| Gregor Richards | 0210592 | 2018-09-20 21:58:31 -0400 | [diff] [blame] | 1 | RNNoise is a noise suppression library based on a recurrent neural network. |
| Jean-Marc Valin | 3d260c4 | 2017-09-01 17:45:11 -0400 | [diff] [blame] | 2 | |
| 3 | To compile, just type: |
| Jan | 3d2acb8 | 2017-10-03 14:14:06 +0200 | [diff] [blame] | 4 | % ./autogen.sh |
| Jean-Marc Valin | 3d260c4 | 2017-09-01 17:45:11 -0400 | [diff] [blame] | 5 | % ./configure |
| 6 | % make |
| 7 | |
| 8 | Optionally: |
| 9 | % make install |
| 10 | |
| 11 | While it is meant to be used as a library, a simple command-line tool is |
| 12 | provided as an example. It operates on RAW 16-bit (machine endian) mono |
| 13 | PCM files sampled at 48 kHz. It can be used as: |
| 14 | |
| Danielius Visockas | 40a3816 | 2021-01-22 16:22:28 +0200 | [diff] [blame] | 15 | ./examples/rnnoise_demo <noisy speech> <output denoised> |
| Jean-Marc Valin | 3d260c4 | 2017-09-01 17:45:11 -0400 | [diff] [blame] | 16 | |
| 17 | The output is also a 16-bit raw PCM file. |
| Petter Reinholdtsen | bd7a29c | 2021-01-18 09:36:04 +0100 | [diff] [blame] | 18 | |
| 19 | The latest version of the source is available from |
| 20 | https://gitlab.xiph.org/xiph/rnnoise . The github repository |
| 21 | is a convenience copy. |