Can't get past speaker identification (error 3221225794)

[mnemonictranslation]

Hi there. I saw that someone has had the same error code before, but I understand the solution had been already implemented so that the software adapted by itself.

This is what my log looks like:

=== Starting queue ===
Processing 1 transcription job(s)

Starting job: 2026-05-28 18-37-12.mp3

Converting audio...
Audio conversion finished

Speaker identification...
Loading pyannote
Diarization worker exited unexpectedly (code 3221225794). UI remains responsive.
Error in step 2 - identifying the speakers.: Subprocess terminated unexpectedly

=== Queue processing complete ===
Total jobs: 1
Completed: 0
Failed: 1
Canceled: 0
Total processing time: 0:06

I'm on windows 10. My processor is an AMD Ryzen 5 5600G and my GPU is a Radeon RX 460 Series.

Setting speaker identification to "none" shows this error instead: "Transcription worker exited unexpectedly (code 3221225794).
Subprocess terminated unexpectedly"

Thank you.

[kaixxx]

Did you install the CUDA version? That's not compatible with you AMD Radeon GPU and might explain the issue.

If this is not the issue, please check the latest log in %LOCALAPPDATA%\noScribe\noScribe\log\ and post the results here, as they might give a little more insight in what's happening.

[mnemonictranslation]

Thank you for your response.

Yes, I installed the general purpose version.

Here are the results of the last log:

whisper beam size: 1
whisper temperature: 0.0
whisper compute type: default
timestamp_interval: 60000
timestamp_color: #78909C
=== CPU FEATURES ===
System: Windows
VendorId : AuthenticAMD
num_virtual_cores : 12
num_physical_cores : 6
num_threads_per_core : 2
num_cpus : 1
cache_line_size : 64
cache_L1_size : 0
cache_L2_size : 0
cache_L3_size : 0
OS_x64 : True
OS_AVX : True
OS_AVX512 : False
MMX : True
x64 : True
ABM : True
RDRAND : True
BMI1 : True
BMI2 : True
ADX : True
PREFETCHWT1 : False
MPX : False
SSE : True
SSE2 : True
SSE3 : True
SSSE3 : True
SSE4.1 : True
SSE4.2 : True
SSE4.a : True
AES : True
SHA : True
AVX : True
XOP : False
FMA3 : True
FMA4 : False
AVX2 : True
AVX512f : False
AVX512pf : False
AVX512er : False
AVX512cd : False
AVX512vl : False
AVX512bw : False
AVX512dq : False
AVX512ifma : False
AVX512vbmi : False
AVX512vbmi2 : False
AVX512vnni : False

Converting audio...
Audio conversion finished

Transcription...
Loading whisper
ERROR: Transcription worker exited unexpectedly (code 3221225794).

[kaixxx]

Thank you. Unfortunately, we don't get much more information from that.

Both the speaker identification process as well as the transcription process seem to fail quite early. The return code 3221225794 points to a DLL loading issue during initialization. I have retested the newest noScribe version on my old Win10 machine, and it works fine.

Maybe an incomplete download and installation could be the culprit? Please go to C:\Program Files (x86), right click on noScribeand select properties. Now check the size of the folder, it should be around 3.5 GB.

Also, you can start noScribe.exe from the command line (a Windows terminal) and look if any additional error messages pop up there during transcription. This whole issue is a little bit of a mystery...