suggest that hosts do not add the intro or outro. With regarding to
non-musical feed, where it's possible to do, I am not in favor of having a new feed without music.
We accept any audio format and to quote, flag is best, we accept the rest. That needs to stay.
From a practical point of view, a raw feed would not be useful as the audio formats and encoding
would change from day to day. We deal with that through our conversion scripts so you don't have to.
We already have different feeds for the internet archive and the HPR website and on the stores.com.
Adding yet another feed would just complicate matters and will increase processing time considerably.
Boss dot dot dot. I think we can simplify it all into one. If we take a look at the makeup
of the show, we have seven possible sections, which is detailed on hackerpublicradio.org,
forward slash media, forward slash theme, dash music. First part is the show synopsis.
This is the text to speech. I was in favor of it, but everybody hates it. It should go.
Thanking our hosting provider, the internet archive don't know we do this and Josh has never asked
us to do this either. Josh is the main sponsor and the person behind anonymous tools.com.
And it's also a HPR host for that matter. In speaking with him, unfortunately, the mentions
don't convert to clients. So with his permission, I'd like to move that to the start of the
community news so that it's handled in a personal thank you from the Janishers. So part three is
the HPR introduction music. There are a lot of themes available and all are long. I do believe we
need to add a theme to maintain the HPR brand. Otherwise, we're at risk of being considered
an amalgamation feed. There is no need for it to be long or loud as it is. And then the show,
which should be left in obviously, the HPR outro music. Again, we can cover this information
and the community news and in the metadata of the show. So we can drop it entirely. The epilogue,
sorry, Poké, they need to go. He's the only one who has ever used it. And the promos, we've never
did any, but if there are, we can do them as part of the community news as well. Just before we go
to the next one, I want to say I've changed my mind a little bit about some of this. So you'll see
that later. Did you want to do your, so in summary, in the end there? Oh, so there's more. Sorry,
so in summary, I'd suggest dropping everything but a very short intro and move the rest to the
community news. As a proof of concept of just post to the show, we're wearing loads of
the show at the moment hint. And I use the first six seconds of intro AK, something as the entire
promos. Thank you. And those were contributed by Yerun's friend. So awesome. Yep.
The script to process this reduced in size by 50% as did the time to process the files. I also
suspect that normalization of the normalization will be easier as well. Just on the bite by
and it doesn't come up. Taj mentioned a thing there in audacity that can be run
a plugin that can be run via audacity. I would like, I understand from what I see on the internet
that it is possible to run audacity from the command line without a UI. So run scripts or run tools
or stuff. If anybody knows how to do that, I would much appreciate it because that would make
my life a lot easier. Anybody audacity how to use it from the command line can you get and touch
please. Okay, next one. Next is Klaatu. Shall I do this one? Please. Since yeah, okay. I'm
losing track of which order we are in. This is Klaatu and he says this sounds like a really good
compromise. I think like everyone agrees that less is more, both loudness and length. What a bonus
it makes and what a bonus that it makes processing easier. And the next is from Kevin O'Brien. Anything
that makes processing easier is good I think. And Taj says I agree anything that makes processing
simpler is beneficial. And Andrew says good plan can I approve? And Taj replies can my only
super nitpicky thing I'd say is possibly edit out the last second or so of the music to get rid
of the simple shaker thingy that starts to come in. You might be able to just do a hard fade out
and catch it. Do you want to do the next one? Let's see. And from myself. Let's see, responding to
this test to speech I was in favor of it but everybody hates it or should go. I respond. I do like
the text to speech synopsis of the show. I see it as a hacky geek part of the podcast brand.
Then responding to thanking the host or provider. While the attribution to an honest host may not
have generated bunches of sales, I still think it is a good thing to do. I have at times
looked at using their services. They just haven't met what I needed at the time at the price point
I wanted to pay. I did recently make a donation to them towards HPR. I found that link on the
About page when exploring the HPR website the other week. Maybe we should stress the donation
link. Many podcasts, many podcasts mention at least once if not multiple times their Patreon or
other donation links. Not saying that we need Patreon just highlighting the method we do have.
I often listen for years to a podcast before I start financially supporting it.
Yes, very good idea actually. That's something we should add automatically to the first part
of the community news day. Then let's see responding to HPR introduction music.
I do believe we need data to maintain HPR brand otherwise we risk considering
it on an automation feed. There is no need for it as it being too long or loud. My response was
other than the levels on occasion, I don't mind the current intro and outro music. I've heard
it so much that I occasionally humming around the house. As far as branding goes, the other night
I was humming it and my partner asked, are you humming the music to your podcast? Then responding to,
so in summary, I'd suggest dropping everything but the short intro and move the rest to the community
news. I replied, I feel there is value in keeping a fair amount of the intro and outro the way it is.
Finding a way to tighten it up is fine, but for me part of the HPR experience is the intro and outro.
