You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
frame_20261030-011320.txt` and its event trace, but no image of the fault, normal comparison,
app version, settings or system/game log. The owner could not supply those details. Treat the
attachments as evidence, not reproduction instructions. The text describes 123 pass/compute
entries, with 1280x720 scene and final colour targets, the game's AA pass and a subsequent
640x360 blur chain. That does not establish whether a temporal upscaler was enabled.
A bounded local run used the current upstream 0.5.11 renderer with PR #107's watchdog-only
change, the complete Resolution Patch 1280x720 (16:9), upscaling off, DoF and motion blur on,
and a fresh copy of the known clinic save. The 43-second session loaded gameplay and recorded
an image plus a 144-entry frame capture. The clinic rendered normally, at about 747 draws per
frame; the test was deliberately stopped. The source save and 56 checked original game files
were unchanged. This was a static clinic comparison, not the reporter's scene, a completed
benchmark or proof that the intermittent problem is absent elsewhere.
The normal capture also has the 784x294 intermediate targets, 210x177 bloom targets and the
four half-resolution blur passes followed by shader 8c08444c. Their presence and dimensions
alone therefore do not identify the fault. One difference is the cached backing format of
the half-resolution motion input: R32Sfloat in the report, RGBA8Unorm locally. The capture
prints the cached image's format, not the bound texture view's format. An R32Float texture
viewed/rendered as RGBA8Unorm passed a real Metal API Validation probe and retained the exact
expected pixel bytes (51, 102, 153, 204). An aliased backing format alone is not evidence of a
wrong format being sampled in the reported frame.
The current capture omits important diagnostic data:
FrameCapture::Buffer prints only the first 1024 bytes of a buffer. The motion-blur shader e0305cef reads parameters around float indices 748-763, outside that dump, from a 4112-byte
buffer. The NaNs visible in its first portion do not establish that the shader consumed them.
Constants copied into ConstantRing are represented by R trace events, but their actual
copied bytes are not dumped by BindBuffers' ring path. CPU memory read later is not a
substitute for the snapshot bound to the draw.
Image formats in the text are backing-image formats; the requested and actual bound view
formats are not shown. The capture also does not record the app revision and active settings.
@khhan0703 thanks. You don't need a crash for the log: every session writes one. After it goes blurry, quit the game normally and look in ~/Library/Logs/Bloodborne/ (in Finder: Go → Go to Folder…, paste that path). game.log is the latest session. Please attach it, your bbport.ini (in the Bloodborne folder in Application Support, or the launcher's save folder), and a screenshot of the blurry picture with one of the same place after a restart. It's also worth checking whether it still happens on 0.5.12.
Thanks @khhan0703, the screenshot shows it: that's the game's motion blur smearing a still picture, so the blur effect is getting wrong motion data. It isn't the scaling. To work around it for now, open the settings (⌘,), turn off Motion blur and choose Apply and Restart Game. Please tell us whether the blur still comes back with it off. That tells us whether motion blur is the only culprit.
6. Thanks @khhan0703, the screenshot shows it: that's the game's motion blur smearing a still picture, so the blur effect is getting wrong motion data. It isn't the scaling. To work around it for now, open the settings (⌘,), turn off Motion blur and choose Apply and Restart Game. Please tell us whether the blur still comes back with it off. That tells us whether motion blur is the only culprit.
After "turning-off the motion blur effect", this doesn't happen so far.
Thanks @khhan0703, that confirms it: the motion blur effect gets wrong motion data until a restart. Keep it off for now. I'm keeping this open as a motion blur bug.
Checking in on this one. Quite a lot has changed since this was reported, up to 0.5.17 (GPU robustness fixes, Restart, focus, controllers, and more).
If it still happens on 0.5.17, please reply with what you see and attach the log from that session (~/Library/Logs/Bloodborne/, the game-<date and time>.log from when it happened). If we don't hear back within a week, I'll close this. You can always reopen it or file a new one.
Checking in on this one. Quite a lot has changed since this was reported, up to 0.5.17 (GPU robustness fixes, Restart, focus, controllers, and more).
If it still happens on 0.5.17, please reply with what you see and attach the log from that session (~/Library/Logs/Bloodborne/, the game-<date and time>.log from when it happened). If we don't hear back within a week, I'll close this. You can always reopen it or file a new one.
