So the tone of this blog post is going to be a little bit different. And the reason why is because I’m not actually typing this blog post. I’m speaking it. I’m sitting at my desk right now with my microphone in front of my face, and I’m talking and watching as the words appear inside of Emacs. It’s really freaking cool.

Now the question is, why have I set this up? Well, the answer is, is that I want to start scripting some of my videos. Not only for the Linux Cast, but for my history channel, and maybe another channel that I’m thinking about creating. I want to be able to actually script, and then read the scripts in front of a camera. When I tried this by writing out a script in traditional fashion with my keyboard, and then reading it out afterwards, it wasn’t a good experience. It never has been. I write really well. I have very good grammatical skills, for the most part, when I’m typing. But when I speak, I don’t speak like that. My grammar is not very good. I say words out of order a lot. I have filler words. And I say things that don’t make any sense. And I also repeat myself quite a bit. It’s just kind of my speech style. If you watch any of my videos on the Linux channel, you’ll know this about me, that I ramble a lot. And that’s my normal speech pattern. It’s not the way that I write.

So when I went to read out a script that had been written in the normal fashion with my keyboard, it sounded nothing like me. It sounded… Well, it sounded scripted. And it didn’t sound good. So what I decided I wanted to do was create a speech-to-text module for NixOS using Whisper so that when I sit in front of Emacs to create a script, I would just pull out my microphone and actually talk about the things that I want to talk about, and then I’d have everything written down. I could edit it up a little bit afterwards so that I could take out parts that are repetitious or whatever. But I’d still have something that actually sounded like me at the end of the day.

And I have to admit, it works really freaking well. Now, is it perfect? No. While I’m talking, the text that’s outputted has no punct uation or paragraphs or anything like that. So I can’t really see where I’ve gone wrong. And once Whisper has actually gone through and made it better, there’s still some problems. So punctuation gets put in weird spots. There’s some spaces that needs to be worked out. And I still don’t talk in a way that makes sense a lot of the time. There are sentences that just kind of come out of my mouth in the wrong order. It’s a dyslexic thing or something. I don’t know. I do have dyslexia, but usually it doesn’t translate into speech for me. But apparently it does. My point is that it’s not perfect, but it’s actually damn good. So I’ve taken two pictures here. One of them is while I’m talking into this live. And it’s just one big wall of text. And the other is after Whisper has cleaned it up. And you guys will actually see the output because that’s what you’re reading. But you kind of see the differences there. It’s actually really good. And one of the things I’ve noticed now that I’ve actually used this a couple times for videos is that it’s a lot easier for me to read stuff when it sounds like me. Instead of being so, you know, grammatically correct and lacking my filler words.

Now the question is, can I use this for other things other than scripts? Because you guys are reading this right now. And this is on my personal blog. And normally when I put out a blog post, I want it to be at least somewhat readable in terms of grammar and spelling and all that stuff. Because I’m typing it, I’ll run it through language tool or some other spell checker, grammar checker or whatever. And it will come out reasonably well. But I will say that I just talked a thousand words into existence in two minutes. I can’t type that fast. Despite the fact that I can type fast, I can’t type that fast. I have to admit that that’s pretty appealing. Being able to sit here in front of my microphone and put out a blog post in just a few minutes without actually having to type. That’s pretty cool.

What I don’t know is if it’s good enough. I don’t know if I can find it in myself to post something consistently that isn’t well enough written that people would want to actually read it. I don’t know about that. I’m going to try it a couple times because I have some blog posts that I’ve wanted to do, but I’ve been too lazy to type. So maybe I’ll just sit here in front of a microphone and speak those blog posts into existence and be happy about it. We’ll see.

Another thing that I think might actually happen here is that it might help me talk better. Even with this, I’ve tried to start thinking about what I’m saying out loud before I say it. Now, it feels like as a 41, nearly 42-year-old man, I should have come up with this idea long before now, but apparently it’s a new concept for me. Thinking before I talk. Being able to see what comes out after I’ve spoken it has led me to believe that I can do better by just kind of thinking before I speak the words out loud. And this is the second version of this blog post, and I think it’s going to be better than the first one because I did have that experience where I can think about what I say out loud before I say it, and then the outcome of it is actually a little bit better. Now, is it perfect? No. It still sounds like I’m talking to you guys through a microphone, and it’s just a bunch of text. But still, it’s better than it was, and I think that that could lead to some self-improvement in terms of how I actually speak out loud. We’ll see, but I think that that’s the case.

So that’s the end of the blog post. I’m going to go ahead and see how this turns out. I’m not going to do any editing other than punctuation. I’ll try to fix some of the punctuation, but other than that, I’m going to leave it as is. You guys tell me what you think, either here or on Mastodon. I appreciate it. We’ll see you next time.