Skip to content

OLED display type sends unsupported "-y o" flag to spotread for i1d3 meters #7

Closed

Description

@SirColeslaw

What happened

  1. Set Display Type to "OLED" for my AORUS FO32U2P (QD-OLED) in the calibration setup, with a matching .ccss meter correction selected
  2. Started an HDR calibration (target: HDR Desktop PQ (sRGB gamut))
  3. Expected the measurement to begin; instead it failed immediately: "spotread exited (code 1) before producing a measurement", Diagnostic: "Failed to locate display type matching 'o'"

Gloam-Diagnostics-20260915-100654.zip

Gloam version

1.9.1.0

Windows version

Windows 11 25H2, build 26200

GPU and driver version

NVIDIA GeForce RTX 5090, driver 32.0.16.1088 (nvlddmkm.sys 2026-07-28)

Display model

Gigabyte AORUS FO32U2P (QD-OLED, 32" 4K, HDR peak ~1080 nits per EDID), Display Port 2.1b

Was Windows HDR on?

HDR on

How did you install Gloam?

Setup.exe (recommended installer)

Diagnostics bundle

  • I have attached a diagnostics bundle to this issue

Calibration details

Calibrite ColorChecker Display Plus (identifies as i1Display3, HW Rev B-02, FW v2.28), target: HDR Desktop PQ (sRGB gamut), preset: Adaptive, meter correction: "AORUS FO32U2P - i1 Pro 2.ccss" (i1 Pro 2 spectral sample, downloaded via the in-app DisplayCAL database browser)

Anything else

Possible root cause: "OLED" display type maps internally to spotread flag -y o. Per spotread's own usage output for this instrument, the valid -y values for i1D3 are only:

-y n|l i1D3: Non-Refresh display [Default,CB1]
r|c i1D3: Refresh display [CB2]
1 i1D3: LCD CCFL PVA (Resolve JVC CPF)
l|c Other: l = LCD, c = CRT

There is no 'o' option for this instrument family. Since OLED is an emissive/refresh-type panel, -y r (Refresh) looks like the intended target rather than -y o.

Reproduced identically on two separate attempts (colorimeter.log, 09:59:06 and 10:05:41), same invocation both times:
spotread.exe -v -N -c 1 -e -y o -X "...AORUS FO32U2P - i1 Pro 2.ccss"

Secondary, likely unrelated: before each attempt, app.log shows ColorProfileRemoveDisplayAssociation failing with HRESULT 0x800707DF (ERROR_PROFILE_NOT_ASSOCIATED_WITH_DEVICE) while disabling a pre-existing foreign MHC2 profile ("HDR-kalibriertes Profil 06.09.2026 1080 Nits.icc"), created by Microsoft's own "Windows HDR Calibration" app (Microsoft Store) rather than by Gloam or DisplayCAL. Gloam's own log notes this is expected on Windows 11 25H2 and falls back successfully, so it doesn't appear to block calibration — flagging it in case it's related.

Activity

  1. davidtorcivia commented on Sep 15, 2026

    @davidtorcivia
    Collaborator

    Good catch, thanks. Confirmed.

    The o/e/b letters only exist when X-Rite's EDR files have been imported into Argyll. My machine has them, a fresh install doesn't, so this broke OLED, LCD LED and wide gamut on every clean install, not just OLED.

    One correction: Argyll treats OLED as non-refresh, so the base for an i1D3 is n, not r. With a .ccss selected it doesn't matter anyway, spotread takes both the spectral cal and the refresh mode from the file.

    Fix is in #8 and will ship in 1.9.2. It now picks -y from the table spotread prints for the connected meter.

    The 0x800707DF error is a known 25H2 quirk and harmless.

  2. davidtorcivia commented on Sep 15, 2026

    @davidtorcivia
    Collaborator

    1.9.2 is out with the fix. Installed copies will pick it up on next launch.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

bugSomething isn't working

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions