Analogue Audio on Floppy Disks

Storing analogue audio on floppy disks using an unmodified drive.

[@wichtelchen]

For years, I thought, I was the only one who stored music on floppy disks. But in the 90's, I did it in the regular digital way using a custom codec I made in Qbasic. But storing it in an ANALOGUE way on a storage that is meant to be digital... without modifieing he drive.... is just... damn... genius :D

[@michalzustak8846]

What was your codec like? Please elaborate, this is cool as hell.

[@wichtelchen]

@@michalzustak8846 When I tinkered it, I was 14 years old xD. It was in the 90s. The computer we had was a 486DX66 without internet connection. I tried to figure out how MP3 works and I tried to recreate it. But I was totally wrong (MP3 works completely different). But my implementation did its job (Basic, with many goto statements in it xD). Today, I would say, it is some sort of 1-Bit ADPCM. But in contrast to other ADPCM codecs, the step size was chosen on time domain only, because there is only one single bit per sample which you need to determine the direction the signal goes (up or down). Stepping into the same direction again increased the step size. Stepping back and forth reduced it. The "sound driver" for playback and recording was a loop that queried from the parallel port (where the "sound card" was) at a given sample rate. It still worked pretty well at low bit rates (say, 48kbps. Sometimes better than MP3). The bitrate was chosen by the recording sample rate (because of 1 bit per sample). But it sounded hissy and scratchy^^. That didn't bother me much because the way I recorded the music into it was much more lossy: I barely remember the "soundcard" was a parallel-port-"stick" and I used a crappy microphone to record into that from the speakers of a cassette player. I didn't manage to support sound blaster cards because QBasic had no built-in way to access these (and I had no idea how to circumvent that limitation using assembly language xD). As a special feature, the files had something similar to a GIF stored too (usually some pixel art that was dancing to the music), which was displayed during playback. That was an adventure... the pure exotic island solution. xD But I was surprised once I discovered the opus codec shares one strange limitation with my codec: The only supported sampling rates are 24kHz and 48kHz, but not 44.1kHz. It upsampled to 48kHz too when it came to encoding of 44.1kHz material (which never really happened because it just queried the parallel port sound card directly 48k times per second). :D.

[@jamesslick4790]

WAYYY Back in the early 80's when my computer was a TRS-80 Model III, (using 5.25" disks), I pondered recording (analog) audio to such a disk/drive. I brought it up to a friend of mine (who actually worked in a recording studio), He pointed out that the overall magnetic surface area,being tiny compared to a reel (or even cassette) of tape, it would likely only hold a few seconds of audio, so it was a useless idea. I didn't CARE that it was useless, Just had to be COOL! , Now 35+ years later, someone DID it! and on a disk with even less surface area- Still COOL!

[@NinjaSushi2]

lol Hell yeah. Life came full circle.

[@willyarma_uk]

Very clever! The I/O level might be digital but the timing is fully analogue, brilliant way of exploiting that fact by using FM.

[@danielmcanulty1562]

Oh wow, great work using both sides of the diskette! I always blew out both sides with my writing technique.

[@nochan99]

Giving the floppytron a run for it's money!

[@Popclone]

Still surprised this video dose not have 12 million, oh well, it will one day :)

[@itsmememeking3387]

I have been searching every where for anyone to record audio onto a floppy disc only u did it!!!! Pls make a tutorial

[@HelloKittyFanMan.]

Well, _analog_ audio.

[@simonemastroianni1985]

I have a bunch of floppy drives (3.5" drives) that I want to experiment. The problem is that i cannot find anything on the internet about how to rotate the spindle motor. How do you spin that motor? The only thing that i can find is modding the drive internally or get a 5.25" drive. Thanks

[@AureliusR]

Huh? The spindle motor is driven by a control signal on the connector... just look up the floppy drive cable pinout and the protocol specification. It's absurdly simple.

[@EdwinvandenAkker]

This is so cool! Next steps: • use a compression chip on data to store more audio • use a buffer to avoid those syncin-pulse-stops 🤓

[@straightpipediesel]

That's basically copying a MP3 to a floppy.

[@HelloKittyFanMan.]

Wow, that's cool! So in this mode the drive doubles as a metronome!

[@Popclone]

MetroBeat. Lol

[@CassetteMaster]

That is freaking amazing!! Sometime, I plan on doing something similar ( did a little, but have not picked back up on it yet) to make a floppy recorder that uses analog tape recorder circuitry hooked straight to the head.

[@kitsune-denshi]

Let me know how this goes, that was another avenue I was considering but didn't want to go through the hassle of modifying the drive too much. It would be interesting to see what you can come up with in terms of spindle speed and head positioning.

[@eDoc2020]

Jeri Ellsworth did this years ago with interesting effects. You can see the video on her channel. It has the video ID Xpr7B-7BFP4.

