Vh%dZddlZddlmZddlZddlmZddlm Z ddl m Z Gdd ej j Z dd ejdededededeedefdZdZddZddZdS)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 neZdZdZddedededeffd Zd Zdd ej d e ede fdZ dZ xZS) ResampleFracz? Resampling from the sample rate `old_sr` to `new_sr`. = ףp=?old_srnew_srzerosrolloffcPtt|trt|tst dt j||}||z|_||z|_||_ ||_ | 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$old_sr and new_sr should be integersN) super__init__ isinstanceint ValueErrormathgcdr rrr _init_kernels)selfr rrrr __class__s U/var/www/html/movieo_spanner_bot/venv/lib/python3.11/site-packages/julius/resample.pyrzResampleFrac.__init__s6 &#&& Ej.E.E ECDD Dhvv&&m m    c`|j|jkrdSg}t|j|j}||jz}t j|j|jz|z |_tj |j |j|jz }t|jD]}| |jz ||jz z|z}| |j |j}|tj z}tj||jz dz dz}t||z}|||||dtj||jdddS)Nkernelr)r rminrrceilr_widthtorcharangefloatrangeclamp_picosrdiv_sumappendregister_bufferstackview)rkernelssridxitwindowr s rrzResampleFrac._init_kernelsCs{ ;$+ % % F dk * * dl&i T[ 82 =>> lDK<t{)BCCIIKKt{## # #ADK#dk/1R7A$*dj11A LAYq|A~..1F!WWv%F KK % % % NN6 " " " " Xu{7';';'@'@aQS'T'TUUUUUrNFx output_lengthfullc:|j|jkr|S|j}|jd}|d|}t j|dddf|j|j|jzfd}t j||j|j}| ddt|dddgz}tj |j|z|jz }tj |} tj|} ||r| n| } nC|dks|| krt!d | tj|} |rt!d |d d| fS) a Resample x. Args: x (Tensor): signal to resample, time should be the last dimension output_length (None or int): This can be set to the desired output length (last dimension). Allowed values are between 0 and ceil(length * new_sr / old_sr). When None (default) is specified, the floored output length will be used. In order to select the largest possible size, use the `full` argument. full (bool): return the longest possible output from the input. This can be useful if you chain resampling operations, and want to give the `output_length` only for the last one, while passing `full=True` to all the other ones. r!N replicate)mode)striderrrz$output_length must be between 0 and z0You cannot pass both full=True and output_length.)r rshapereshapeFpadr$conv1dr transposelistr% as_tensorr#longfloorrtensor) rr8r9r:r?lengthysyfloat_output_lengthmax_output_lengthdefault_output_lengthapplied_output_lengths rforwardzResampleFrac.forwardrs ;$+ % %H IIb& ! ! E!AAAtG*t{DK$+,EF[ Y Y Y XaT[ 9 9 9 LLA   & &tE#2#J'7'72$'> ? ?#odkF.BT[.PQQ!J':;;@@BB % ,? @ @ E E G G  9=$X$5$5CX ! ! Q  -2C"C"CWDUWWXX X$)L$?$? ! U !STTT,,,,--rc t|S)Nr)rs r__repr__zResampleFrac.__repr__s4   r)r r )NF)__name__ __module__ __qualname____doc__rr'rrr%TensorrboolrQrS __classcell__)rs@rr r s$$s$C$$5$$$$$$L-V-V-V^#.#.#.hsm#.RV#.#.#.#.J!!!!!!!rr r r Fr8r rrrr9r:cbt||||||||S)aR Functional version of `ResampleFrac`, refer to its documentation for more information. ..warning:: If you call repeatidly this functions with the same sample rates, then the resampling kernel will be recomputed everytime. For best performance, you should use and cache an instance of `ResampleFrac`. )r to)r8r rrrr9r:s r resample_fracr]s5 ><w 7 7 : :1 = =aPT U UUrctjd|zdzd}|ddd}tj| dz|dz d|z}|tjz}t ||zddd}|S)NrF)periodicr?r!)r% hann_windowlinspacerr*rr1)rwinwinoddr6r s r_kernel_upsample2_downsample2rfs  AIME : : :C AYF v|US[!e)< > >sABBw G L Ze ZUY Z Z ZC QH"%%%A 16 5 "   rc |jddzdkrtj|d}|ddddf}|ddddf}|j^}}t||}|tj|dd|||dddfjg||Rz}|jg|dRd S) a Downsample x by a factor of two. The output length is half of the input, ceiled. Args: x (Tensor): signal to downsample, time should be the last dimension zeros (int): number of zero crossing to keep in the sinc filter. This function is kept only for reference, you should use the more generic `resample_frac` one. This function does not perform anti-aliasing filtering. r!rr)rr.Nrrhra)r?rArBrfr\rCr1mul)r8rxevenxoddrkrlr rms r _downsample2rss wr{Q! E!V   c33Q3hKE S!$Q$Yr{sK    $$$$$$!!!!!58?!!!F5:DI V VU\ V3 V V V,1 V!)# V=A V V V V""))))))r