o i@sdddlmZmZddlZddlmZmZmZ   ddej dej dee d ee d ee d ej f d d Z  ddej dej dee d ee d ej f ddZ  ddej dej dee d ee d ej f ddZ    ddej dej dee d ee dee dee d eej dffddZ    ddej dej dee d ee dee dee d ej fddZdS) )OptionalTupleN)sdr_losssdrbss_eval_sourcesFestref zero_meanclamp_dbpairwisereturncCst||d|||dS)Nr) filter_lengthr r r )r)rr r r r rW/home/ubuntu/.local/lib/python3.10/site-packages/fast_bss_eval/numpy/scale_invariant.py si_sdr_losssrc Cst||d||dddS)NrFT)rr r return_perm change_signrrr r r rrrsi_sdr_pit_losssrcCst||||ddS)NT)r r r )rrrrrpairwise_si_sdr_loss)srrr.c Ct||dd||||dS)a Compute the scale-invariant signal-to-distortion ratio (SI-SDR) only. This function computes the SDR for all pairs of est/ref signals and finds the permutation maximizing the SDR. The order of ref/est follows the convention of bss_eval (i.e., ref first). Parameters ---------- ref: numpy.ndarray (..., n_channels_est, n_samples) The estimated signals est: numpy.ndarray (..., n_channels_ref, n_samples) The groundtruth reference signals filter_length: int, optional The length of the distortion filter allowed (default: ``512``) use_cg_iter: int, optional If provided, an iterative method is used to solve for the distortion filter coefficients instead of direct Gaussian elimination. This can speed up the computation of the metrics in case the filters are long. Using a value of 10 here has been shown to provide good accuracy in most cases and is sufficient when using this loss to train neural separation networks. zero_mean: bool, optional When set to True, the mean of all signals is subtracted prior to computation of the metrics (default: ``False``) clamp_db: bool, optional If provided, the resulting metrics are clamped to be in the range ``[-clamp_db, clamp_db]`` load_diag: float, optional If provided, this small value is added to the diagonal coefficients of the system metrics when solving for the filter coefficients. This can help stabilize the metric in the case where some of the reference signals may sometimes be zero return_perm: bool, optional If set to True, the optimal permutation of the estimated signals is also returned (default: ``False``) change_sign: bool, optional If set to True, the sign is flipped and the negative SDR is returned (default: ``False``) Returns ------- cisdr: numpy.ndarray, (..., n_channels_est) The SDR of the input signal perm: numpy.ndarray, (..., n_channels_est), optional The index of the corresponding reference signal. Only returned if ``return_perm == True`` rN)r use_cg_iterr r rrr)r rr r rrrrrsi_sdr2s8rTcompute_permutation load_diagc Cr)a This function computes the SI-SDR, SI-SIR, and SI-SAR for the input reference and estimated signals. The order of ref/est follows the convention of bss_eval (i.e., ref first). Parameters ---------- ref: numpy.ndarray (..., n_channels_est, n_samples) The estimated signals est: numpy.ndarray (..., n_channels_ref, n_samples) The groundtruth reference signals zero_mean: bool, optional When set to True, the mean of all signals is subtracted prior to computation of the metrics (default: ``False``) clamp_db: bool, optional If provided, the resulting metrics are clamped to be in the range ``[-clamp_db, clamp_db]`` compute_permutation: bool, optional When this flag is true, the optimal permutation of the estimated signals to maximize the SIR is computed (default: ``True``) load_diag: float, optional If provided, this small value is added to the diagonal coefficients of the system metrics when solving for the filter coefficients. This can help stabilize the metric in the case where some of the reference signals may sometimes be zero Returns ------- si-sdr: numpy.ndarray, (..., n_channels_est) The scale-invariant signal-to-distortion-ratio (SDR) si-sir: numpy.ndarray, (..., n_channels_est) The scale-invariant signal-to-interference-ratio (SIR) si-sar: numpy.ndarray, (..., n_channels_est) The scale-invariant signal-to-interference-ratio (SIR) perm: numpy.ndarray, (..., n_channels_est) The index of the corresponding reference signal (only when ``compute_permutation == True``) rN)rrr r rr)r)r rr r rrrrrsi_bss_eval_sourcesvs/r)FNF)FN)FNFF)FNTN)typingrrnumpynpmetricsrrrndarrayboolfloatrrrrrrrrrs      G