A New Version of the Same Version

By Shamus Posted Monday Nov 9, 2020

Filed under: Rants 58 comments

Like I said last week, the plan was to have Ross on the podcast and talk about his involvement with Watch Dogs Legion. Since you’re reading this and not listening to a podcast, that obviously didn’t work out. The reasons for why it didn’t work out are fairly stupid and complicated and will require the following long-winded explanation…

We use the program Ventrilo to record our show. Vent is ancient by internet standards. In its heyday it was used as voice chat by raiding guilds in World of Warcraft. The trick with Vent is that you need to pay someone to run a server for you. This means it costs money, which in turn means it has faded into obscurity as free services like Discord and Steam chat gave everyone the ability to talk for free.

For the past few weeks I’ve been getting alarmed emails from my Vent provider, warning me that I’m going to need to upgrade soon. That struck me as a little weird. I pay them to run the Vent server, so it seems like I ought to have the option to NOT upgrade right now. Just, you know, don’t upgrade my server and my client should continue to work.  Specifically, I wanted to finish the episode with Ross before doing the upgrade. It’s bad enough I force guests to install this ancient and needlessly uglyThe interface looks straight out of Windows 98. software when they already have newer alternatives installed, and the last thing I wanted was to do that to someone in the middle of a technical shake-out period.

My hosting provider didn’t make this easy for me. They sent an email announcing that this forced upgrade would happen at the end of October. Then just before the day in question, it got pushed back to early November. I shrugged. I only use it once a week, so I figured I’d just wait until they send out the “All done!” email before doing the upgrade on my end.

They did not send such an email. It just happened without fanfare. Which, fine, I guess? That’s not unreasonable. I was just caught off-guard because it’s not what I expected.

But then Ross emailed me, saying he couldn’t connect to the server. Did the upgrade happen and I missed it?

I jumped over to the Vent site and checked. Nope. The latest version for download was 3.1.0, and  that’s the version I’m already running. So the mandatory upgrade hasn’t happened yet.

Right?

The next day Ross still couldn’t connect. Just to make double-sure this wasn’t a new version problem, I decided to connect myself. And then I discovered I couldn’t.

So nobody can connect, but there’s no new version out? What’s going on here?

Next up was a long chain of misunderstandings and misguided troubleshooting as I tried to figure out what was going on.  Eventually I realized that I needed to upgrade to the new version… of 3.1.0!

Whenever you make a new version, you increment the version number! That’s literally why version numbers exist! Vent even has a three-digit version number. Just call the new one 3.1.1 or whatever. And if your new version breaks compatibility with the old, then maybe you ought to make this a major release and call the new one 3.2.0 or something. But the one thing that you shouldn’t do is come out with a new version of the same version number, because that’s nonsense.

But fine. Let’s say you’re bound to some external numbering system. Maybe you’re (somehow) forced to keep your version numbers in lockstep with (say) a third-party provider to keep the versions matching between client, server, and protocol. That’s a terrible way to run things, but whatever. If that’s the case you could at least append something to the version number to help your users out.  Maybe adopt a system of sub-versions like 3.1.0.1 or 3.1.0a. Whatever. You just need some way for people to tell two distinct and incompatible versions apart.

Paul and Ross: To fix this, go to the Vent download page, and scroll down to the bottom where it says “Version 3”. Yeah, just “3”. Then skip the paragraphs of ominous import regulations and go down to the bottom. You’ll see that the “latest version” is 3.1.0, which is supposedly the version you already have. But it isn’t! This version of 3.1.0 is somehow newer than the version 3.1.0 you’re using now. Just download this and install it over the old one.

You’ll know it worked, because… uh? Well, I guess you can’t check. The program looks the same and the About screen shows the exact same info. I guess you’ll be able to connect now, unless something else is preventing you from connecting.

Done? Great. We’ll try this again same time next week.

Shamus, Why Don’t You Dump Vent?

A fair question. This thing is old, ugly, outdated, and it costs money when the alternatives are free. The vast majority of podcasters seem to talk over Discord. Everyone records their audio locally in Audacity, and then after the show they all send their audio to the host. If everyone else does that, then why am I clinging to this silly system?