[@eDoc2020]

Or the full link is kzbin.info/www/bejne/jqHVaHVjbKd5hpY

[@jacobsteel]

Thank you so much for sharing this. I wondered for years how this would sound! :D

[@papafrank808]

Have you see the Floppy Beatbox from Techmoan? Midi on a Floppydisc is very interesting.

[@HelloKittyFanMan.]

MIDI on a floppy[space]disk is "interesting"? Well, sure it was when I first saw MIDI keyboards that came with floppy disk drives, and then eventually got one, but by the time I had one they were commonplace, and now MIDI keyboards have USB or/and SD card slots instead.

[@michalnemecek3575]

10:37 if this ever became a commercial product (even for memos), its creators would probably either create their own drive that would output data during the sync pulse, or put in some chip that would interpolate the data, just like CD players do with damaged discs.

[@lizichell2]

This is a great little experiment

[@HelloKittyFanMan.]

Rewatching, and adding this comment to mine from 4 years ago: Some of them do have their own brains. For example, many Commodore drives and compatible models do.

[@newkfromrotterdam]

Nicely figured out and good soundquality! (as far as i can tell from that microphone recording of that little speaker) Maybe you can circumvent that pulse in the audio signal like Jeri Ellsworth did in her "Floppy Drive Reverb" video..

[@kitsune-denshi]

Jeri hooked up to the analogue heads directly, whereas I am going through the full digital interface of the drive. This blanks out the data lines during the index pulse, so sadly not an easy fix.

[@mercuryvapoury]

what happens if yiou put a magnet on only part of the recorded disk? Do you just get silence, or any tupe of strange interference?

[@lwilton]

Thank you! A little insanity is always nice to see! :-) Boy that sure takes me back. I remember fighting floppy controllers form various makers to deal with alien formats and the like for what seemed like years at the time. And probably was. But that was in the 8" and 5.25" days more than 3.5" days. Out of curiosity what was on that Unisys disk? I wasn't able to make out the part number or description.

[@kitsune-denshi]

Thank you - the Unisys disk is a rather boring MS DOS 3.3, so nothing special I'm afraid.

[@jamesslick4790]

@@kitsune-denshi DOS is BOSS! 👍😊

[@briant333]

That is cool. The only way I ever managed to play audio from a floppy disk out of a pc was to copy a low quality mp3 onto a floppy disk, then using a ls120 drive and usb to ide adapter i played the mp3 file on a car stereo with usb input.

[@wichtelchen]

In the mid-90's, I cobbled together a custom 1bit-ADPCM codec (using QBasic) that made it possible to store up to 5mins of digital audio onto a floppy disk. :D MP3 did (and does) glorious fail at lowest bit rates. The most difficult thing was to bring the music into the computer because the computer we had in the 90's only had floppy drive and a small hard drive.

[@alexandruianu8432]

@@wichtelchen These days, you could use the Opus codec to write good quality audio at lower bitrates. I've gotten 6 minutes on a floppy at good quality by down mixing to mono. Around 55 kb/s (3.5 min) is where it starts getting better to go mono rather than stereo. Quite interesting to see that the bigger drives got, we also improved our ability to use smaller ones.

[@Ice_Karma]

It sounds like there's two dropouts per track step; are you recording on both sides of the disk before stepping to the next track?

[@kitsune-denshi]

Yes, it's one side, then the other side, and then step to the next track.

[@AiOinc1]

I'd always wondered if this was possible, and I'd had thoughts of doing it to ye olde MFM/RLL hard disk drives, which work in much the same way.

[@retrotechandelectronics]

I tried this on commodore drives like 30 years ago!!!

[@manobit]

What an interesting work!

[@coondogtheman]

How long can you record for? you could maybe eliminate that pulse noise with some type of filter.

[@kitsune-denshi]

32 seconds, which is determined by the rotational speed (300rpm, 200ms), number of tracks (80) and sides (2). Which, incidentally, is why I chose a 3.5" drive since they rotate more slowly than 5.25" drives (360rpm). The clicking pulse could probably be improved somewhat by filtering.

[@coondogtheman]

@@kitsune-denshi if you spun the disk more slowly would the sound quality be the same maybe worse. Thinking of the pxl-camera that recorded on audio tape.

[@lizichell2]

Amazing

[@swilwerth]

Maybe sample and hold to avoid the track switch noise.

[@ildafons]

How many minutes of analogue audio can be recorded on 3.5?

[@Popclone]

An idea, you can make a mix in a program like Fruitloops and mix it with the thump of the floppy drive head beat. The drive thump will have to be original independent as its now. Real time.

[@10p6]

Interesting experiment, :-)

[@herrschwarz485]

is the latter recorded song esperanto anthem ? It sounds so similar !

[@nikotinko]

It's Ode to Joy from Beethoven, you probably heard it hundreds of times.

