2022 TPV Developer meeting summary, week #3

The following notes are taken from the TPV Developer meeting held on Friday, January 21st, 2022.

These meetings are generally held every other week.  They are recorded by Pantera Północy, and her video of the meeting is embedded at the end of this report – my thanks to her for allowing me to do so – and it is used with the chat log from the meeting and my own audio recording to produce this summary, which focuses on the core topics discussed.

SL Viewer

[Video: 0:10-7:37]

• The two Maintenance RC viewers Jenever and Koaliang, have been combined into a single Maintenance RC viewer, version 6.5.3.567451, issued on January 20th.

Important note: the above viewer version has a significant issue that may prevent users on this cohort from logging in. See this report for details. The recommendation is for those wishing to avoid the issue is to download and install the current release viewer (or if experiencing issues, to contact Support and request and inventory fix). 

The rest of the current list of official viewers remains as:

• Release viewer: version version 6.5.2.567427 – Mac Voice hotfix viewer, January 13.
• Release channel cohorts (please see my notes on manually installing RC viewer versions if you wish to install any release candidate(s) yourself).
  • The Tracy Integration RC viewer version 6.4.23.563771 (dated Friday, November 5) issued Tuesday, November 9.
• Project viewers:
  • Mesh Optimizer project viewer, version 6.5.2.566858, dated January 5, issued after January 10.
  • Performance Improvements project viewer version 6.5.2.566967, dated December 17.
  • Performance Floater project viewer, version 6.4.23.562625, issued September 2.
  • Legacy Profiles viewer, version 6.4.11.550519, dated October 26, 2020.
  • Copy / Paste viewer, version 6.3.5.533365, dated December 9, 2019.

General Viewer Notes

• The combined Maintenance RC viewer is likely the next viewer in line for promotion.
• The Graphics improvements viewer still has some bug to be fixed prior to moving to RC status. In particular, Euclid Linden is working on fixing the frame stall issue resulting from a media texture update. Essentially, if vsync is enabled, then command buffer resources aren’t as unbounded as they are with vsync disabled, resulting in textures copied to it a call to update media textures exhausts all available resources, effectively blocking it until it is flushed, rather than the buffer simply being discarded as is with case with vsync disabled.

Upcoming Feature Work

[Video: 1:30-1:51]

• Vir re-iterated that 2022 should see the viewer progress with new graphics features.
• Further performance improvements beyond those currently within the Performance Improvements viewer are in the planning stages.
• However, nothing is available for open discussion by LL at this point in time.

Animation Override Discussion

[Video: 11:07-52:00]

• A discussion on improving avatar animations through the use of a cap / reliable messaging between the viewer and the server to directly replace the default server-side avatar animations (walk, run, sit, swim, fly, etc), with custom animations  – as per a scripted Animation Override – but avoiding the need  to use llSetAnimationOverride and scripted HUDs (as is currently the case).
• See also  BUG-230100.
• No work is currently planned for this, but interest was expressed in how it might work.
• Firestorm has such a capability (adopted by some other TPVs), but the implementation isn’t widely favoured.
• Another suggested option would be to make animations assets that define an animation set (e.g. a “Stand” asset that can be a container for a set of animation stands + the timing for running them + defining if they should be randomly or sequentially played).
• A further advantage is that as well as removing the need for scripted attachments, it would allow everything to be drive via a viewer UI element, offloading work from the server to the Viewer (and potentially into an off-thread).
• A problem in using a cap is that the viewer could hit it fairly readily, particularly if cycling through a large set of animated stands, for example.
• There are also edge cases were scripts may still be required, but these are not inimical to the development of a client-side AO system able to work directly with the server-side animation graph.
• The discussion lays out the benefits for more of a client-side AO control capability, and I refer you to it for the in-text comment – not that towards the end of the discussion, things turned towards some WIBNIs (as in, “Wouldn’t It Be Nice If the system could dynamically adjust walk speed to avatar size”), although elements like this would require a much more intensive overhaul of the animation system.
• Please refer to the video for the full text of the discussion.

In Brief

• [52:32-End] A short discussion on the benefits of LL defining an Area Search.
  • While Firestorm has an Area Search, it tends tow spam an entire region with requests for each individual prim hover properties (and is subject to draw distance and interest list).
  • ObjectNavMeshProperties, however (with changes and a throttle), could provide an alternative and potentially more preferable solution.
  • The core of this discussion is in text, with little input from LL – please refer to the video.
• It has been reported via the forums that there is Mac OS Monterey performance issue associated with the viewer. At the time of writing, it is not clear how widespread this is or if a bug report has been raised, although the issue appears to be with a Monterrey OS capability, rather than the viewer – please read the forum post for more.