This week’s newsletter describes a new technique for compressing Bitcoin transactions and summarizes an idea for privacy-enhanced transaction cosigning. Also included are our regular sections announcing new releases and release candidates and describing notable changes to popular Bitcoin infrastructure software.


  • Bitcoin transaction compression: Tom Briar posted to the Bitcoin-Dev mailing list a draft specification and proposed implementation of compressed Bitcoin transactions. Smaller transactions would be more practical to relay through bandwidth constrained mediums, such as by satellite or through steganography (e.g., encoding a transaction in a bitmap image). Traditional compression algorithms take advantage of most structured data having some elements that occur more frequently than other elements. However, typical Bitcoin transactions consist largely of uniform elements—data that looks random—like public keys and hash digests.

    Briar’s proposal addresses this using several approaches:

    • For the parts of a transaction where an integer is currently represented by 4 bytes (e.g., transaction version and outpoint index), these are replaced by a variable-length integer that can be as small as 2 bits.

    • The uniformly distributed 32-byte outpoint txid in each input is replaced by a reference to the location of that transaction in the block chain using its block height and location within the block, e.g. 123456 and 789 would indicate the 789th transaction in block 123,456. Because the block at a particular height can change due to a block chain reorganization (breaking the reference and making it impossible to uncompress the transaction), this method is only used when the referenced transaction has at least 100 confirmations.

    • For P2WPKH transactions where the witness structure needs to include a signature plus a 33-byte public key, the public key is omitted and a technique of reconstructing it from the signature is used.

    Some other techniques are used to save a few extra bytes in typical transactions. The general downside of the proposal is that converting a compressed transaction back into something that full nodes and other software can use requires more CPU, memory, and I/O than processing a regular serialized transaction. That means high-bandwidth connections will likely continue to use the regular transaction format and only low-bandwidth transmission will use compressed transactions.

    The idea received a moderate amount of discussion, mostly around ideas for saving a small amount of additional space per input.

  • Privacy enhanced co-signing: Nick Farrow posts to the Bitcoin-Dev mailing list about how a scriptless threshold signature scheme like FROST could improve the privacy of people who use co-signing services. A typical user of a co-signing service has multiple signing keys that are stored separately for security; but, to simplify normal spending, they also allow their outputs to be spent by a combination of some of their keys plus one or more keys held by one or more service providers who only sign after authenticating the user in some way. The user can bypass the service provider if needed, but the service provider makes operations easier in most cases.

    With scripted threshold signature schemes like 2-of-3 OP_CHECKMULTISIG, the service’s public key must be associated with the output being spent, so any service will be able to find the transactions it signed by looking at onchain data, allowing it accumulate data about its users. Worse, all currently used protocols we’re aware of directly reveal user transactions to the service provider before signing, allowing the service to refuse to sign certain transactions.

    As Farrow describes, FROST allows hiding the signed transaction from the service at every step of the process, from generation of an output script, to signing, to publication of the fully signed transaction. All the service will know is when it signed and any data the user provided to authenticate themselves with the service.

    The idea received some discussion on the mailing list.

Releases and release candidates

New releases and release candidates for popular Bitcoin infrastructure projects. Please consider upgrading to new releases or helping to test release candidates.

  • Libsecp256k1 0.4.0 is the latest release of this library for Bitcoin-related cryptographic operations. The new version includes a module with an implementation of ElligatorSwift encoding; see the project changelog for more information.

  • LND v0.17.0-beta.rc2 is a release candidate for the next major version of this popular LN node implementation. A major new experimental feature planned for this release, which could likely benefit from testing, is support for “simple taproot channels”.

Notable code and documentation changes

Notable changes this week in Bitcoin Core, Core Lightning, Eclair, LDK, LND, libsecp256k1, Hardware Wallet Interface (HWI), Rust Bitcoin, BTCPay Server, BDK, Bitcoin Improvement Proposals (BIPs), Lightning BOLTs, and Bitcoin Inquisition.

  • Bitcoin Core #28354 changes the default value of -acceptnonstdtxn to 0 on testnet, which is the default on all other networks. This change may help applications avoid creating transactions that are nonstandard and thus rejected by default nodes on mainnet.

  • LDK #2468 allows users to provide a payment_id which is encrypted in an invoice request’s metadata field. LDK checks the metadata in received invoices and will only pay if it recognizes the id and hasn’t already paid another invoice for it. This PR is part of LDK’s work toward implementing BOLT12.