Starting a cassette label in the Japanese noise underground rarely begins with a manifesto. More often it begins with a stack of blank tapes bought at a discount electronics shop in Akihabara or Den Den Town, a second-hand dual-deck recorder, and a stack of J-cards printed on a home laser printer that jams every third sheet. The romance of the format—the hiss, the limited edition, the hand-numbered shell—obscures the actual labour: packing, addressing, calculating postage to seventeen different countries, and fielding emails about lost packages that the Japanese postal system has quietly consumed.
Every release demands decisions that are neither artistic nor glamorous. How many copies to make. Whether to use clear or coloured shells. Whether to include a download code when the entire point of the format is its physical limitation. Whether to sell through a distro or handle orders directly through PayPal—a platform that, as discussed in a previous article on payment processing for noise labels, can freeze accounts without explanation when it detects a sudden spike in small international transactions.
Manufacturing Constraints
The most common mistake new label operators make is underestimating the time required to duplicate tapes. Professional duplication in Japan is expensive and typically requires minimum runs of 100 units. Most DIY labels cannot afford that, nor do they need that many copies. So they turn to real-time duplication: a master tape played into a bank of cassette decks, recording at 1:1 speed. A C60 tape takes sixty minutes to duplicate. A C90 takes ninety. Multiply that by fifty copies and you have spent three full days monitoring levels, flipping tapes, and replacing decks when one of the mechanisms eats a cassette.
Blank tape stock has become unreliable. The last major Japanese manufacturer of ferric tape, TDK, ceased production years ago. Current blanks come from surplus stock, inherited collections, or Chinese factories whose quality control varies wildly. A batch of fifty tapes might yield forty-five usable shells and thirty-eight that record without excessive dropouts. The rest become test tapes, giveaways, or trash. Labels that claim every copy sounds identical are either using professional duplication or lying.
Distribution Realities
Selling directly through a personal website or a Big Cartel storefront gives the label maximum control over pricing and presentation, but it also means handling every single order.
International shipping from Japan is expensive. A single cassette in a padded envelope costs around 700 yen to send to the United States via small packet, and 1,200 yen via registered mail. The buyer pays this, but the label must absorb the cost of packaging materials and the occasional lost parcel. Many labels now require tracking on all international orders because the number of claims for non-delivery rose sharply after 2020.
Distros offer an alternative. A handful of European and American distros specialise in Japanese noise and experimental music. They buy at wholesale—typically 50% of the retail price—and handle fulfilment themselves. The label loses half the revenue but gains access to a customer base that does not follow Japanese Twitter accounts or check Japanese auction sites. The trade-off is visibility: the distro's catalogue buries the release among hundreds of others, and the label has no direct relationship with the buyer.
The Economics of a Small Run
A typical small-run cassette release in the Japanese noise scene costs roughly the following to produce, assuming fifty copies:
- Blank tapes: 8,000–12,000 yen (depending on source and quality)
- J-cards printing: 3,000–5,000 yen (home printer, ink, paper)
- Cases: 2,000–4,000 yen (jewel cases or norelco boxes)
- Postage for review copies: 3,000–5,000 yen
- PayPal fees: 3–5% of total revenue
Total production cost: roughly 16,000–26,000 yen. If the label prices each cassette at 1,000 yen and sells all fifty copies, gross revenue is 50,000 yen. After costs, the profit is 24,000–34,000 yen—enough to fund the next release but not enough to pay rent. Most label operators have day jobs. The ones who do not are either independently wealthy or deeply in debt.
Packaging and Presentation
Japanese noise cassettes have a reputation for elaborate packaging: hand-painted covers, inserts on washi paper, obi strips, sealed in plastic with a numbered sticker. This reputation is accurate for a minority of releases. The majority use standard jewel cases with a single-sided J-card printed on matte paper. The elaborate packaging tends to come from artists who already have a following and can afford the extra cost, or from labels that treat packaging as the primary artistic statement rather than the music.
The practical consideration is weight. A cassette in a jewel case with a multi-page booklet weighs roughly 80 grams. Shipping five of them pushes the package into a higher postage bracket. Labels that prioritise elaborate packaging often lose money on every international sale because the shipping cost exceeds the margin. Some have started offering a "no packaging" option: the cassette in a plastic sleeve, shipped in a bubble mailer, no case, no insert. It is not collectible, but it is affordable.
Dealing with Collectors
The collector market for Japanese noise cassettes has grown significantly since 2015. Discogs listings, eBay auctions, and dedicated Facebook groups track every release. A cassette that sold for 1,000 yen upon release can appear on a resale site for 8,000 yen two years later if the artist gains notoriety or the run sells out quickly. This creates tension. Some label operators deliberately limit runs to 30 copies to create scarcity. Others press 100 copies and watch 70 of them sit unsold in a cardboard box under the bed.
Collectors also drive demand for represses. A label that refuses to repress a sold-out release is accused of elitism. A label that does repress is accused of diluting the value of the original. There is no correct answer. The only honest approach is to decide at the outset whether the label exists to serve listeners or to serve collectors, and accept the consequences of that choice.
Archiving and Documentation
Few DIY labels maintain proper archives. Masters are stored on hard drives that fail. J-card artwork exists only as a PDF on a laptop that was replaced three computers ago. Contracts with artists are verbal or exchanged over now-deleted Twitter accounts. This fragility is not unique to Japanese noise—it is endemic to the entire DIY ecosystem—but it is more acute in a scene where many participants treat the release as ephemeral by design.
Some labels have begun uploading digital masters to the Internet Archive as a hedge against physical decay. Others refuse, arguing that the release is the physical object and a digital file is a different work. Both positions are defensible. The practical outcome is that a significant portion of Japanese noise releases from the 2010s are already unlistenable outside of private collections. The tapes shed oxide. The shells warp. The masters are lost. The music exists only in the memory of people who were there.
The Future of the Format
Cassette production for noise and experimental music shows no sign of stopping, but the infrastructure is thinning. Fewer shops carry blank tapes. Fewer postal services offer affordable small-packet rates. Fewer listeners own functioning tape decks. The format persists because it is cheap to produce, easy to distribute in small quantities, and physically distinct from streaming. It will not disappear, but it will become more expensive and more difficult to do well.
One label in Osaka now includes a QR code on every J-card linking to a digital backup of the master. Another in Kyoto prints a small field recording of the duplication process—the clicks and hiss of the decks—that plays when the code is scanned. These small gestures acknowledge the fragility of the format without pretending the tape will last forever. The cassette becomes a letter, not a product. When a listener in Buenos Aires or Berlin writes to say the tape arrived, that is the transaction that matters. The economics are broken, but the connection is not.