MagicBeanDev/Hookshot

Archived

Watch

Fork 0

foliage is still disapearing, some flower and branches glow #667

Closed

opened 2026-11-08 16:23:47 +00:00 by tyanybloomu-cmd · 6 comments

tyanybloomu-cmd commented

2026-11-08 16:23:47 +00:00

(Migrated from github.com)

Builder version

2.0.1

How you built

Hookshot Builder app (Windows)

Game files given to it

A .nup

The error

foliage is still disapearing, some flower and branches glow

Visual Studio or Build Tools version, and Windows SDK version

No response

Operating system and free space on the output drive

windows 11 180 free gb

Before you send it

I'm on the newest release, or it still fails on it.
I searched the open issues for this error.
I'm not attaching game files or anything built from them.

### Builder version 2.0.1 ### How you built Hookshot Builder app (Windows) ### Game files given to it A .nup ### The error ```text foliage is still disapearing, some flower and branches glow ``` ### Visual Studio or Build Tools version, and Windows SDK version _No response_ ### Operating system and free space on the output drive windows 11 180 free gb ### Before you send it - [x] I'm on the newest release, or it still fails on it. - [x] I searched the open issues for this error. - [x] I'm not attaching game files or anything built from them.

👍 2

hellknight0000 commented

2026-11-08 16:31:39 +00:00

(Migrated from github.com)

hookshot_log.1.txt
hookshot_log.2.txt
hookshot_log.3.txt
hookshot_log.txt

https://youtu.be/0GzDiqAwMl0

Same issue

<img width="1539" height="967" alt="Image" src="https://github.com/user-attachments/assets/e35554af-28b8-443f-8f26-375fbb281bf1" /> <img width="1539" height="967" alt="Image" src="https://github.com/user-attachments/assets/36f8e2a4-200e-4cf9-af4c-3eefc83cfff6" /> <img width="1539" height="967" alt="Image" src="https://github.com/user-attachments/assets/06ced52c-ff65-45ea-808a-1f724b382528" /> [hookshot_log.1.txt](https://github.com/user-attachments/files/34547142/hookshot_log.1.txt) [hookshot_log.2.txt](https://github.com/user-attachments/files/34547143/hookshot_log.2.txt) [hookshot_log.3.txt](https://github.com/user-attachments/files/34547145/hookshot_log.3.txt) [hookshot_log.txt](https://github.com/user-attachments/files/34547144/hookshot_log.txt) https://youtu.be/0GzDiqAwMl0 Same issue

namphamsociu commented

2026-11-08 17:15:20 +00:00

(Migrated from github.com)

Same issue on v2.0.2

<img width="1501" height="870" alt="Image" src="https://github.com/user-attachments/assets/a17dd080-93c3-4f60-8f59-2b9fd1b4d709" /> <img width="1503" height="870" alt="Image" src="https://github.com/user-attachments/assets/109e0660-bc46-418f-bef7-eff9b0ac2107" /> Same issue on v2.0.2

BernabeCR93 commented

2026-11-08 17:23:15 +00:00

(Migrated from github.com)

Same here (RTX 3060 Ti, Ryzen 5 3600, Windows, a source build of main as of today). Some findings in case they help:

A radius around Link where everything near "washes out". Inside a sphere that moves with Link, nearby undergrowth loses its stems and leaves (ferns keep only their flowers), trees lose their canopy (their shadow shows bare branches), and the shadows inside that radius go too. Walking, the edge of the sphere moves ahead of Link and "washes" what it reaches. Outside it, everything draws. Since the shadows go with it, it looks like something tied to the near region (nearest shadow cascade, the player-centred grass interaction map, or a player-centred target reused between frames) rather than the foliage meshes themselves.

Ferns and far trees drawn like mirrors (white facing the sun, sky blue when looking up; the colour comes back when the camera moves). Logging every storage buffer a draw could not bind showed the far-foliage vertex programs 95db514a35353e28 and d76e73cfbe651494 (and 68642bb810ee3e7c, 6b4dea601798c255, 739acd678ae08655, d905f52b7517c681, 51183fe750d5d0c8, 4f0c0b40ab3a32cd) binding a "storage buffer" whose address is float data: c4+0x40 addr 0x430c0000c2480000, c5+0x60 addr 0x10001c00a size 0x3faaaaab. The address was tracked without the NVN bias to a constant buffer other than c0 while part of it comes from memory, so it was never bound and read zeros. Resetting that result for graphics stages in global_memory_to_storage_buffer_pass.cpp (next to the c1 rule: index != 0 and Origins(...).load, left to the run-time search) took "storage buffer NOT bound" from ~38,000 a session to 0, and trees and bushes draw with their colours; the ferns still look reflective.

Tried, no change: RX_VK_SSBO_STRETCH=1, HOOKSHOT_OCCLUSION_OFF=1, RX_VK_FOLLOW_MESH=1 (more flashes), 8 instead of 4 global_tables (more black ground patches).

Workaround: Hyrule Field's ResFoliageScene has no LodSetting. Adding one with Dist_Undergrowth0 = 0.5 (and growing that file's entry in ResourceSizeTable.Product.111...rsizetable.zs, or the game closes while the map loads) keeps the far models up close; trees then keep their canopy.

