Changes to rc_buffer_size for hardware cbr encoding causes increased latency on client

Is there an existing issue for this?

• I have searched the existing issues

Is your issue described in the documentation?

• I have read the documentation

Is your issue present in the nightly release?

• This issue is present in the nightly release

Describe the Bug

Since c72aeef67bbcc31406525eda46e6586f8cb769d7 introduced a change to rc_buffer_size, I get increased decoding latency on a client running Windows 10 and an Intel Celeron 2957U. Prior to this commit, I was getting <1ms decoding latency, same as with regular Gamestream. After this change, latency spikes to 35+ms, with frequent freezes. Happens on bitrates exactly 100mbps or above (99.5mbps works with no issues).
Streaming at 1080p@60 h264 with NVENC.

Changing the line back to ctx->rc_buffer_size = bitrate / ((config.framerate * 10) / 15); in the latest nightly fixes the issue.

Probably related to issue #478, though this seems to be the opposite.

Expected Behavior

Same decoding latency as with Gamestream (<1ms) for high bitrates (above 100mbps).

Additional Context

Updated Windows 10 client PC, using latest Moonlight, 150mbps bitrate at 1080p@60.
Intel Celeron 2957U.

Host Operating System

Windows

Operating System Version

Windows 11

Architecture

64 bit

Sunshine commit or version

0.17.0

Package

Windows - portable

GPU Type

Nvidia

GPU Model

RTX 3080

GPU Driver/Mesa Version

528.24

Capture Method (Linux Only)

No response

Relevant log output

No response