When I select Z-Depth or Position under auxilliary render data, transparency maps are ignored for those layers: the image should show exactly what I mean.

This really limits the usefulness of these potentially very useful features.

got that too, bb said to use atmopshere object depth cue instead (turn all lights and ambient mats off)
http://www.renderosity.com/mod/forumpro/showthread.php?thread_id=2797605
would like to see the auxiliary fixed though so it's all in a single render :)

April,

I tested the scene using that technique and it worked as it should. Just got a reply from SM support saying it's been reported to the project team so hopefully it'll be fixed in SR1.

(unless I'm mistaking what you mean of course)

Here's what I mean.
This is the actual render - back 2 cubes are at the same distance from the camera

"They only give depth data, something that is transparent but the same "depth" as something opaque should have the same greyscale value ." 

That would be pretty much my understanding as well. Carrara also appears to ignore transparency and I see some posts on the web suggesting the same for VRay and MentalRay.

it is much more useful to me personally to use transparency but it does make sense to keep consistent with "industry standards" if any exist.  how about an option for the aux psd z-depth to use transparency or not? lol

note the thread I linked, my original problem with it was it was not anti-aliased, so it was jaggy even on fully opaque objects. I hope that's fixed (too) ...

In this example, quickly cobbled together in Vue, there's a figure with transmapped hair (aargh! sorry April, should have used one of yours!), and two cubes in the background, one solid, one glass.

Everything in the Z-depth render is as it should be: the alpha mask for the hair has been recognised; the cubes have the correct tones to record their positions in the Z-plane.

Using the Z-depth render to create depth of focus effects - or any of the Position channels for other effects - is highly compromised if alpha masks are ignored.

An alpha mask isn't the same as transparency though for a render engine - Vue has separate sections for alpha and transparency.

Quote - Using the Z-depth render to create depth of focus effects - or any of the Position channels for other effects - is highly compromised if alpha masks are ignored.

I'm not sure that's the case (at least when I've used it)  When I pick the grey level to use as the focal point in the lens blur filter, the result is pretty much what I'd expect it to be

I think your bug is more to do with alpha masks not being rendered correctly in ZDepth passes rather than transparency per-se.
In which case yeah, SM have to fix it