o 2wi%@sdZddlZddlmZddlZddlmZddlm Z ddl m Z Gdd d ej j Z dd ejdededededeedefddZddZdddZdddZdS)a  Differentiable, Pytorch based resampling. Implementation of Julius O. Smith algorithm for resampling. See https://ccrma.stanford.edu/~jos/resample/ for details. This implementation is specially optimized for when new_sr / old_sr is a fraction with a small numerator and denominator when removing the gcd (e.g. new_sr = 700, old_sr = 500). Very similar to [bmcfee/resampy](https://github.com/bmcfee/resampy) except this implementation is optimized for the case mentioned before, while resampy is slower but more general. N)Optional) functional)sinc simple_reprc sbeZdZdZddedededeffdd Zd d Zddej de ede fddZ ddZ ZS) ResampleFracz? Resampling from the sample rate `old_sr` to `new_sr`. = ףp=?old_srnew_srzerosrolloffcs^tt|trt|tstdt||}|||_|||_||_ ||_ | dS)a Args: old_sr (int): sample rate of the input signal x. new_sr (int): sample rate of the output. zeros (int): number of zero crossing to keep in the sinc filter. rolloff (float): use a lowpass filter that is `rolloff * new_sr / 2`, to ensure sufficient margin due to the imperfection of the FIR filter used. Lowering this value will reduce anti-aliasing, but will reduce some of the highest frequencies. Shape: - Input: `[*, T]` - Output: `[*, T']` with `T' = int(new_sr * T / old_sr) .. caution:: After dividing `old_sr` and `new_sr` by their GCD, both should be small for this implementation to be fast. >>> import torch >>> resample = ResampleFrac(4, 5) >>> x = torch.randn(1000) >>> print(len(resample(x))) 1250 z$old_sr and new_sr should be integersN) super__init__ isinstanceint ValueErrormathgcdr r r r _init_kernels)selfr r r rr __class__L/home/ubuntu/sommelier/.venv/lib/python3.10/site-packages/julius/resample.pyrs     zResampleFrac.__init__cCs|j|jkrdSg}t|j|j}||j9}t|j|j||_t |j |j|j }t |jD];}| |j||j|}| |j |j}|tj 9}t||jdd}t||}||||q5|dt||jdddS)Nkernelr)r r minrrceilr _widthtorcharangefloatrangeclamp_picosrdiv_sumappendregister_bufferstackview)rkernelssridxitwindowrrrrrCs     "zResampleFrac._init_kernelsNFx output_lengthfullc Cs|j|jkr|S|j}|jd}|d|}tj|dddf|j|j|jfdd}tj||j|jd}| ddt |dddg}t |j||j}t |} t |} |durk|rh| n| } n|dkss|| krztd | t |} |rtd |d d| fS) a Resample x. Args: x (Tensor): signal to resample, time should be the last dimension output_length (None or int): This can be set to the desired output length (last dimension). Allowed values are between 0 and ceil(length * new_sr / old_sr). When None (default) is specified, the floored output length will be used. In order to select the largest possible size, use the `full` argument. full (bool): return the longest possible output from the input. This can be useful if you chain resampling operations, and want to give the `output_length` only for the last one, while passing `full=True` to all the other ones. rN replicate)mode)striderrrz$output_length must be between 0 and z0You cannot pass both full=True and output_length.)r r shapereshapeFpadr!conv1dr transposelistr" as_tensorr longfloorrtensor) rr5r6r7r;lengthysyfloat_output_lengthmax_output_lengthdefault_output_lengthapplied_output_lengthrrrforwardrs&   *$ zResampleFrac.forwardcCst|S)Nr)rrrr__repr__szResampleFrac.__repr__)r r )NF)__name__ __module__ __qualname____doc__rr$rrr"TensorrboolrMrN __classcell__rrrrrs  &/%rr r Fr5r r r rr6r7cCst||||||||S)aR Functional version of `ResampleFrac`, refer to its documentation for more information. ..warning:: If you call repeatidly this functions with the same sample rates, then the resampling kernel will be recomputed everytime. For best performance, you should use and cache an instance of `ResampleFrac`. )rto)r5r r r rr6r7rrr resample_fracs rWcCsdtjd|ddd}|ddd}t| d|dd|}|tj9}t||ddd}|S)NrF)periodicr?r)r" hann_windowlinspacerr'rr.)r winwinoddr3rrrr_kernel_upsample2_downsample2s  r_cCsv|j^}}t||}tj|dd|||ddddfjg||R}tj||gdd}|jg|dRS)a Upsample x by a factor of two. The output will be exactly twice as long as the input. Args: x (Tensor): signal to upsample, time should be the last dimension zeros (int): number of zero crossing to keep in the sinc filter. This function is kept only for reference, you should use the more generic `resample_frac` one. This function does not perform anti-aliasing filtering. rrpadding.N)dim)r;r_rVr=r?r.r"r-)r5r othertimeroutrHrrr _upsample2s 6rfcCs|jdddkrt|d}|ddddf}|ddddf}|j^}}t||}|tj|dd|||ddddfjg||R}|jg|dRd S) a Downsample x by a factor of two. The output length is half of the input, ceiled. Args: x (Tensor): signal to downsample, time should be the last dimension zeros (int): number of zero crossing to keep in the sinc filter. This function is kept only for reference, you should use the more generic `resample_frac` one. This function does not perform anti-aliasing filtering. rrr)rr.Nrr`rZ)r;r=r>r_rVr?r.mul)r5r xevenxoddrcrdrrerrr _downsample2s  * rj)r r NF)r )rRrtypingrr"torch.nnrr=corerutilsrnnModulerrSrr$rTrWr_rfrjrrrrs.