Because of risk.

Over the years I discovered that I really, REALLY hate it when we lose a show. It’s only happened a few times, but it’s been incredibly painful and I never want to do it again. It sucks to burn an entire evening and have nothing to show for it, and it sucks even more when we come together the following week and try to recreate those lost conversations. It’s confusing and awkward and it’s never as good as the original.

Here’s the thing: Vent is an exceptional recording tool for this sort of thing. When you hit the record button, it will record everyone’s contribution in an audio channel that’s isolated from everyone else. So if someone lags and their audio gets out of sync, or if people lag out and accidentally talk over each other, then it’s easy for the editor to sort that out because the voices aren’t all mixed together. If one person is too loud and another is too quiet, the editor can fix it in post because they’re not all smashed into a single audio file.

The important thing is that every participant gets a copy of the entire conversation. If Paul’s laptop crashes and it corrupts the recording in progress, that’s no problem because I still have clean recordings for both of us.

But let’s say we record on Discord instead. We can do the naive thing and record the raw audio of the discussion. But then we have everyone’s voices mixed together and we can’t fix problems in post. That sucks, so let’s not do that. Instead,  we each record our own audio locally.

Let’s say the odds of one person losing their recordingA crash, disconnect, Windows Update ambush, or a power outage, or just forgetting to hit “record” at the start of the show. are 1:100. In Vent the only way we can lose the entire show is if everyone hits that unlucky 1:100.  In a show with just Paul and I, our mutual redundancy means the odds that we both lose everything are 1:10,000. That’s pretty safe!

Now in a scenario when we each only record our own audio, the odds work against us in the other direction. If anyone involved loses their recording, then the whole show is a bust. My audio alone is worthless, because I’m obviously speaking to someone. So the odds of failure are now 1:50, which, on a weekly podcast, is about once a year.

The odds get even more extreme in a show with 3 people. A discord-style podcast has a 1:33 chance of failure, and a Vent podcast has a 1:1,000,000 chance of failure.

In Conclusion

Please use version numbers for your software. If you’re already using version numbers, then please don’t forget to increment the number when you release something new. And if you work for lunatics who won’t let you do that for some reason, then append sub-versions to save your poor users from confusion. Barring that, just copy the WordPress system and give each version a cute name. WP likes to name their major releases after jazz musicians, but you can do whatever you like. Name versions after colors. Ice cream flavors. Sports teams. Mythological creatures. Varieties of mold. Porn stars. Serial killers. Plagues. Whiskey brands. There’s so much fun you can have with naming versions. Express yourself!

Just don’t give a new version of [thing] the same number as old version of [same thing], because there’s already enough insanity and chaos in the world.

Thank you.

 

Footnotes:

[1] The interface looks straight out of Windows 98.

[2] A crash, disconnect, Windows Update ambush, or a power outage, or just forgetting to hit “record” at the start of the show.



From The Archives:
 

58 thoughts on “A New Version of the Same Version

  1. MerryWeathers says:

    I ate a taco today, still not into it.

    1. Zeta Kai says:

      That statement is a bit of a non-sequitur, but I can’t argue with your premise.

  2. Dreadjaws says:

    OK, but I have serious trouble believing that the one pro this Ventrilo software has is exclusive to it. There has to be other software that does the same and isn’t ancient and crazy with its updates. I don’t use recording software (more than once I’ve had the desire to start a podcast with friends or something like that, but I absolutely and fiercely detest the recorded sound of my own voice, so I actively avoid recording myself talk as much as I can), so I certainly can’t provide an example, but come on! There has to be. Surely someone else in the comments can provide an alternative so you don’t have to keep suffering with this nonsense.

    1. Lino says:

      I absolutely and fiercely detest the recorded sound of my own voice

      That’s actually very common – it’s called voice confrontation. Whenever I’ve heard people talk about it, they always say that the more recordings you make, the more used to it you become.

