
The following notes were taken from my audio recording + the video recording by Pantera (embedded at the end of this summary) of the Third-Party Developer meeting (TPVD) held on Friday, July 5th, 2024. My thanks to Pantera as always for providing it.
Meetings Purpose
- The TPV Developer meeting provides an opportunity for discussion about the development of, and features for, the Second Life viewer, and for Linden Lab viewer developers and third-party viewer (TPV) / open-source code contributors to discuss general viewer development. This meeting is held once a month on a Friday, at 13:00 SLT at the Hippotropolis Theatre.
- For both meetings: dates and times are recorded in the SL Public Calendar, and they re conducted in a mix of Voice and text chat.
- The notes herein are a summary of topics discussed and are not intended to be a full transcript of the meeting.
Official Viewers Status
- Release viewer: version 7.1.8.9375512768, formerly the Graphics Featurettes RC viewer dated June 5 and promoted June 10th.
- Release channel cohorts:
- WebRTC Voice RC, version 7.1.9.9688089989, July 1.
- Atlasaurus RC (object take options; improved MOAP URL handling), version 7.1.9.9620320242, June 27.
- Maintenance B RC (usability updates / imposter changes) 7.1.9.9555137545, June 21.
- Maintenance C RC (reset skeleton in all viewers), version 7.1.9.9469671545, June 14.
- Project viewers:
- None.
General Viewer Notes (Both Meetings)
- The switch to working from multiple viewer RC branches to a single development branch is continuing.
- This will mean in future there are likely to be fewer RCs in the pipeline than has bee the case for roughly the last decade (and sees the visible aspect of viewer development and release process swing back towards how it appeared prior to the switch to using RC channels).
- Parallel tracking of viewer development will continue for a while, given the fact there are currently four RC viewers in flight.
- The above change-over will not prevent contributions being accepted.
WebRTC
Summary
- A new project intended to move Second Life away from reliance on the Vivox voice service and plug-in, and to using the WebRTC communications protocol (RTC=”real-time communication”). Roxie Linden is leading this work.
- Key benefits:
- WebRTC supports a wide range of real-time communications tools in common use (e.g. Google Meet), supporting audio, video and data communications, and is thus something of a “standard” approach.
- Offers a good range of features: automatic echo cancellation, better noise cancellation and automatic gain control, much improved audio sampling rates for improved audio quality.
- Opens the door to features and capabilities to voice services which could not be implemented whilst using Vivox.
- In addition:
- LL are are of some of the security concerns around WebRTC voice (e.g. risk of eavesdropping, exposure of users’ IP addresses, etc), and is actively working to block these through the use of an internal proxy service.
- LL will be looking to Linux devs to help give feedback on how well WebRTC is working on their Linux viewers
- Feature requests for WebRTC should be made via the WebRTC board on the SL Feedback Portal.
Status
- Work has been on stabilising WebRTC and getting the viewer to RC status so that TPVs can look at it.
- Overall, development work is in a “wrapping-up” phase.
- Currently, LL is looking at August for a potential deployment across all of SL on the server-side.
- This will follow the usual approach of roll-out to the simulator RC channels first, then to the SLS Main channel.
- As a result, there will be some short-term issues around peer-to-peer, Group and ad-hoc voice connections between those on regions running the two different voice services (Vivox and WebRTC).
- Depending on how the deployment goes (e.g. first to a single RC, then multiple RCs, then the SLS Main channel), it is hoped that any such issues will only be for around 2 weeks.
- Viewers adopting the WebRTC code prior to or during this deployment period will be able to process both WebRTC and Vivox voice.
glTF Update
- As a result of the Firestorm PBR release, Runitai Linden has been revisiting the issue of memory use in the PBR-enabled viewer code on lower-specification computers.
- Geenz Linden is continuing to work on various glTF extensions, including glTF index of refraction (IOR) and transmission – with the work on the latter potentially being wrapped up.
- Cosmic Linden has been tweaking the PBR terrain work for PBR transforms on terrain (one transform per material).
- [Video 39:44-41:14] requests have been made for reflection probes in shapes other than cubes and spheres (e.g. cylinders, triangles, hemispheres) to account for more awkward interior space shapes (e.g. in roof areas). This is viewed as “highly unlikely” at the Lab.
- However, the mirror capability may be extended to include sphere reflection probes at some point “if all goes well”.
PBR Terrain Painting
- This is the next planned project for Cosmic Linden, and is in the very early stages of planning, so things are subject to potential change.
- Currently, the thinking is:
- The four PBR materials currently used for PBR terrain would remain available for use / painting.
- The painting element would allow a user to define how these materials are mixed, rather than having to rely purely on the the height map.
- E.g. if you have a paint map for a region, you’ll be able to blend the materials based on that, rather than having to use the height map, and define where areas of grass or rock or dirt, etc., appears on the ground.
- The paint map is likely to initially be on the basis of one blended texture at region level (not parcel), although the resolution of the texture is still TBA at the time of writing.
- The permissions for terrain painting will be based on ability to edit the height map (if you can alter the latter through the Region settings, then you’ll be able to use the terrain painting capability).
- Terrain painting will be a significant departure in how terrain texturing has been managed, requiring a new entity to be introduced. This is also still being thought through, but it is unlikely it will be a new asset type stored on the asset servers.
- [Video: 16:15-16:29] No decision has been made on making terrain painting open to scripted control.
- Cosmic is open to feedback on how this might be used, if enabled (e.g. a scripted explosion leaving a detonation mark on the ground for a period of time).
In Brief
- Reflections turn black when zooming in close is an issue which appears to be related to the use of mirror probes as well as “normal” reflection probes. If you are impacted by this, add your vote.
- Auto-exposure under PBR (adjusting the general scene brightness when looking at one or more very bright objects in the scene), with people interpreting it as a bug, particularly with the release of Firestorm PBR.
- It’s been suggested that a wiki page on how auto-exposure works should be produced, to which people can be pointed to help them understand what it is and how it works. Or potentially a debug setting or similar to disable (with a warning things may not display correctly as a result).
- Canny feedback has been requested on issues being encountered / suggestions for related features (such as the setting noted above).
- This led to an extending general discussion on lighting and rendering – please refer to the video.
- RLV/RLVa adoption by the Lab is still in discussion and described as something the Lab wants to do, but it does represent an extensive change to their viewer.
- As such, it is likely that as contributions are made, they will be pulled into the official viewer incrementally (presumably with some going behind debug flags until such time as they can be properly enabled).
- Overall, it now appears the Lab does want to support the full RLV/RLVa feature set, rather than just a sub-set thereof as had been previously indicated as a possible route.
- Opening the PBR terrain painting to scripted control led to a conversation on scripted weather systems – such as being able to change the ground appearance to match the season; having the ground appear to be covered by snow when it snows, etc. There have been requests for this, but it is not something the Lab is currently working on.
Next Meeting
- Friday, August 2nd, 2024 at the Hippotropolis Theatre.
† The header images included in these summaries are not intended to represent anything discussed at the meetings; they are simply here to avoid a repeated image of a gathering of people every week. They are taken from my list of region visits, with a link to the post for those interested.