{"id":20214318,"url":"https://github.com/yl4579/hiftnet","last_synced_at":"2025-04-05T04:11:45.948Z","repository":{"id":195695026,"uuid":"693453235","full_name":"yl4579/HiFTNet","owner":"yl4579","description":"HiFTNet: A Fast High-Quality Neural Vocoder with Harmonic-plus-Noise Filter and Inverse Short Time Fourier Transform","archived":false,"fork":false,"pushed_at":"2025-01-14T00:47:28.000Z","size":57821,"stargazers_count":157,"open_issues_count":6,"forks_count":13,"subscribers_count":9,"default_branch":"main","last_synced_at":"2025-03-29T03:11:18.740Z","etag":null,"topics":["deep-learning","speech-synthesis","text-to-speech","tts","vocoder","vocoders"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/yl4579.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2023-09-19T04:12:34.000Z","updated_at":"2025-03-27T09:34:33.000Z","dependencies_parsed_at":"2025-01-14T01:41:21.787Z","dependency_job_id":null,"html_url":"https://github.com/yl4579/HiFTNet","commit_stats":null,"previous_names":["yl4579/hiftnet"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/yl4579%2FHiFTNet","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/yl4579%2FHiFTNet/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/yl4579%2FHiFTNet/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/yl4579%2FHiFTNet/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/yl4579","download_url":"https://codeload.github.com/yl4579/HiFTNet/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":247284949,"owners_count":20913704,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["deep-learning","speech-synthesis","text-to-speech","tts","vocoder","vocoders"],"created_at":"2024-11-14T06:15:16.163Z","updated_at":"2025-04-05T04:11:40.937Z","avatar_url":"https://github.com/yl4579.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# HiFTNet: A Fast High-Quality Neural Vocoder with Harmonic-plus-Noise Filter and Inverse Short Time Fourier Transform\n\n### Yinghao Aaron Li, Cong Han, Xilin Jiang, Nima Mesgarani\n\n\u003e Recent advancements in speech synthesis have leveraged GAN-based networks like HiFi-GAN and BigVGAN to produce high-fidelity waveforms from mel-spectrograms. However, these networks are computationally expensive and parameter-heavy. iSTFTNet addresses these limitations by integrating inverse short-time Fourier transform (iSTFT) into the network, achieving both speed and parameter efficiency. In this paper, we introduce an extension to iSTFTNet, termed HiFTNet, which incorporates a harmonic-plus-noise source filter in the time-frequency domain that uses a sinusoidal source from the fundamental frequency (F0) inferred via a pre-trained F0 estimation network for fast inference speed. Subjective evaluations on LJSpeech show that our model significantly outperforms both iSTFTNet and HiFi-GAN, achieving ground-truth-level performance. HiFTNet also outperforms BigVGAN-base on LibriTTS for unseen speakers and achieves comparable performance to BigVGAN while being four times faster with only 1/6 of the parameters. Our work sets a new benchmark for efficient, high-quality neural vocoding, paving the way for real-time applications that demand high quality speech synthesis.\n\nPaper: [https://arxiv.org/abs/2309.09493](https://arxiv.org/abs/2309.09493)\n\nAudio samples: [https://hiftnet.github.io/](https://hiftnet.github.io/)\n\n**Check our TTS work that uses HiFTNet as speech decoder for human-level speech synthesis here: https://github.com/yl4579/StyleTTS2**\n\n## Pre-requisites\n1. Python \u003e= 3.7\n2. Clone this repository:\n```bash\ngit clone https://github.com/yl4579/HiFTNet.git\ncd HiFTNet\n```\n3. Install python requirements: \n```bash\npip install -r requirements.txt\n```\n\n## Training\n```bash\npython train.py --config config_v1.json --[args]\n```\nFor the F0 model training, please refer to [yl4579/PitchExtractor](https://github.com/yl4579/PitchExtractor). This repo includes a pre-trained F0 model on LibriTTS. Still, you may want to train your own F0 model for the best performance, particularly for noisy or non-speech data, as we found that F0 estimation accuracy is essential for the vocoder performance. \n\n## Inference\nPlease refer to the notebook [inference.ipynb](https://github.com/yl4579/HiFTNet/blob/main/inference.ipynb) for details.\n### Pre-Trained Models\nYou can download the pre-trained LJSpeech model [here](https://huggingface.co/yl4579/HiFTNet/blob/main/LJSpeech/cp_hifigan.zip) and the pre-trained LibriTTS model [here](https://huggingface.co/yl4579/HiFTNet/blob/main/LibriTTS/cp_hifigan.zip). The pre-trained models contain parameters of the optimizers and discriminators that can be used for fine-tuning.  \n\n## References\n- [rishikksh20/iSTFTNet-pytorch](https://github.com/rishikksh20/iSTFTNet-pytorch)\n- [nii-yamagishilab/project-NN-Pytorch-scripts/project/01-nsf](https://github.com/nii-yamagishilab/project-NN-Pytorch-scripts/tree/master/project/01-nsf)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fyl4579%2Fhiftnet","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fyl4579%2Fhiftnet","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fyl4579%2Fhiftnet/lists"}