      But then again, I’ve only heard humans talk about it. I don’t know how well that applies to Metroids. I didn’t even know you guys could talk! I thought you communicated exclusively through sucking people’s life force…

      1. Dreadjaws says:

        We can all talk. We just don’t do it because we hate the sound of our own voice.

        1. CloverMan-88 says:

          We hear the reverb in our skull, so we all think we have deeper voices – and as humans we prefer lower pitched voices. So everyone hates recordings of themself.

    2. Uristqwerty says:

      For performance in large calls, most other software probably mixes all the audio together on the server, so 10 people means 10 uploads and 10 downloads, rather than 10 uploads and 90 downloads (server relays a copy of everyone else’s audio streams), or a full P2P 90 uploads and 90 downloads of everyone sending directly to each other.

      Of the remaining software, well, writing more than two audio tracks (Left and Right, or if you’re really being fancy you might get a few extra tracks for 5.1 surround sound) to a single file is very uncommon, so even if the file format supports it, the API of whatever library is doing the writing might not, or might make the task substantially harder. It’s far easier to just grab the single stream after everything’s been mixed together and shove it into the bog-standard audio file output with default settings.

  3. Lino says:

    I just went to Ventrilo’s site, and I got such a nostalgia trip to the Internet of the Early 2000’s! What I loved most were the themes that are JUST LIKE all those crazy Windows Media Player skins! I wonder which one Shamus uses – my money’s on the one in the middle, becuase it looks kind of sci-fi-ish. Although my personal favourite is the Egyptian-looking one. So cool!

    And I feel like there’s something to be said about a program that’s fairly priced, and does what it says on the tin without engaging in all that Facebook-wannabe, users-as-a-product schtick everyone likes to do these days. I hope they manage to keep the service running (and learn how to use version numbers :D)!

    1. Hal says:

      I love how their news section is about the use of Vent as the official tool of Quakecon.

    2. Chris says:

      I love websites like that. No weird rescaling if you have a different size screen, no dropdown menus that have a weird delay or are horrible to navigate. No menus that are stuck at the top of your screen if you scroll down. It’s sad that design like that no longer is used.

  4. Mr. Wolf says:

    This is “New Comment” because it is the newest comment.

    If it is not the newest comment it shall be renamed “Old New Comment”.

    If this comment is edited it shall be the “New New Comment”.

    If is edited but not the newest comment ot shall be renamed “New Old New Comment”

    And that is how versioning works.

    1. BlueHorus says:

      Now all we need is for the site to crash whenever someone tries to read your comment, and we’ll have the full experience!

  5. Ronan says:

    I think mumble has the same multitrack recording feature you need, but it also comes with the same caveats as ventrilo: it comes from a distant past and is not user friendly at all, and you need to rent a server.

    1. killphi says:

      Yes, Mumble has that multitrack recording feature.
      In regards to that feature, this guide is accurate. The rest is … well, ten years old.
      https://www.mumble.info/blog/for-the-record/

      Also, I would like to mostly disagree about the caveats.

      While I can neither refute that Mumble does come from a distant past and mostly still looks the part, it is IMO not worse than TeamSpeak in this regard, which I sometimes see still used today. Most of the “ancient” look comes from the overall very simple/simplistic Qt-based UI with a “native Windows” look for pre-10 aesthetics of “native Windows”.

      I’m not saying it’s the best in terms of user-friendliness, but it’s not _not_ user friendly. It does what it needs to and is very straightforward about that.
      Personally, I never had any usability issues, but I know people who did.
      Anecdotally, I can say that the sentiment of “it’s user-friendly” correlates with being on the tech-savvy side of gaming enthusiasts and an attitude of being willing to actually give it a shot, instead of dismissing it due to either looks or in some cases initial audio problems which may or may not have been the fault of Mumble.
      (It was an audio issue that happened on some computers for some programs, and was relatively consistent in that; at the time, Windows 10 was new and Mumble was still in the process of upgrading the Qt version of their codebase between 1.2.X and 1.3.X, recommending the builds from the beta branch)

      As for renting a server, yes, you can rent a server. But you don’t need to, you can run one yourself.
      It is actively developed open source software. It comes bundled with the server as an optional install in the regular installer. And as far as I’m aware Mumble/Murmer is found in most widely-used Linux distributions’ default package managers.