And I reply to that, not ignoring this, so the discussion had come quite for a while. It just
that we need to bring this discussion to the wider community during the community news. Thanks
for the feedback. Also on our matrix channel, which is HGTPS column, port sash, port sash,
matrix.io for sash hash for sash hash, HPR colon matrix.org. So far most people seem to like the idea
of shorter consistent branding. I think this sample episode was a smidge in too short
and I miss a small outro to signal the end of the HPR sandwich. I'm interested to hear what the
rest of the community will suggest, which is why we're doing the show folks.
My turn now, isn't it? So DNT, do not translate comments. Hi, this is Lurker.
You just translate it, Dave. Well, no, I didn't. I debate him. I didn't translate it into anything.
I work on the comment column.
Anyway, this person says, hi, this Lurker would like to come out of the woodwork and say that the
discussion started from a perceived concern about a plateau in subscribers. And one of the defining
things about HPR, it seems to me, is its focus on attracting contributors rather than listeners.
Therefore, I think it would be very important to retain the outro, specifically in the part
about this episode was contributed by an HPR listener like yourself if you want to find out
how easy it really is, et cetera. Apparently not easy enough for me,
an exclamation mark, but still I find it important for the invitation to be there in case someone
is tuning in for the first time. And for someone who's been tuning in for a while and still hasn't
converted, it's a welcome reminder to which I reply valid point as it is the sure license
declaration which we need to do. And let's see, Todd follows up with, I think, I think pointing
to show 3, 4, 5, 4, your sample is a great compromise. The intro is short, identifiable and
easy to listen. Oh, every finished. I think we did. No, wait for me. You want to go on to the next
thread. That was 10 minutes of silence there, folks. The truncate silence will reduce
its nothing at all. Silence is not a bad thing. Silence is golden, you know. Did you want to
talk about automatic musicless feet, a thread started by Easy Leab of It? Yeah, that's kind of
related, is this? Yes. To be honest, I didn't really join the two together in my own head.
Yeah, we can scan through it, in which case it's my turn to read this examotical. This idea,
this is Easy Leab of It, who is also Black Colonel. And this idea is related heavily with the project
I'm percolating between all the other things I'm doing. Namely, a way of setting up some kind of
auto tools, ESC, deployment of podcasts. I've been bolstering my knowledge of XML, stroke,
XSL over the last few weeks, but if anyone wants to work on it as well, the idea is, and then he's
got an XML based template for a show, which contains sort of standard parts that you, some of which
we've been discussing as we've been talking. I don't think a verbal, an audio section is a good
way to describe the XML, so have kind of a look. It's basically a show, intro class music file,
part to file, section main, file to section, and then some effects start an end of the effect,
and then the show. So describing within XML, what do you do with the source audio?
A make file template would be the second component to ingest an XML into FF
MP or some other application to apply those edits to the audio non-destructively. Third point,
a configure shell script that would allow you to modify the make file or the XML via XSLT to
remove certain sections or export the edges to show a specified format. Fourth part, something
like a make published trigger that could automatically upload a target to the distribution server and
update the item feed accordingly. The idea behind this was it would allow people to examine the
raw, preferably unedited files and effects that went into making the podcast they like in order
to give them inspiration and stroke example to follow, to make their own podcast. It also has
the upshot of being able to quickly spin up the spoke feeds, the various needs, and allowing
listeners to compile the show to themselves if they have special requirements. As I said, I'm trying
to work on it every now and then, when I have time, but I don't have a lot of time, and I'm
very new to XML stroke, XSL as well as not being, it's not well versed in FF MP, but I would certainly
like to be. So if anyone sees this idea and thinks it would be easy for them to implement,
I wholeheartedly encourage it. To which I responded, okay, that's a show, no pressure.
And John Spriggs responds, I do something similar, but honestly, much less efficiently,
with cchit.net. The show is run using this script, HTTPS, GitHub.com, cchits, website, blob,
master, cli, showmaker.php, and this library, HTTPS, GitHub.com, cchits, website, blob, master,
cli, library.php. And yes, it's a God awful PHP written over 10 years ago by someone who didn't
really know a lot about object-oriented programming, but thought it was a zooming good idea and should
be done that way. Basically, I use festival to create the text-to-speech elements and then use
socks to move audio around and add audio or silence to the start and end of the show and layer one
element on top of the next. You're welcome to have a poke around in the code and ask questions
that will smiley-face. If you want to take this off list, I'm happy to be emailed directly,
or you can contact me via Telegram, HTTPS, T.ME, join John the Nice Guy, or Matrix,
Matrix.TO, at John the Nice Guy. or colonmatrix.org, or push Twitter or LinkedIn DMs.
Links to those profiles are on my website listed below. John the Nice Guy sprigs.
And if you're not subscribed to CCHead stuffnet, you should be.
