How do I set up EXR output in Blender?

This lesson assumes you’ve read What the hell is an EXR file in Blender?. Or at least know the gist that EXR keeps your raw render data and PNG bakes it into pixels. If not, start there.

Now here’s how to actually set EXR up in Blender.

The three settings that matter

Open Output Properties and the choices come down to three rows.

File Format → OpenEXR. This is the format that keeps your render data intact: scene-referred values, including any pixels that came in brighter than white. PNG and JPG cannot hold that. (OpenEXR MultiLayer is the multi-pass sibling, covered in a future lesson.)

Color Depth → Float (Half). Half float gives you about four orders of magnitude of brightness range. Way more than any monitor can show and plenty for grading. Full float exists for the rare case you need more precision, at roughly double the file size.

Codec → ZIP or PIZ. Both are lossless. ZIP is faster to write, PIZ compresses smaller on grainy renders. Either is a safe default. Avoid None unless you want files that are needlessly huge for no reason.

When you probably do not need EXR

EXR is not mandatory for every render. You can usually skip it when:

• the render is final and you’re exporting it,
• the image is only for review or approval,
• it’s a quick preview or playblast,
• or the shot is simple and you know you won’t comp it.

The point is not “always use EXR forever.” The point is: if the render is still a working asset, EXR is usually the safer master file.

Treat the EXR as a working master

Render to EXR. Bake PNG or JPG only when you need a deliverable. Keep the view transform (Color Management → View Transform → AgX) as a monitoring choice on top of the EXR, not a permanent change to the file.

That distinction is what makes EXR worth the disk space. If you bake a display look into the saved file, you have effectively committed to a grade. The first time someone asks “can we lift the shadows?” two weeks later, you have to find the original .blend and re-render.

Common mistakes that create avoidable pain

1. Treating 16-bit PNG as if it were equivalent to EXR because both say 16-bit.
2. Saving only a PNG sequence for a shot that still needs compositing.
3. Using full float on everything when half float would have been enough.
4. Baking a display look too early, then trying to make big grading changes later.

Bottom line

The renderer already did the expensive work. EXR is how you avoid throwing the useful part of that work away.

The next step up from this is keeping every render pass in one file (diffuse, glossy, depth, normals) so you can grade and recomp them without re-rendering. That is the MultiLayer EXR. Its own lesson, coming soon.