      Depending on the the circumstances, setting up a server can be a bit of a hassle, but not moreso than “setting things up on the computer for the first time” in general. It is something anyone mildly comfortable with reading documentation, editing a config file and potentially home router port forwarding can do.
      Assuming Linux as the server OS, running it on an “outside” server replaces home router stuff with minor Linux administration.
      It is not very resource heavy, especially for just a few people. Any server with install/shell access will do.
      For example, you could throw it on an on-demand cloud server instance that you manually turn on before the session. Or that you have scripted/scheduled to turn on 5-10 minutes before a session.

      Sorry for all the rambling, but I do think that Mumble is a better choice than Ventrilo, given the circumstances.
      Yes, it’s more of a “sidegrade”, but I do think that the replacement problems are either lesser versions of Ventrilo’s problems, or just fewer of them in the first place.

      And most importantly, the one thing Shamus really cares about:
      Sane version strings.

      EDIT PS:
      since semantic versioning was mentioned below, Mumble does not follow semantic versioning. 1.2.X to 1.3.X was a major upgrade that broke backwards compatibility.
      Otherwise, they do follow the route of “new version increments the version number”, except for unstable builds, which append the abbreviated git commit hash to the version number.

      1. Nick-B says:

        I run my own murmur server on my own computer. For my close friends, they know to chat with me they just log onto it. I just slapped a shortcut to the murmur server in my own computer’s startup folder and just remember to update my domain name/IP every now and then. (My isp uses true dynamic IP assignment. If my internet flickers for even a MOMENT, I am assigned a completely random IP address. If their servers are having problems and dump me every few days, I get a new IP).

        I used it because of some WOW friends needed it to raid. It ended up being pretty good in terms of voice quality and – more importantly – latency. I don’t know how accurate the claim is, but low latency was a big plus to me. Now, nobody knows what the heck Mumble is, but I love it still.

        1. Scerro says:

          To be fair, Mumble’s touted killer feature was that it had low latency. I heard plenty about it but it didn’t have enough draw for most guilds in the WOTLK – Cataclysm era.

          The primary thing that made Discord explode was how fast and easy it is for anyone new to get into a voice channel. You share a URL, they click it. They click the button to unmute themselves. It’s a 30 second process. Meanwhile mumble just took too much because of the full download/install/put in server/user/pass info. People (even 15-30 yr olds) are lazy bums.

          1. Abnaxis says:

            I’m pretty sure this is why you aren’t going to find the features from Vent in any modern software–the technical setup to enable those features means it can’t be streamlined as hard, so it won’t be as widely adopted

          2. killphi says:

            I think that “does not require initial installation” combined with “just works* (* in almost all regular cases)” is _the_ biggest feature of Discord.

            At the very least, that was why it won me over:
            It was just easier to get people on there, therefore everyone else switched to it, therefore I had the choice between being alone on Mumble, or having a general IM/VoIP platform for online gaming stuff and friends.
            Once critical mass in my social circle was reached, it was a no-brainer.

            Personally, I still think Mumble is the best choice for “just VoIP”, ignoring the initial bump of “barrier to entry”
            (installing, potentially setting it up, getting used to it; that bump is relatively small for most people, but it is there), but a large part of that is admittedly for personal ideology reasons.

  6. Armstrong says:

    You’d think someone would have come up with a new dedicated podcasting program by now, especially considering the podcasting boom going on.

  7. Mattias42 says:

    Hmm… that central server, everybody gets a recording copy thing honestly sounds pretty dang amazing for even semi-commercial content. I can see why Ventrilo lingers, and you’d be loathe to give up that killer feature.

    Question, though: Can’t you guys just… buy a small server yourself, and run the program on that? Wouldn’t that long-term be cheaper then renting? Or is it a proprietary code type thing?

    Been ages since I used Vent, but I vaguely recall my brother using it for WoW raids and such, and getting very angry if you turned off his computer because that was how they were running their whole guild’s communications. Could have been pirated, of course, but never really played WoW so the details kinda flew over my head.