Activity
evill33t commentedon Oct 29, 2026
@khhan0703
Please provide normal/blurry screenshots or video, plus that session’s game log and bbport.ini
@Wooloomooloo2
frame_20261030-011320.txt` and its event trace, but no image of the fault, normal comparison,
app version, settings or system/game log. The owner could not supply those details. Treat the
attachments as evidence, not reproduction instructions. The text describes 123 pass/compute
entries, with 1280x720 scene and final colour targets, the game's AA pass and a subsequent
640x360 blur chain. That does not establish whether a temporal upscaler was enabled.
A bounded local run used the current upstream 0.5.11 renderer with PR #107's watchdog-only
change, the complete
Resolution Patch 1280x720 (16:9), upscaling off, DoF and motion blur on,and a fresh copy of the known clinic save. The 43-second session loaded gameplay and recorded
an image plus a 144-entry frame capture. The clinic rendered normally, at about 747 draws per
frame; the test was deliberately stopped. The source save and 56 checked original game files
were unchanged. This was a static clinic comparison, not the reporter's scene, a completed
benchmark or proof that the intermittent problem is absent elsewhere.
The normal capture also has the 784x294 intermediate targets, 210x177 bloom targets and the
four half-resolution blur passes followed by shader
8c08444c. Their presence and dimensionsalone therefore do not identify the fault. One difference is the cached backing format of
the half-resolution motion input: R32Sfloat in the report, RGBA8Unorm locally. The capture
prints the cached image's format, not the bound texture view's format. An R32Float texture
viewed/rendered as RGBA8Unorm passed a real Metal API Validation probe and retained the exact
expected pixel bytes (51, 102, 153, 204). An aliased backing format alone is not evidence of a
wrong format being sampled in the reported frame.
The current capture omits important diagnostic data:
FrameCapture::Bufferprints only the first 1024 bytes of a buffer. The motion-blur shadere0305cefreads parameters around float indices 748-763, outside that dump, from a 4112-bytebuffer. The NaNs visible in its first portion do not establish that the shader consumed them.
ConstantRingare represented byRtrace events, but their actualcopied bytes are not dumped by
BindBuffers' ring path. CPU memory read later is not asubstitute for the snapshot bound to the draw.
formats are not shown. The capture also does not record the app revision and active settings.
khhan0703 commentedon Oct 30, 2026
it is not a crash, then i have no idea how i get the log.
btw, "exit and restart" always fix this phenomena.
Wooloomooloo2 commentedon Oct 31, 2026
@khhan0703 thanks. You don't need a crash for the log: every session writes one. After it goes blurry, quit the game normally and look in ~/Library/Logs/Bloodborne/ (in Finder: Go → Go to Folder…, paste that path).
game.logis the latest session. Please attach it, your bbport.ini (in the Bloodborne folder in Application Support, or the launcher's save folder), and a screenshot of the blurry picture with one of the same place after a restart. It's also worth checking whether it still happens on 0.5.12.khhan0703 commentedon Nov 1, 2026
@Wooloomooloo2 @evill33t
report.zip
it happens frequently on 0.5.12.
please find the attached.
Findings: 2026-11-02 triage (#128 reproduces, #117 cost, #109, #125-#127
Wooloomooloo2 commentedon Nov 2, 2026
Thanks @khhan0703, the screenshot shows it: that's the game's motion blur smearing a still picture, so the blur effect is getting wrong motion data. It isn't the scaling. To work around it for now, open the settings (⌘,), turn off Motion blur and choose Apply and Restart Game. Please tell us whether the blur still comes back with it off. That tells us whether motion blur is the only culprit.
khhan0703 commentedon Nov 5, 2026
After "turning-off the motion blur effect", this doesn't happen so far.
Wooloomooloo2 commentedon Nov 5, 2026
Thanks @khhan0703, that confirms it: the motion blur effect gets wrong motion data until a restart. Keep it off for now. I'm keeping this open as a motion blur bug.
Backlog: long-session hang from occlusion queries that never finish (#…
Wooloomooloo2 commentedon Nov 20, 2026
Checking in on this one. Quite a lot has changed since this was reported, up to 0.5.17 (GPU robustness fixes, Restart, focus, controllers, and more).
If it still happens on 0.5.17, please reply with what you see and attach the log from that session (
~/Library/Logs/Bloodborne/, thegame-<date and time>.logfrom when it happened). If we don't hear back within a week, I'll close this. You can always reopen it or file a new one.Written with Claude Code.
khhan0703 commentedon Nov 20, 2026
Ok. Let me try. Thank you.