Audio buffer is not finite everywhere

[ludzeller]

The problem has been reported a number of times, is there a fix in sight? I am running the tool on a number of short and longer mp3 files up to 60 minutes.

Audio buffer is not finite everywhere这个怎么解决呀！ · Issue #1007

Broken Audio Files Error · Issue #621

[Error] Audio buffer is not finite everywhere · Issue #487

ParameterError: "Audio buffer is not finite everywhere" · Issue #454

ParameterError: Audio buffer is not finite everywhere · Issue #393

Last Error Received:

Process: MDX-Net

If this error persists, please contact the developers with the error details.

Raw Error Details:

ParameterError: "Audio buffer is not finite everywhere" Traceback Error: " File "UVR.py", line 6584, in process_start File "separate.py", line 487, in seperate File "separate.py", line 354, in final_process File "separate.py", line 418, in write_audio File "separate.py", line 389, in save_with_message File "separate.py", line 396, in deverb_vocals File "separate.py", line 1388, in vr_denoiser File "lib_v5/spec_utils.py", line 340, in cmb_spectrogram_to_wave File "librosa/util/decorators.py", line 104, in inner_f File "librosa/core/audio.py", line 606, in resample File "librosa/util/decorators.py", line 88, in inner_f File "librosa/util/utils.py", line 294, in valid_audio "

Error Time Stamp [2023-12-21 16:12:55]

Full Application Settings:

vr_model: UVR-DeEcho-DeReverb aggression_setting: 0 window_size: 512 mdx_segment_size: 256 batch_size: Default crop_size: 256 is_tta: False is_output_image: False is_post_process: False is_high_end_process: False post_process_threshold: 0.2 vr_voc_inst_secondary_model: No Model Selected vr_other_secondary_model: No Model Selected vr_bass_secondary_model: No Model Selected vr_drums_secondary_model: No Model Selected vr_is_secondary_model_activate: False vr_voc_inst_secondary_model_scale: 0.9 vr_other_secondary_model_scale: 0.7 vr_bass_secondary_model_scale: 0.5 vr_drums_secondary_model_scale: 0.5 demucs_model: Choose Model segment: Default overlap: 0.25 overlap_mdx: Default overlap_mdx23: 8 shifts: 2 chunks_demucs: Auto margin_demucs: 44100 is_chunk_demucs: False is_chunk_mdxnet: False is_primary_stem_only_Demucs: False is_secondary_stem_only_Demucs: False is_split_mode: True is_demucs_combine_stems: True is_mdx23_combine_stems: True demucs_voc_inst_secondary_model: No Model Selected demucs_other_secondary_model: No Model Selected demucs_bass_secondary_model: No Model Selected demucs_drums_secondary_model: No Model Selected demucs_is_secondary_model_activate: False demucs_voc_inst_secondary_model_scale: 0.9 demucs_other_secondary_model_scale: 0.7 demucs_bass_secondary_model_scale: 0.5 demucs_drums_secondary_model_scale: 0.5 demucs_pre_proc_model: No Model Selected is_demucs_pre_proc_model_activate: False is_demucs_pre_proc_model_inst_mix: False mdx_net_model: UVR-MDX-NET Inst HQ 3 chunks: Auto margin: 44100 compensate: Auto denoise_option: None is_match_frequency_pitch: True phase_option: Automatic phase_shifts: None is_save_align: False is_match_silence: True is_spec_match: False is_mdx_c_seg_def: False is_invert_spec: False is_deverb_vocals: True deverb_vocal_opt: All Vocal Types voc_split_save_opt: Lead Only is_mixer_mode: False mdx_batch_size: Default mdx_voc_inst_secondary_model: VR Arc: 1_HP-UVR mdx_other_secondary_model: No Model Selected mdx_bass_secondary_model: No Model Selected mdx_drums_secondary_model: No Model Selected mdx_is_secondary_model_activate: False mdx_voc_inst_secondary_model_scale: 0.9 mdx_other_secondary_model_scale: 0.7 mdx_bass_secondary_model_scale: 0.5 mdx_drums_secondary_model_scale: 0.5 is_save_all_outputs_ensemble: True is_append_ensemble_name: False chosen_audio_tool: Manual Ensemble choose_algorithm: Min Spec time_stretch_rate: 2.0 pitch_rate: 2.0 is_time_correction: True is_gpu_conversion: True is_primary_stem_only: False is_secondary_stem_only: True is_testing_audio: False is_auto_update_model_params: True is_add_model_name: False is_accept_any_input: False is_task_complete: False is_normalization: True is_wav_ensemble: False is_create_model_folder: False mp3_bit_set: 320k semitone_shift: 0 save_format: FLAC wav_type_set: PCM_16 cuda_set: Default help_hints_var: True set_vocal_splitter: No Model Selected is_set_vocal_splitter: False is_save_inst_set_vocal_splitter: False model_sample_mode: False model_sample_mode_duration: 30 demucs_stems: All Stems mdx_stems: All Stems

[ludzeller]

I converted the input flac to wav and turned off the vocal splitter auto deverbing and then it ran through nicely.

Still, I encountered more problems when attempting to batch deverb the resulting files, also introducing new issues such as silent or even fully glitched outputs. Has it to do with the VR Architecture somehow, were the deverbing is happening, so that that produces invalid buffers at time leading to all kinds of unwanted behaviour downstream in file compression and saving?

[ChrisTG742]

To the coders: Maybe this here could be helpful solving the issue: https://stackoverflow.com/questions/57097840/librosa-parametererror-audio-buffer-is-not-finite-everywhere-cutting-wav-files

[ludzeller]

Yes, it seems like there are NaNs in the buffer, but the question then is where do they come from and why?

[ChrisTG742]

There is one hint about the NaNs in the SO-article I linked. They say it is connected to the usage of integers instead of floats. @Anjok07 is there any type-conversion from float to int in your code?

[ChrisTG742]

Interestingly, the file "lib_v5/spec_utils.py" is completely missing in my installation here (Windows installer), though it runs without any problems or errors. @Anjok07 could it be that the code-base of the Windows-installer differs from what there's currently in your repo?

[Abedalhkeem-z]

Any updates on solving the problem?

[ANDYVDL]

This is still an issue