  8. Moss says:

    And if your new version breaks compatibility with the old, then maybe you ought to make this a major release and call the new one 3.2.0 or something.

    Gaaah! I’m sorry Shamus but I can’t help myself from being smug and contrived about this.

    Semantic versioning is a versioning system designed for APIs, but which a lot of applications have borrowed for their own. Its rules are

    Given a version number MAJOR.MINOR.PATCH, increment the:

    MAJOR version when you make incompatible API changes,
    MINOR version when you add functionality in a backwards compatible manner, and
    PATCH version when you make backwards compatible bug fixes.

    If the new Ventrilo version breaks compatibility, it should be bumped to 4.0.0, not 3.2.0.

    There’s nothing that says Flagship Industries have to use semantic versioning for Ventrilo (or even for their APIs). In fact I don’t know what versioning system they use. But it still physically hurt my know-it-all wise-guy sense of self when you wrote that line.

    1. King Marth says:

      I love that the explicit alternative to semantic versioning is called “sentimental versioning”, where you use numbers in some other arbitrary way rather than to convey meaningful, consistent information. The best of these is TeX (a mathematical typesetting language), where its version number increases in length with each release, getting ever closer to pi (versions 3, 3.1, 3.14, 3.141, etc).

      Which is to say, if you’re going to be difficult, at least do so in a beautiful way.

    2. Shamus says:

      I didn’t mention it in the article, but the wrinkle here is that there’s ALREADY a version 4.0, so incrementing the major version wouldn’t work in this case.

      1. Echo Tango says:

        Is it newer than both of the “3” versions…or just one of them? (I’m pretty curious to know how badly messed up all their version numbers are. :)

      2. Steve C says:

        And it is a mystery why Vent lost its dominate position to Discord.

    3. Echo Tango says:

      Ventrilo’s probably just using whatever numbers they feel like, without any rhyme or reason. ^^;

      1. Kyle Haight says:

        At least they’re using numbers. There’s an API where I work that has three versions so far. They are referred to as v1, v2 and vSuperbowl.

        I have also seen cases where the branch number of a newer software release is smaller than the branch number of the previous release in the series.

    4. Sartharina says:

      I don’t think I’ve ever seen backwards compatibility preserved between patches in consumer software, because by the nature of the beast, that backward compatibility unfixes any bugs that are fixed. Or maybe the backward compatibility is API-only.

      Most software still uses the MAJOR-MINOR-PATCH, but compatibility breaks every single time: Major is for major overhauls of the product. Minor is New Features the user interacts with. Patch is bugfixes and backend work – but Backward Compatibility is not preserved in any instance. So no, the version shouldn’t have been bumped to 4.0.

      Your know-it-all wise-guy self is wrong.

  9. Asdasd says:

    That is a great feature; I’m surprised it isn’t more widely available elsewhere. Fond memories of hanging out on ventrilo and that dum um-BAH dum sound when you log in.

  10. Thomas says:

    I’d love to know what convoluted chain of events led to people deciding a new version should have the same version number as the previous version.

    Somewhere this makes sense to someone, and they’re probably so far down the rabbit hole that the daylight has redshifted.

    1. Eric Fletcher says:

      The number was probably assigned at the start of the work period, with a plan to be fully compatible (just a minor compatibility fix)
      Then, something ended up breaking, but it was too late to change the version number.

    2. Paul Spooner says:

      For the daylight to have redshifted, they would need to be moving down the rabbit hole at a high velocity. Also, it would need to be a very straight rabbit hole.

      1. Philadelphus says:

        Well, if we take general as well as special relativity into account, the further down the rabbit hole they are the more the sunlight will be gravitationally blueshifted, regardless of their velocity…though given Earth’s much smaller gravitational potential, it will never overcome the ~630 m/s gravitational redshift the sunlight endures climbing out of the Sun’s gravitational well in the first place.

    3. Erik says:

      Since there’s already a 4.x out, anything on 3.x will be a maintenance release. That (sadly but typically) implies less care taken than on the main line of release.

