Forty-Two Marks
A group submitted three hours of recordings with forty-two marked regions. We identified forty-one of them. The forty-second is nine tenths of a second long.
A group calling itself a paranormal research society sent the Survey three hours and twenty minutes of audio on a memory card, with a covering letter, a hand-drawn plan of the hall, and a printed list of forty-two timestamps they wanted looked at. The letter was courteous, the plan was good, and the list was numbered and cross-referenced to the recordings, which is more than we get from most professional referrals.
They wanted to know what was on the tape. This is what was on the tape.
The review
Marks were reviewed blind of the group's own annotations, then compared. Each region was listened to at normal speed and at a quarter speed, viewed as a spectrogram, and checked against the two hours of the recording either side of it. Where a source could be identified elsewhere in the same recording with the same signature, that is recorded as the attribution.
| Time | Instrument | Reading | Entry |
|---|---|---|---|
| — | 09 | 14 marks | Building services. Boiler firing, boiler pump, and a fridge compressor in the kitchen with a 240 s duty cycle that appears throughout. |
| — | 09 | 9 marks | The group themselves. Whispering, clothing, a chair, and one clearly audible instance of a member saying “did you get that” four seconds after the mark they made. |
| — | 09 | 6 marks | Borborygmus. Digestive noise, one person, close to the recorder. |
| — | 09 | 5 marks | External. Two car doors, one aircraft, one dog, and a wood pigeon on the roof at 03:11 which is on the recording for eleven minutes and was marked once. |
| — | 09 | 4 marks | Recorder artefacts. Automatic gain control pumping after loud events, twice; handling noise, twice. |
| — | 09 | 3 marks | Structural. Timber and a radiator cooling. |
| 02:14:33 | 09 | 1 mark | Not attributed. 0.9 s. Band-limited 240–1100 Hz, three amplitude peaks, no harmonic structure resolvable at this level. |
Full attributions for all forty-two marks are in the file. The six marks attributed to borborygmus are at 00:41:12, 01:07:55, 01:38:02, 02:03:19, 02:44:50 and 03:01:07, and are the same speaker throughout.
The forty-second
It is nine tenths of a second long. It sits 6 dB above the local floor at the time, which is within the range of the digestive events and below the range of the whispering. It has three amplitude peaks about 240 ms apart, which is roughly a syllable rate, and which is also roughly the rate at which a great many things that are not speech happen.
The group heard two words in it. Both members who wrote to us afterwards heard the same two words, which they had discussed with each other before writing.
The Survey's reviewer heard nothing in particular in it, having been given the mark without the suggestion. When told the two words afterwards he heard them immediately and was subsequently unable to stop hearing them, which is the ordinary and well-documented behaviour of a human being and not a fact about the recording.
Why the determination says “very probably”
Because it is not certain, and the Survey does not have a way of making it certain, and writing certainly in a file where the evidence supports probably is how an archive stops being worth anything.
The reviewer's recommendation was that the mark be recorded as unattributed and the probable attribution stated. The reviewing archivist queried the phrase on the grounds that it invites exactly the reading the group put on it. The phrase stands, with the query on the file, which is why you are reading about the query.
The part that is actually useful
Forty-one of forty-two marks were made by people who were sitting in the room where the sources were, at the time the sources made the noise, and did not notice. That is not a criticism of them; it is how attention works. You cannot hear a fridge you have been sitting next to for ninety minutes. The recorder can.
This is the strongest argument for recording at all, and it is not the argument people usually make for it. A recorder is not a device for capturing things the room does not contain. It is a device for capturing the things the room does contain that you have stopped being able to hear — and if you do not go and identify every one of them before you start marking regions, you will mark the fridge, and you will mark it as a voice, because at −49 dBFS with the gain up a fridge sounds exactly like somebody standing in the kitchen deciding whether to speak.
Cause established.
Forty-one of forty-two marked regions are attributable to identified ordinary sources. The forty-second is 0.9 s long, of insufficient duration and bandwidth to characterise, and is very probably the same source as marks 12, 19, 24, 28, 31 and 37.
The Thinplace Survey is invented and so is this file. The case files are horror fiction written in the format of a field report. The instruments they describe are real and are explained in full; the incidents are not.
Fiction