o ziC@szddlZddlmZddlZddlmZddlmZddlmZGdddej Z Gd d d ej Z Gd d d ej Z dS) N)List)nn) AudioSignal STFTParamscszeZdZdZddgedddddddf d eed ej d e d e d e de de de de ffdd Z dedefddZZS)MultiScaleSTFTLossa%Computes the multi-scale STFT loss from [1]. Parameters ---------- window_lengths : List[int], optional Length of each window of each STFT, by default [2048, 512] loss_fn : typing.Callable, optional How to compare each loss, by default nn.L1Loss() clamp_eps : float, optional Clamp on the log magnitude, below, by default 1e-5 mag_weight : float, optional Weight of raw magnitude portion of loss, by default 1.0 log_weight : float, optional Weight of log magnitude portion of loss, by default 1.0 pow : float, optional Power to raise magnitude to before taking log, by default 2.0 weight : float, optional Weight of this loss, by default 1.0 match_stride : bool, optional Whether to match the stride of convolutional layers, by default False References ---------- 1. Engel, Jesse, Chenjie Gu, and Adam Roberts. "DDSP: Differentiable Digital Signal Processing." International Conference on Learning Representations. 2019. h㈵>?@FNwindow_lengthsloss_fn clamp_eps mag_weight log_weightpowweight match_stride window_typec sHtfdd|D|_||_||_||_||_||_||_dS)Nc g|] }t||ddqS) window_length hop_lengthrrr.0wrrO/home/ubuntu/.local/lib/python3.10/site-packages/audiotools/metrics/spectral.py 6z/MultiScaleSTFTLoss.__init__..) super__init__ stft_paramsrrrrrr) selfrrrrrrrrr __class__rr!r%)s   zMultiScaleSTFTLoss.__init__xyc Csd}|jD]D}||j|j|j||j|j|j||j||j|j  |j  |j|j  |j  7}||j ||j|j7}q|S)a?Computes multi-scale STFT between an estimate and a reference signal. Parameters ---------- x : AudioSignal Estimate signal y : AudioSignal Reference signal Returns ------- torch.Tensor Multi-scale STFT loss. ) r&stftrrrrr magnitudeclamprrlog10r)r'r*r+losssr r r!forwardFs  zMultiScaleSTFTLoss.forward__name__ __module__ __qualname____doc__rL1LossrinttypingCallablefloatboolstrr%rr3 __classcell__r r r(r!r s>  rcseZdZdZddgddgedddddd d d gd d gd f d eed eedej de de de de de de dee dee de ffdd Z dedefddZZS)MelSpectrogramLossaCompute distance between mel spectrograms. Can be used in a multi-scale way. Parameters ---------- n_mels : List[int] Number of mels per STFT, by default [150, 80], window_lengths : List[int], optional Length of each window of each STFT, by default [2048, 512] loss_fn : typing.Callable, optional How to compare each loss, by default nn.L1Loss() clamp_eps : float, optional Clamp on the log magnitude, below, by default 1e-5 mag_weight : float, optional Weight of raw magnitude portion of loss, by default 1.0 log_weight : float, optional Weight of log magnitude portion of loss, by default 1.0 pow : float, optional Power to raise magnitude to before taking log, by default 2.0 weight : float, optional Weight of this loss, by default 1.0 match_stride : bool, optional Whether to match the stride of convolutional layers, by default False Pr r r r r Fr,Nn_melsrrrrrrrrmel_fminmel_fmaxrc sZtfdd|D|_||_||_||_||_||_||_| |_ | |_ ||_ dS)Ncrrrrrr r!r"r#z/MelSpectrogramLoss.__init__..) r$r%r&rDrrrrrrErFr) r'rDrrrrrrrrrErFrr(rr!r%|s    zMelSpectrogramLoss.__init__r*r+c Csd}t|j|j|j|jD]Q\}}}}|j|j|jd}|j|f||d|} |j|f||d|} ||j | | |j  |j | |j  |j 7}||j| | | 7}q |S)a*Computes mel loss between an estimate and a reference signal. Parameters ---------- x : AudioSignal Estimate signal y : AudioSignal Reference signal Returns ------- torch.Tensor Mel loss. r,)rrr)rErF)ziprDrErFr&rrrmel_spectrogramrrr/rrr0r) r'r*r+r1rDfminfmaxr2kwargsx_melsy_melsr r r!r3s  zMelSpectrogramLoss.forwardr4r r r(r!rAbsP     #rAcsDeZdZdZ ddededeffdd Zd ed efd d ZZ S) PhaseLossa0Difference between phase spectrograms. Parameters ---------- window_length : int, optional Length of STFT window, by default 2048 hop_length : int, optional Hop length of STFT window, by default 512 weight : float, optional Weight of loss, by default 1.0 r r r rrrcs t||_t|||_dS)N)r$r%rrr&)r'rrrr(r r!r%s zPhaseLoss.__init__r*r+c Cs|j}||j|j|j||j|j|j|j|j}||tj kdtj7<||tjkdtj8<|j |j }}|j|||}||d }|S)a.Computes phase loss between an estimate and a reference signal. Parameters ---------- x : AudioSignal Estimate signal y : AudioSignal Reference signal Returns ------- torch.Tensor Phase loss. r) r&r-rrrphasenppir.minmaxmean) r'r*r+r2diffx_minx_maxweightsr1r r r!r3s zPhaseLoss.forward)r r r ) r5r6r7r8r:r=r%rr3r@r r r(r!rNs rN) r;rnumpyrQtorchrrrModulerrArNr r r r!s    Wa