      I personally suspect that the new 3.1.0 was built from the same core source code as the old 3.1.0, but was rebuilt for some reason. And when it rebuilt, it sucked in updated external libraries for some component, and *there* is where the incompatibility lies. If I had to guess, someone found a critical security bug in a communications library that was used in 3.1.0 and updated the server to fix it. Then just rebuilt the client to match, and because there were “no changes” to the application source code, they didn’t roll the version despite the changes to the library. Leading to (as ever) Shamus getting screwed over for our collective amusement.

    4. evilmrhenry says:

      I assume a developer just forgot to increment the version number. The real question is why this isn’t automated, but I suspect the answer is that they’re using the same processes they used back when this was new software.

  11. Ashen says:

    I haven’t used Ventrilo for ages, but it used to come with server software too. Is that no longer the case? I remember running a server just for me and two other people back in the WoW days which seems similar to your use case.

    Also, A New Version of the Same Version is a perfect description of Watch Dogs Legion itself.

  12. Leeward says:

    semver.org

    This is not spam, I promise. Read it. It will make the world better.

  13. SoulMan says:

    FYI, in case it hasn’t been mentioned yet, there is a discord bot that supposedly does multi-track/channel voice recording. I’ve not used it so YMMV. Not sure the policy on posting links, so just google “discord + craig bot” or “discord + multi channel record”.

    1. Echo Tango says:

      You’re allowed to post links as much as you want, as long as you’re not spamming. Although note, more links or links from more than one site, seems to set off Shamus’ spam-bot detector more frequently. For the record, this appears to be the bot for Discord. Multi-track recording is one of their first bullet-points for features! :)

  14. Kathryn says:

    Lack of configuration management is often an issue. A while back, we (work) bought an RFID reader and tested it out thoroughly to prove it works in our application (stringent environmental requirements). It was great, so we bought some more. And they…didn’t work. Turned out the vendor had changed one of the components but didn’t change the top level part number. How…how do you not change the part number? It’s no longer the same part. I mean, if you swap out a resistor for an identical one from a different manufacturer, fine, you can have a materials equivalency matrix documenting that as an acceptable substitution, but this was one of the major components (I want to say it was the optical reader), and changing it changed the function of the device sufficiently that it no longer worked in our environment. You can’t make a change that significant and not roll the dash number.

    This was years ago, it wasn’t even my project, and my eye is still twitching. Configuration management, people. It’s key.

    1. Richard says:

      This is extremely common.

      A purchaser substitutes a component with an ‘equivalent’ that’s either cheaper or shorter lead time, and suddenly it doesn’t work anymore.

      Usually the ‘new’ component arrives some time before they run out of the original stock, the stock is mixed and then you end up with a large overlap of serial numbers which may or may not work…

  15. Hey, Shamus, I did some googling and apparently there is a discord bot available that makes it possible to separate audio kinda the way you want to? https://craig.chat/home/

    If that works for you, it might help you break your Vent dependence.

    1. Echo Tango says:

      LoL, ninja’d by someone down farther in the comments! :)

      1. Hah. I know literally nothing about it, but if it works properly it seems like it could be useful.

  16. CJK says:

    Ventrilo takes very few resources to host – indeed, if your ISP lets you open and forward ports on your router it’s trivial to run it yourself. It kinda makes sense to do that if the server will only be used for your own conversations – if your internet goes down you’re not recording anyway so it doesn’t matter that nobody else can connect.

    Alternately you should be able to get it running comfortably on a pretty modest VPS (“sign up to Linode today, and use offer code…”)

  17. wumpus says:

    I really don’t miss having long discussions in engineering meetings about how to handle version numbers, eventually coming to agreement, implementing the system, and then, a few weeks from shipping, having marketing come in and say, “You can use whatever system you want for versioning, but we’re releasing at 1.0, so you need to have that on the app.”

  18. Paul Spooner says:

    Well, I followed the instructions, and am able to connect! Talk to you all next week.

  19. Xeorm says:

    I’m really interested to hear the story on how that all went down. Why even have version numbers if you’re not going to keep them updated?

