Given that I've been doing almost nothing else for the last few days, I guess it's time I posted something about the new toy I've cryptically and irritatingly referred to a couple of times in recent posts. I really wanted to get a bit further in my experiments before committing to print but it's either this or some more speculation about Stars Reach and I'm pretty sure no-one wants that.
So, first up, just what is this new toy? Didn't I say it had something to do with AI? Not that again! I thought we'd all agreed it wasn't fun any more and not to bang on about it.
Ah, yes, but that was when the LLMs and their pals had mostly gotten their act together. They'd stopped behaving like surreal toddlers or like science fiction pets who could almost talk but not quite. They'd nearly reached the competent-but-boring stage and the novelty had worn off, yet they still hadn't ascended to genuine reliability.
Instead, they'd managed to plateau into being not much fun any more but also not quite reliable enough to be genuinely useful, while at the same time achieving a level of cultural ubiquity that made them obnoxious. It was probably time to leave them to cook a little longer in the hope they'd eventually either start working properly or go away, neither of which seemed all that likely to happen any time soon.
So that's what I did, mostly. I stopped playing around with the newer versions of the AIs I already knew about and didn't bother looking for any new ones. I kept a vague eye on developments via the Theoretically Media YouTube channel but that was about all.
My thoughts moved on to other things, one of which was a project I've been fiddling about with for years, something much more old-school than all this new-fangled artificial intelligence nonsense.
For around a decade, from the late 1970s to the late 1980s, I was somewhat actively involved in making music. I played guitar and sang in various bands - or would-be bands, not al of which ever left the rehearsal rooms.
It wasn't much of a musical career. There was a lot more practicing than performing and much of the time the songs I was singing were written by other band-members, not by me. God knows why they thought I'd do a better job of it it than they could but that's how it was.
Even though I didn't get many of them onto the set-lists, I did still write quite a few songs of my own and I occasionally used to record myself singing them. I didn't have anything even as sophisticated as those old four-track Tascam machines everyone had back then, although I thought often enough about buying one. I never did, though, so I just recorded myself through the inbuilt mic of my cheap boombox or ghetto blaster or whatever we called those things before it wasn't cool to call them that any more.
You can imagine what the sound quality was like. Also I mostly got the urge to do it when I came home after a night out, so some of the results weren't even listenable the next day, let alone forty years later. I also liked to buy those big packs of very cheap blank tapes, the ones that stretch and break. It's not hard to guess how any that survive are likely to sound after four and a half decades.
Because I'm a pack-rat and I never throw anything away, I still have almost every tape I made. Almost. One or two seem to have gone missing and of course one of those I remember being of particular interest. I persuaded my ex to sing a couple of songs on it. I always imagined most of my songs to be sung with a female voice but none of the bands I ever played in had girl singers. Probably just as well or I would have been out.
That one seems to have vanished. It's hardly surprising. They have not been carefully archived, just dumped into boxes and shoved under the bed. Some have lost their cases. It's a wonder there's as many left as there are. I also have nearly all the lyrics, handwritten originals, in a battered, yellow binder.
At various times over the last fifteen years or so I've made half-hearted attempts to "do something" with them. I once got so far as to listen to the whole lot and scribble the odd note about what was on them but I only ever listened to the minimum necessary to add labels like "1970s" or "horrible". Nothing useful like a track list.
I would quite like to know what's on them before it's too late. I'm aware how fragile and ephemeral magnetic tape, especially the cheap stuff used in 1970s and '80s compact cassettes, can be, which is why I've had the idea of digitizing my "archive" just about since I learned it was possible. I've actually tried to do it a few times but never with much success.
For a long time there were always tape players of different sizes and qualities lying about the house. I've connected most of them up to different computers over the years but largely failed to get anything that seemed worth the considerable effort it took to get them to talk to each other. I've bought devices, cables and contraptions that were supposed to make transferring recordings from tape to digital files easy but none of them did a very good job and all of them were more trouble than it felt the output was worth.
Last year, I tried once again, only to find that none of the old tape players I had worked any more and the various devices I'd bought were no longer compatible with modern operating systems. I had several very frustrating sessions, trying to get anything to speak to anything else, entirely without success.
Allegedly. Sounded too good to be true. I was highly suspicious. But there were lots of reviews praising it and some of them might not even have been by bots. Plus it was only about £25. Seemed worth a try.
It was. It works. Exactly as described.
It's not perfect. The controls are a bit fiddly, just as several people complained in the reviews. Sometimes you have to press the buttons a few times to get them to respond. The build is cheap and the sound quality, were you to want to listen to the tape through the speaker or headphones, is poor. But it works.
Play a tape, press record and a digital file is created on the USB drive supplied. Or any USB drive. Take the USB out and plug it into a PC and presto! There's the contents of the tape, all dressed up as an MP3. Granted, sometimes it's an MP3 that Audacity doesn't recognize but a quick pass through VLC's converter and it's all good.
Ah, yes. Audacity. I've installed a copy on every PC I've used for as long as I can remember but I've rarely found much use for it until now, when I'm starting to discover just how versatile and useful it is. As you might expect, some of these tapes are not in the greatest of shape after nearly half a century. They vary from tinny but clear to something that sounds like faint whisperings from a far-off star.
Audacity can fix all of that. At least, it can fix everything well enough for the purposes I require, which are twofold:
- I want to know what's on every tape
- I want to turn anything that's worth keeping into something that's also worth listening to.
The first is easy. I just have to be able to listen to them without having to actually listen to them. No-one wants to hear a drunken twenty-something caterwauling over the tuneless strumming of an out-of-tune guitar. That it's my caterwauling and strumming makes it no more appealing.
Having the files in digital form means I can skip through the running order in a few seconds instead of having to wait for the excruciatingly slow transport of a cheap cassette deck to trundle the tape back wards or forwards, while I constantly stop and start the machine to see if there's anything worth hearing. I also only have to listen to a few seconds to know if it's worth carrying on.
Once I find a take of a song I think might be worthy of further attention, I can excise it, name it and store it separately. There is a function on the device itself that supposedly separates the individual songs on the tape into discrete MP3s as it goes but, while I imagine that works perfectly well for pre-recorded tapes with definite gaps between tracks, it doesn't do much for the kind of stop-start home recordings I'm dealing with. I tried it but it turned out to be easier to just scan through the digital file visually and identify the breaks that way.
So much for finding the songs that have potential. That's the easy part. The hard part is turning them into something anyone would want to hear.
I'll be honest. Even with the full force of nostalgia in play and even given these are my songs and it's me singing them, there's precious little I'd want to hear twice and hardly anything I'd be willing to play to anyone else. I was never a very good singer or a very good guitar player and these primitive recordings, many made when I was under the influence of something or other, do not make me sound any better than I was.
For that, I always needed the help of others, preferably people who knew how to play. In my mid-sixties and with forty years between now and the last time I knew anyone who fancied pretending to be in a band, where am I going to find help like that?
And that, as I'm sure you've guessed, is where AI enters the story. But that's a story that will only be told when I'm good and ready. It shouldn't be too long, not with all the time I'm spending on it.
When it comes, there will be examples.
Consider yourself warned.
Posts
No comments:
Post a Comment