seshat VS nnnoiseless

> Can you already say something about encrypted search support in Element X?

It's planned, and should be way easier to do than on old Element given we can hook the clientside search indexer (https://github.com/matrix-org/seshat), which happens to be in Rust, directly into matrix-rust-sdk and thence both EIX and EAX. Whereas previously we'd have had to do custom integs with Seshat on EI and EA like the current one on Element Desktop, which just becomes a maintenance nightmare. Hopefully should be ready in the next months.

> Also since you mentioned it, is multiaccount support something that we can realistically expect to see in Element X someday (maybe even in a v1.0?). If I remember correctly for the previous iterations of the clients the status was mostly "maybe one day" or "the architecture does not really support it".

Definitely. We actually had it pre-Element in Matrix Console, and then it got lost in Element. The hope is to get it into Element X relatively early.

https://github.com/matrix-org/seshat - we’ve had it for years on desktop; just need to hook it up on mobile. it’s currently blocked behind unifying sdks, so we don’t have to do it twice in iOS & Android.

Excellent noise cancellation (and background noise elimination, microphone scratch noise elimination etc). Ideally you need something like https://krisp.ai/ or https://workspaceupdates.googleblog.com/2021/06/background-n... in the mix - but doing this in an E2EE-friendly and privacy preserving manner is Hard. However, just like we solved E2EE full text search by doing it clientside and making the indexes gossipable between your clients (https://github.com/matrix-org/seshat), we'll have a go at doing something similar for this problem too.

The competitive gap with Discord in terms of media quality is probably something like:

* Need a low-latency SFU. This should be very doable; not only are there a lot of good FOSS SFUs to build on top of these days, the history of the Matrix team is actually that we built VoIP stacks fulltime before we shifted focus to Matrix, and we've built MCUs and media servers of all flavours in the past. MSC3401 should also give us a competitive edge given latency will be automagically minimised by using the physically closest decentralised SFU, and letting anyone bring their SFU to the party.

* Needs a SFU with good rate control (and/or FEC). This is probably the single most important thing to get right in terms of quality. Signal wrote up a good overview of why: https://signal.org/blog/how-to-build-encrypted-group-calls/

* Excellent noise cancellation (and background noise elimination, microphone scratch noise elimination etc). Ideally you need something like https://krisp.ai/ or https://workspaceupdates.googleblog.com/2021/06/background-n... in the mix - but doing this in an E2EE-friendly and privacy preserving manner is Hard. However, just like we solved E2EE full text search by doing it clientside and making the indexes gossipable between your clients (https://github.com/matrix-org/seshat), we'll have a go at doing something similar for this problem too.

* Excellent automatic gain control. The importance of normalising/compressing everyone's audio so they're equivalent loudness is really important.

We're also in the process of adding in spatial audio (unsure if Discord has that) which should help a tonne with distinguishing the different audio feeds.

We can probably also be more bullish about supporting new audio codecs like Lyra.

Messages are stored in encrypted form on the Beeper server and the Beeper client has a local search index (the same one used by Element desktop: https://github.com/matrix-org/seshat)

1. So the SFUs we're currently looking at are yours, ion-sfu (and/or galene) and mediasoup. Honestly we haven't finished looking at how they compare for rate control, but the Pion team seems very interested in ensuring they have good rate control.

2. From context I think you're talking about noise cancellation here? I assumed that some of the more exotic ML-based ones ran serverside, which obviously is incompatible with E2EE. It sounds like there are a bunch of options for running WASM-based intelligent noise cancellation clientside though, especially with MediaStreamTrackProcessor and friends. rnnoiseless as a pure Rust->WASM port of rnnoise looks fun, for instance: https://github.com/jneem/nnnoiseless

3. True, although given Google are highly motivated to make AEC work properly in WebRTC, I guess I'm hoping that they'll continue improving it, much as they have been. I certainly never want to have to write or integrate one ever again :D

Very nice! I've just updated [`nnnoiseless`](https://github.com/jneem/nnnoiseless) to rustfft 5.0, and got a cool 20% overall improvement.