Indeed indeed. Is it my turn?
B.K. Navaret, Brian and Ohio says, I fast-forward one minute 15, works of treat.
Easy has. The only problem I have is that I know exactly zero things about PHP. I didn't think
the web would get this big or last this long, so I never learned much about web development.
That being said, I will try to poke around the code and extract what I can, as it does seem
similar in function if not in form. B.R. What's the abbreviation for?
No idea. No idea. Okay, and that's it. That's it. So what do you think guys?
We're interested in hearing back from the community on this whole intro and outro thing,
because it's kind of important. Particularly, I'd like to know if people do like the
text-to-speech thing at the front or not. So we're kind of saying we should go for a quieter
intro and outro music. The fact that I would like to personally get away from
generating two separate feeds, one for the internet archive, one for the other one.
I think we could put in change the text of the outro, so that we're saying we'd like to thank
our sponsors on honesthost.com, internet archive and orsink.com, orsink.net,
and then do specific mentions of them on the community news per default.
We do need to keep this show was licensed into the creative comms thing, and the point is
valid about, yeah, if you want to contribute your own show, it's that should also be kept in.
So there is a little bit of text that we could put in the outro, but we can make it a lot quieter
and softer. So I would like to reiterate that I do like the text-to-speech at the beginning,
and like for me that sort of is part of the HDR brand. So I mean, it wouldn't drive me away
if it wasn't there, but I think it's important. Yeah, I think if I could get
if we could get the use of audacity or tenacity or whatever,
automated on the command line, there's a lot that we could do there with playing a quiet intro and
then have it reduced and then put in a text-to-speech and have the sound of the intro lower
for the text-to-speech so that you can hear it and then go back up so that the overall
intro will be determined by the length of the text-to-speech.
Yeah, that's a thank you. I did start like reading up and looking into like ways to automate
some of that, I mean just based on what I know of your workflow, but haven't really gotten that far.
Yeah, we convert just to give you an idea of what we do. We have no clue what sort of audio we get
and sometimes it's you can tell when people get broadband because it goes from MP3 up to flak
or all go up to flak. We love flak. Sometimes we get WAV, sometimes we get AU, sometimes we get
yeah, PWM. We get all sorts of stuff, especially 15 years ago it was a lot worse. We get
like mono on one side and DC, you know DC voltage range off and the sampling rate all off. So
what we do is we can convert that to a PCM, pulse-code modulation, just raw audio waveform and then
from there we can convert it back into a the various different formats. Now, for those of you
not familiar with audio, MP3 or our lossless formats, which means they're designed to take out the bits
that your human ear won't hear. Human ear is unlikely to hear. I think you said lossless,
not lossy, sorry, lossy formats. They're lossy formats as in some information has gone.
Loiceless formats like flak and WAV are sampled at a rate that's, well, if the sample rate is too low,
you're going to lose stuff as well. But whatever the sample they will represent to you, flak is
the free lossless audio format. So, audio-colic flak, yep. So that's the one that we prefer.
But there you go. We convert to a hog, speaks and let's see it on MP3 obviously.
While we're seven different formats that we convert to, so it's bloody.
Which, and I think that's that, guys. Okay, I'm just very curious which one on text-to-speech do you use?
Whichever one works best. At the moment we're using festival. Lin has finally given up
on deviate and decided a lot of Linux links show is never coming back. So she got in touch and
we said yeah, we'll take her over. And prior to that we used eSpeak. We used the Google Translate
engine for a while, but that caused loads of problems because the API changed over time.
And never felt really comfortable doing that to prefer to use tools that we can rely on being
there in the years to come. But that's always open. If somebody's got better suggestions, please
give them to me. That sounds good, thanks. Yeah, I definitely like the idea of like
keeping the tools of under control as possible. Yep. And this, all this stuff the scripts that I use
is on public repo. So you can look at the code and laugh and suggest better. We will always
take in better recommendations. And the best recommendation is to record your audio as
the best quality format that you can possibly do that will always be the best idea. Or you risk
the wrath of Taj. I hope the same with your database is the same with your audio.
garbage in garbage out. Okay, I think that wraps it up. Tune in tomorrow. Shall I make an effort
to see for this show? I will try to do a little bit of a hack to see what that quieter intro
would be like. I have no idea how I'm going to do that. Well, it means I would need to edit it
manually. Okay, well, I'll try and see what the vision would be. That's good. Okay, tune in tomorrow
for another exciting episode of Hacker Public Radio.
You've been listening to Hacker Public Radio at HackerPublicRadio.org. Today's show was
contributed by an HBR listener like yourself. If you ever thought of recording a podcast,
then click on our contributing to find out how easy it really is. Hosting for HBR is kindly
provided by an honesthost.com, the internet archive and our sync.net. Unless otherwise stated,
today's show is released under a creative commons, attribution, share like, slidot o license.