docs(cnn_v3): blender4 alias + Blender 4.5 LTS requirement for training data

1 files changed, 15 insertions, 7 deletions

diff --git a/cnn_v3/docs/HOWTO.md b/cnn_v3/docs/HOWTO.md
index 0cf2fe5..983e8b7 100644
--- a/cnn_v3/docs/HOWTO.md
+++ b/cnn_v3/docs/HOWTO.md
@@ -82,15 +82,23 @@ CNN v3 supports two data sources: Blender renders and real photos.
 
 ### 2a. From Blender Renders
 
+Requires **Blender 4.5 LTS** — Blender 5.x compositor does not route per-pass
+render data yet (only Combined is exported).
+
 ```bash
-# 1. In Blender: run the export script (requires Blender 3.x+)
-blender --background scene.blend --python cnn_v3/training/blender_export.py \
-    -- --output /tmp/renders/ --frames 200
+# macOS: add to ~/.zshrc or run once per shell session
+alias blender4="/Applications/Blender_4.5.8_LTS.app/Contents/MacOS/Blender"
+
+# 1. Export G-buffer passes to multilayer EXR
+blender4 -b scene.blend -P cnn_v3/training/blender_export.py \
+    -- --output /tmp/renders/
 
-# 2. Pack into sample directory
-python3 cnn_v3/training/pack_blender_sample.py \
-    --render-dir /tmp/renders/frame_0001/ \
-    --output dataset/blender/sample_0001/
+# 2. Pack each EXR into a sample directory
+for exr in /tmp/renders/*.exr; do
+    name=$(basename "${exr%.exr}")
+    python3 cnn_v3/training/pack_blender_sample.py \
+        --exr "$exr" --output /tmp/renders/$name/
+done
 ```
 
 Each sample directory contains: