Why did the Atari 8-bit computers make beeping noises while accessing the disk and cassette drives?

Question

If you've ever used an actual Atari 8-bit computer with either a cassette or disk drive, you would have noticed that it made a variety of "beeping" noises while accessing the devices. Specifically, loading a cassette made a "cluuug" sound repeatedly with small pauses and the disk drives made "bee-dee-bee-dee-bee-dee" type sounds while reading and "plut-a-plut-a-plut-a" type sounds while writing.

While it was good to have an audio indication as to what was going on, I never understood why the machine made the noise. It certainly wasn't the sound of the data coming off the device (since that'd presumably sound more like white noise), but the devices weren't as slow as (say) the Commodore 64 & 1541 where you could go have a cup of coffee in the other room and listen for when your game was finished loading.

What was the reason?

I don't think I've used an Atari emulator which reproduces these noises - I'd assume it is because the authors didn't see the need.

As a side note, I seem to recall that different drive manufacturers (e.g. Atari VS Rana) may have had different sounds. Is my memory correct? If so, that would say something more about the sound being a product of the device rather than the Atari.

UPDATE 2017-01-18: There are two answers to this question which suggest "the sound of the data" either leaking through or otherwise. I recently brought my Atari 410 deck back to life and have been loading converted CAS-to-WAV files via an iPod and a headphone jack-to-cassette adapter.

My memory from years ago was that you only heard a mix of silence and the "clooooooog" sound when the tape was being loaded. I was a bit surprised that instead of silence I was hearing the high pitched screech of the cassette data while loading tapes with the iPod, but it dawned on me that its because these converted WAV files are being played in stereo. So instead of data on track 1 and a silent track 2, I'm hearing the data screech from the stereo conversion since the normally silent other track is passed through to the TV speakers (as it would have been used for loading music, instructions, etc).

So what does this mean? You normally did NOT hear the data screech of a cassette, only the "cluuuuug" sound. When does that sound get made? If we assume the carrier signal (or '0') is 5327Hz and the "mark" (or '1') is 3995Hz, the sound only gets made when it detects a data stream (i.e. a mixture of 0's and 1's), though I suspect it is only making the sound when it detects a "mark" but perhaps rounded off a bit since it isn't as fine grained as I hear from the data stream.

All that being said, it doesn't answer the disk drive noise side of the question. Perhaps it is just simply more of "data packet being received over SIO" indication sound?

Answer 1

The beep comes from the OS ROM, and it is actually derived from the timing of the start and stop bits of each byte shifted into POKEY, this is determined from the interaction of the interrupts generated. The frequency output is nominally half of the operating bit rate, e.g. approximately 960Hz for a 19200 baud transmission (e.g. from a disk drive at normal speed). It is, in essence, the interrupt timer frequency, and it's spread across the last two channels (so that 16 bits can be used to define the divisor frequency)

You also hear a slight clicking sound before each beep, which is POKEY being re-set slightly before sending the 5 byte command frame to the peripheral. This also means that "writing" to the peripheral doesn't actually output any sound except for the initial re-set of pokey to emit the command frame. You can observe this by sending a very large block (say 16 kilobytes of data) over SIO at once with no interruptions.

This also means, the "raspberry" you hear, is actually a repeated hammering of the POKEY being re-set to send each 5 byte command frame.

Finally, the "sounder" can be enabled/disabled by setting the SOUNDR register (which lives in page zero at address 65 or $41 to zero)

Answer 2

The loading code does it; you can see that the CPU writes to one of the audio registers during the load if you disassemble the ROM. The "feature" could be disabled through code although it was usually reassuring during tape load.

As far as tapes go, the program was stored on one track and the second track was free for audio; I have only seen this feature used in educational products (learning languages, learning basic programming, etc).

Answer 3

For the tape it's mostly leakage from the other stereo track. Data is recorded on one track (in blocks) and the other track is output to the TV audio. This was intended to allow software to have an audio track (think of educational tapes)

I'll check tonight to see how bad it is on some pre-recorded stuff on the Atari 400. I've got a set of "learn french cassettes" in an recent ebay purchase so I'll see if it's as obvious with the other track recorded onto.

Answer 4

Are you sure it isn't the sound of the data? I have listened to the tapes using an audio player, and it sounds about the same. It is useful because you get immediate information about what is or isn't going on, without needing some computer program to interpret and provide information to the user.

On a tape, this is useful because you can tell if you started playing at the right point or not. It also lets you know if the data part of the tape is passed yet the computer had something go wrong and is failing to launch.

On a disk, this is useful because you learn what it sounds like when it's working properly or not working in various ways, and when it's doing something and when it's not. This is very handy for diagnosing problems such as slightly off plug connections or something.

From experience using them for many years, I'd say the audio was really helpful and they would've been more frustrating without the audio hints and reassurance.