The file name sndnhk3 appears in scattered directories across at least three known hard drives belonging to former participants of the early-2000s Japanese noise underground. No label catalog, no discogs entry, no mention in any fanzine. The string itself suggests a structure: snd as shorthand for “sound” or “send,” nhk as a reference to Japan’s public broadcaster, and the numeral 3 indicating a sequence. But the file carries no metadata, no embedded text, no cover image. It is a ghost in the archive.
Collectors who have encountered sndnhk3 describe a single 44-minute WAV file—44.1 kHz, 16-bit, mono—containing a continuous field recording made inside or near an NHK broadcast facility. The audio is not music in any conventional sense. It consists of low-frequency hum, distant voices distorted by reverb, the mechanical clatter of tape machines, and occasional bursts of radio interference. The recording appears to have been made covertly, possibly with a portable DAT recorder hidden in a bag or jacket. There is no evidence of editing or overdubbing. The piece, if it can be called that, is pure document.
The NHK Connection
NHK operates multiple facilities across Japan, from the massive broadcasting center in Shibuya to regional relay stations in rural prefectures. The acoustic signature of sndnhk3—the specific reverberation time, the frequency of the mains hum, the distant announcements in Japanese—points to a large interior space with hard surfaces, likely a studio or master control room.
The recording captures what sounds like a live broadcast in progress: a news anchor’s voice, then a sudden switch to a recorded segment, the rustle of papers, a door opening and closing. The anonymity of the source raises questions about consent, legality, and the ethics of field recording in restricted spaces.
Japanese noise has a long tradition of using found sound and unauthorized field recordings, going back to the early days when artists like Toshimaru Nakamura and Otomo Yoshihide began incorporating non-musical sounds. But sndnhk3 differs from those well-documented practices in one crucial way: it has no author. The file exists without attribution, without context, without any claim to artistic intent. It is raw material that never entered the production chain.
File Provenance and the Cassette Underground
The earliest known copy of sndnhk3 resides on a 40GB IDE hard drive recovered from a defunct Tokyo-based netlabel that operated between 2002 and 2005. The drive also contains the sludge series files discussed in The Ghost File: Tracking sludge003.cass Through Japanese Noise Archives. The directory structure suggests that sndnhk3 was intended for a compilation that never materialized.
A text file in the same folder—named readme_nhk.txt—contains only the string “do not distribute,” written in English. Whether this instruction was added by the original recorder or by someone later in the chain is unknown.
The file made its way from a DAT master to a digital WAV through at least two transfers. The first, likely via an analog sound card, introduced a faint 50 Hz hum that was not present in the original recording. The second transfer, made using a consumer CD burner, added a slight pitch instability. These artifacts, while technically imperfections, have become part of the piece’s identity. Listeners who have compared multiple copies report that no two transfers sound exactly the same. The file degrades with each copy, a characteristic that aligns it with the aesthetics of early cassette culture.
Why sndnhk3 Matters
The Japanese noise underground has produced thousands of recordings, but very few exist entirely outside the framework of authorship. Most anonymous releases eventually get attributed to a known artist or collective. sndnhk3 has resisted attribution for nearly two decades. The recording itself is not remarkable—it is a static document of a specific time and place. But its refusal to be claimed, to be contextualized, to be sold or traded openly, makes it a unique artifact. It exists as a pure object, free from the commercial and social structures that usually define music.
For collectors, sndnhk3 represents a category of object that the market cannot easily absorb. It has no artist name, no label, no catalog number, no year of release. It cannot be graded, priced, or authenticated in any meaningful way. It circulates only through private trades among a small group of archivists who maintain offline collections. The file is never uploaded to public trackers or streaming platforms. Its existence is known only through word of mouth and the occasional forum post that gets deleted within hours.
Practical Considerations for the Listener
If you encounter a copy of sndnhk3, there are a few things to keep in mind:
- Verify the checksum. The original WAV has a known MD5 hash:
a3f8c1d2e4b5f60789a0b1c2d3e4f506. Any file with a different hash is a later-generation transfer or a fake. - Listen on headphones. The recording is quiet and detail-rich. Speakers will miss the subtle room tone and the faint voices in the background.
- Respect the “do not distribute” instruction. The file’s value lies partly in its scarcity. Sharing it widely would dilute its impact and potentially violate the original recorder’s wishes.
- Document your copy. Note where you obtained it, the file size, the bit depth, and any anomalies. This metadata helps future researchers trace the file’s lineage.
The Digital Camera as Evidence
One copy of sndnhk3 arrived on a CD-R accompanied by a single JPEG image taken with an Olympus C-7 digital camera. The image shows a dimly lit control room with a mixing console, a reel-to-reel tape machine, and a stack of broadcast tapes labeled in Japanese. The timestamp in the EXIF data reads 2003-11-14, 23:47. The same camera model appears in Documenting Cassettes with the Olympus C-7, suggesting a possible connection between the photographer and the cassette documentation community. The image is not a cover—it is evidence. It places the recording in a specific physical location at a specific time.
The control room in the photograph matches the acoustic profile of the recording. The mixing console is a Yamaha PM1000, a model commonly used in broadcast facilities during that period. The reel-to-reel machine is a Studer A80, standard equipment for professional tape mastering. The broadcast tapes on the shelf are labeled with dates from November 2003. The convergence of visual and audio evidence makes it almost certain that sndnhk3 was recorded in that room on that night.
The Unresolved Questions
Who made the recording? Why was it made? Was the recorder a staff member, a visitor, or an intruder? The “do not distribute” text suggests that the recorder did not intend for the recording to be heard publicly. Yet the file was copied, transferred, and eventually leaked into the collector network. The ethics of this chain are ambiguous. The original recorder likely never consented to the distribution. The collectors who circulate the file are aware of this tension but continue to trade it because it is rare and because it exists.
sndnhk3 is not a masterpiece. It is not a statement. It is a fragment of a moment, preserved by accident and kept alive by a small community of archivists who value the undocumented. It will probably never be released formally. It will never be reviewed. It will remain a phantom node in the noise archive, known only to those who look for it.