[@electrifyingelectron9792]

Amazing .. You are a genius .. Keep up the great work .. Thanks a million ..

[@beamfinder8336]

well done

[@sonjahuber3043]

Do you have schematics for that?

[@sepic5407]

up

[@Tomsonic41]

So the audio is recorded in a bunch of concentric tracks, rather than a spiral like on vinyl records. So long as the head moves at the exact same time as when it was recorded, you get a generally uninterrupted audio stream. Would it be possible to make the head move smoothly instead of in 'jumps' so it records a spiral?

[@LazoeJSCREI]

I don't think it's possible, the head is moved by a stepper motor, and like the name implies, it steps. You could possibly get it to do really small steps really fast but that would introduce synchronization problems, and it would never be truly smooth

[@kitsune-denshi]

Interestingly, I *think* the stepper motor wouldn't be the problem. As long as it moves repeatably and follow the same path between tracks, there is no reason why it shouldn't record (and read) as it moves between tracks. The main reason for the annoying beating sound is that the drive electronics don't output valid data during the index pulse, which is long enough to throw off the PLL. I'm not entirely sure if data isn't written during the index pulse, either.

[@alijabari8715]

awesome!

[@duality4y]

do you send the audio as fm to the floppy disk ? do you have any schematics?

[@kitsune-denshi]

Yes, it's simple frequency modulation and I'm afraid I don't have the schematics for it.

[@duality4y]

@@kitsune-denshi thats to bad would have liked to have compared it to mine. I cant get my audio quality good enough. How do you generate the signal? With 555 timer or a pll? (Never mind just watched and you use a pll)

[@kitsune-denshi]

@@duality4y yes, it's a CD74HC4046 PLL - very simple to setup up.

[@samplesmasher]

Jeri Ellsworth did this in 2009...

[@kitsune-denshi]

Yes, although she hooked up the read/write head to the electronics of a tape recorder. There are few more hoops to jump through (frequency modulation) to use the digital interface and an unmodified drive.

[@fedevek]

Great! Can we do the same using a Hard Disk? Can you make a tutorial?

[@kitsune-denshi]

In principle, yes. The same method will work for a hard disk with ST506 or ST412 interface. Back of the envelope calculation suggests that slow-scan video is definitely on the cards with a hard disk.

[@fedevek]

@@kitsune-denshi Can you send me the scheme of the electronic circuit? I'd really like to try your experiment.

[@jamesslick4790]

@@kitsune-denshi WOW! Analog SSTV on an HDD! My geek side of my brain is awake now! 😎

[@matiasdominguez598]

♥

[@cbrunnkvist]

Now imagine if we can encode data as modulated audio signal - then we could use floppy disks to store DATA 🤭

[@Pickelhaube808]

I would want to do this with 8" media...

[@Afrocanuk]

The problem with the 1.44Mb floppy disks is the enormous amount of wasted space on physical disk. This is why it will be a much better idea to put the LS-120 drives back into production. They use 120Mb floppies & are backwards compatible to the 1.44Mb floppy. LS-120Mb superdisks are also an ideal upgrade to the conventional cassette tapes.

[@newkfromrotterdam]

links?

[@sepic5407]

links to schematic?

[@papafrank808]

You have to do that with MOD music. Then you can save much more music to the floppy disk.

[@MegaBobsel]

Yes but that wasn't the purpose of this video. Storing digital files on a floppy is the intended use, storing an analog signal is what makes this interesting

[@HelloKittyFanMan.]

Now maybe you'd suspect that one of us would ask you to cross try those disks, right? Show us what one of those analog music disks pulls up on a computer, both in standard directory view and in sector editor view. And then take a disk with files on it and play it in this! What, oh, this is a year and a half old and you don't still have the circuit built? If so, then will you please rebuild it for this request?

[@kitsune-denshi]

A computer will show an unformatted disk, because it will not be able to find the expected sector markers and other items it would use to recognise the format. The other way round, a computer disk comes out as noise, which I didn't find exciting enough to record. Sorry :)

[@HelloKittyFanMan.]

OK, @@kitsune-denshi, yeah, that makes sense about what this does to the digital formatting. Interesting. Also, when you have people asking if you would please show us what the noise sounds like, even if you don't find that exciting enough for yourself, will you please do it for your viewers?

[@ChrisTheMagical12]

Now mod the drive to make it smooth.

[@fortitude9932]

If it's saved to floppy it's 0 1 digital data

[@kitsune-denshi]

Like someone else said, it's digital (quantised) signal levels but continuous (non-digital) timing, so you can argue either way. However, given that frequency modulation is not typically considered digital, I think it's justified to call this analogue.

[@HelloKittyFanMan.]

"FM-modulated version"? Oops, that would be like saying "frequency-modulation-modulated version." Just "FM version" would suffice.