    Though I wonder if it isn’t because of a hidden patch number being updated. So it reads like 3.1.0.450 to indicate internal versions. So it’s a different version, but not one that the costumer would ever see.

  20. Pink says:

    The older style interface is a plus for me. I find ‘slick modern interfaces’ to be ugly, unusable, resource hogging messes most of the time. And they are usually not made with any but the most casual use in mind.

    “What do you mean, you don’t want to use more resources than photoshop for a less useful chat program!? It has all these social media and spyware features you don’t want!”

    1. Radkatsu says:

      This is why I dropped Skype after 6.21, once they made it impossible to keep using it. Screw 7 and up, that garbage is for kiddies.

  21. Rariow says:

    I have a coworker obsessed with naming his files “final” as soon as they compile. Of course the first compile is never the final version, so he ends up with names like SomethingFinalLast2DoneFinished. Somehow that’s less horrifying to me than upgrading to the exact same version.

    1. Asdasd says:

      A member of the ‘student emailing assignment to self’ school of version naming, I see.

  22. Kyle Johansen says:

    I have to confess my own versioning problems. In my defence, I only do solitary issues. Still, at least leaving it completely aside means not misleading.

  23. Naota says:

    True story, I had the same “3.1.0 but not quite 3.1.0 enough” version as Shamus all this time, because the last time I used it was to join in one of the Good Robot podcasts with Arvind from Pyrodactyl, where it worked fine. I then tried on my ancient laptop – which had the same one. My desktop also still had installed copies of Mumble, Teamspeak, and old Skype kicking around the same folder from my Firearms: Source days. If I remember right, Arvind had the same from his time with the Dystopia modding team.

    Apparently janky voice comm programs are just kind of a way of life if you’ve ever used the Source engine?

  24. TLN says:

    How does everyone else do this, I feel like “2 or more people talking to each other online and recording it” is done by millions of people on a daily basis (even more so this year when people are likely to record podcasts etc. online rather than meeting up). I have to imagine that they aren’t all using Ventrilo or Discord?

    1. CJK says:

      The way regular podcasters that care about audio quality do it is double-ended recording with a safety track. So you have all the guests record their own voice individually, but to safeguard against losing one of the tracks you also record the everyone-mixed-together Skype (or whatever) conversation, so you can edit one participant in from that at reduced quality if you really have to.

      Really-professional podcasters will add a second, external safety track, recording from a pre-out of their audio interface to an external dedicated recorder.

      In practice, the reliability issues that Shamus is concerned about don’t seem to happen anywhere near as often as he suggests. If anything the weekly shows with regular guests establish a routine that helps everyone get it right pretty much every time.

      (The way regular podcasters who don’t care about audio quality do it is simple: just post the Skype recording with minimal editing. I don’t think there are any hugely-popular podcasts like this, but it probably describes the majority of podcasts that exist)

  25. Sean Payne says:

    I was going to say that a stubborn approach to not moving on to new software is self-defeating and then I vaguely remembered you saying in another blog that you like using Jasc Paint Shop Pro 8, which I also refuse to move on from. GIMP is *almost* there, but there are still little functional things in JPSP8 that just work better. It’s just utterly frustrating that Windows 10 demolishes the PSP install every time there’s an update.

    Presuming my memory is correct (less likely these days) and I’m not superimposing my opinions on to things I think you’ve said before, do you still use PSP8 Shamus?

    1. Shamus says:

      I do indeed still use it!

Thanks for joining the discussion. Be nice, don't post angry, and enjoy yourself. This is supposed to be fun. Your email address will not be published. Required fields are marked*

You can enclose spoilers in <strike> tags like so:
<strike>Darth Vader is Luke's father!</strike>

You can make things italics like this:
Can you imagine having Darth Vader as your <i>father</i>?

You can make things bold like this:
I'm <b>very</b> glad Darth Vader isn't my father.

You can make links like this:
I'm reading about <a href="http://en.wikipedia.org/wiki/Darth_Vader">Darth Vader</a> on Wikipedia!

You can quote someone like this:
Darth Vader said <blockquote>Luke, I am your father.</blockquote>

Leave a Reply

Your email address will not be published.