[RETAIL][TWW] relative/absolute location/rotation for attached objects (children)

[ByteWielder42]

Hi all.

I've come across some older threads discussing this exact issue, but I seem to be missing something, which is why I’m starting this new thread. The problem I’m encountering is that object positions change when they’re attached to a parent object (e.g., transport, specific mounts, etc.). The position and rotation (among other things) switch from absolute to relative, based on the parent object. After reversing the latest game client, I found the following:

cg_object has multiple vmethods which calculates and translates position and rotation accordingly to parent (if one exists), but to summarize:

1. Retrieve cmovement from cg_unit + 0xB0 + 0x8].
2. Check if a parent exists (cmovement + 0x0] points to parent data, and cmovement + 0x10] holds the guid of the parent, along with other checks, but I’ll keep it simple).
3. If no parent is present, use the location and rotation from cmovement + 0x20] (for position) and cmovement + 0x30] (for rotation).
4. If a parent is present (this step is recursive), retrieve the translation/world matrix for it. Verify if the matrix was constructed by checking the pointer at parent_cg_unit + 0x1678] (offsets vary by object type). If the pointer is valid, it means a valid world matrix was constructed, indicating that the object has children attached.
5. If the matrix is valid, perform the translation using the matrix, applying matrix4x4 (the parent matrix) * fvector3 (the child’s location).

I’ve replicated this logic externally, but I noticed that the positions are slightly off along the z-axis. Based on the above, I have two questions:

1. Why is my position slightly off? Since the matrix also carries rotation information, am I missing something obvious or making an error (all offsets are current for version 11.0.2.56162)?
2. (Bonus, but not necessary) Is there a way to calculate the absolute position without needing to pull the parent via guid? There seems to be promising data at cmovement + 0x0] when a parent is present.

Thanks in advance. Attaching an image for visual representation:
https://i.imgur.com/ACHtmGi.png

[Lumi666]

I'm having the same issue with my ESP when on a transport.
Camera matrix seems to change on transport, to fix it I tried applying the transport view matrix to the camera matrix. It's working but I end with the same issue as you.
Did you found a solution for your issue ?

[helloworld1024]

The new version of RETAIL seems to have changed the result of cmovement, cmovement + 0x0] does not seem to be points to parent data

Where is it now?