Same here (RTX 3060 Ti, Ryzen 5 3600, Windows, a source build of `main` as of today). Some findings in case they help: **A radius around Link where everything near "washes out".** Inside a sphere that moves with Link, nearby undergrowth loses its stems and leaves (ferns keep only their flowers), trees lose their canopy (their shadow shows bare branches), and the shadows inside that radius go too. Walking, the edge of the sphere moves ahead of Link and "washes" what it reaches. Outside it, everything draws. Since the shadows go with it, it looks like something tied to the near region (nearest shadow cascade, the player-centred grass interaction map, or a player-centred target reused between frames) rather than the foliage meshes themselves. **Ferns and far trees drawn like mirrors** (white facing the sun, sky blue when looking up; the colour comes back when the camera moves). Logging every storage buffer a draw could not bind showed the far-foliage vertex programs `95db514a35353e28` and `d76e73cfbe651494` (and `68642bb810ee3e7c`, `6b4dea601798c255`, `739acd678ae08655`, `d905f52b7517c681`, `51183fe750d5d0c8`, `4f0c0b40ab3a32cd`) binding a "storage buffer" whose address is float data: `c4+0x40 addr 0x430c0000c2480000`, `c5+0x60 addr 0x10001c00a size 0x3faaaaab`. The address was tracked without the NVN bias to a constant buffer other than c0 while part of it comes from memory, so it was never bound and read zeros. Resetting that result for graphics stages in `global_memory_to_storage_buffer_pass.cpp` (next to the c1 rule: index != 0 and `Origins(...).load`, left to the run-time search) took "storage buffer NOT bound" from ~38,000 a session to 0, and trees and bushes draw with their colours; the ferns still look reflective. **Tried, no change:** `RX_VK_SSBO_STRETCH=1`, `HOOKSHOT_OCCLUSION_OFF=1`, `RX_VK_FOLLOW_MESH=1` (more flashes), 8 instead of 4 `global_tables` (more black ground patches). **Workaround:** Hyrule Field's `ResFoliageScene` has no `LodSetting`. Adding one with `Dist_Undergrowth0` = 0.5 (and growing that file's entry in `ResourceSizeTable.Product.111...rsizetable.zs`, or the game closes while the map loads) keeps the far models up close; trees then keep their canopy.

ENPFakeDeveloper commented

2026-11-08 17:46:28 +00:00

(Migrated from github.com)

Seeing this too on AMD Radeon RX 6650 XT (Vulkan 1.4), Windows 11, hookshot v2.0.2 (source 998ef8f9): vegetation is fine at a distance, but up close only the branches remain and the leaves disappear. Details and logs (storage buffer NOT bound (reads zeros) on stage 7 mesh programs) are in #696. I'll add a video here shortly.
Analysis created with Claude; it may contain errors

https://youtu.be/YXcEmNwKbms

Seeing this too on **AMD Radeon RX 6650 XT** (Vulkan 1.4), Windows 11, hookshot v2.0.2 (source 998ef8f9): vegetation is fine at a distance, but up close only the branches remain and the leaves disappear. Details and logs (`storage buffer NOT bound (reads zeros)` on stage 7 mesh programs) are in #696. I'll add a video here shortly. Analysis created with Claude; it may contain errors https://youtu.be/YXcEmNwKbms

👍 1

lucasmazzoni45-del commented

2026-11-08 23:33:33 +00:00

(Migrated from github.com)

RX 7800 XT (RDNA3), v2.0.2-v2.0.3: some flowers/stems glowed neon-green. On v2.0.2-v2.0.4 the log had storage buffer NOT bound (reads zeros) for programs 51183fe750d5d0c8 (stage 7, c5+0x1d0, a float-like address) and da7a7ab28b8e73d9 (stage 1, c5+0). Correction to what I first wrote: on v2.0.5 the lines are still there (40 per session) but the set is different: storage buffer NOT bound (reads zeros) for six mesh (stage 7) programs: 51183fe750d5d0c8 (c5+0x1d0) and 4f0c0b40ab3a32cd (c5+0x228), both with a float-like address, and 68642bb810ee3e7c, 6b4dea601798c255, 739acd678ae08655, d905f52b7517c681 (all c5+0x60, addresses like 0x10001c00a, 0x20001c00a, ... that look like packed floats), 40 lines per session. The character lighting is fine since v2.0.4. The neon glow itself is still present for me on v2.0.5 (see #875).

RX 7800 XT (RDNA3), v2.0.2-v2.0.3: some flowers/stems glowed neon-green. On v2.0.2-v2.0.4 the log had `storage buffer NOT bound (reads zeros)` for programs 51183fe750d5d0c8 (stage 7, c5+0x1d0, a float-like address) and da7a7ab28b8e73d9 (stage 1, c5+0). Correction to what I first wrote: on v2.0.5 the lines are still there (40 per session) but the set is different: `storage buffer NOT bound (reads zeros)` for six mesh (stage 7) programs: `51183fe750d5d0c8` (c5+0x1d0) and `4f0c0b40ab3a32cd` (c5+0x228), both with a float-like address, and `68642bb810ee3e7c`, `6b4dea601798c255`, `739acd678ae08655`, `d905f52b7517c681` (all c5+0x60, addresses like `0x10001c00a`, `0x20001c00a`, ... that look like packed floats), 40 lines per session. The character lighting is fine since v2.0.4. The neon glow itself is still present for me on v2.0.5 (see #875).

NotRexed commented

2026-11-09 05:53:06 +00:00

(Migrated from github.com)

Tracked in #994, follow there. That issue collects every report of floating trees and missing or vanishing vegetation, and this one is linked from it with its GPU, version and log details. It is a known issue in v2.0.6 and not fixed yet. If it still happens on v2.0.6 (https://github.com/NotRexed/Hookshot/releases/tag/v2.0.6; Update next to Play in the launcher), please add a comment there with your hookshot_log.txt.

Commenting is not possible because the repository is archived.