"Defaults to <datasets_root>/SV2TTS/vocoder/. Unused if --ground_truth is passed.") parser.add_argument("-m", "--models_dir", type=str, default="vocoder/saved_models/", help=\ "Path to the directory that will contain the saved model weights, as well as backups " "of those weights and wavs generated during training.") parser.add_argument("-g", "--ground_truth", action="store_true", help= \ "Train on ground truth spectrograms (<datasets_root>/SV2TTS/synthesizer/mels).") parser.add_argument("-s", "--save_every", type=int, default=1000, help= \ "Number of steps between updates of the model on the disk. Set to 0 to never save the " "model.") parser.add_argument("-b", "--backup_every", type=int, default=25000, help= \ "Number of steps between backups of the model. Set to 0 to never make backups of the " "model.") parser.add_argument("-f", "--force_restart", action="store_true", help= \ "Do not load any saved model and restart from scratch.") args = parser.parse_args() # Process the arguments if not hasattr(args, "syn_dir"): args.syn_dir = Path(args.datasets_root, "SV2TTS", "synthesizer") args.syn_dir = Path(args.syn_dir) if not hasattr(args, "voc_dir"): args.voc_dir = Path(args.datasets_root, "SV2TTS", "vocoder") args.voc_dir = Path(args.voc_dir) del args.datasets_root args.models_dir = Path(args.models_dir) args.models_dir.mkdir(exist_ok=True) # Run the training print_args(args, parser) train(**vars(args))
def sythesize_voice(source_voice, script): ROOT_DIR = os.path.dirname(os.path.abspath(__file__)) encoder_model_path = os.path.relpath( "encoder/saved_models/pretrained.pt") ## Info & args parser = argparse.ArgumentParser( formatter_class=argparse.ArgumentDefaultsHelpFormatter) parser.add_argument("-e", "--enc_model_fpath", type=Path, default=ROOT_DIR + "/" + "encoder/saved_models/pretrained.pt", help="Path to a saved encoder") parser.add_argument("-s", "--syn_model_dir", type=Path, default=ROOT_DIR + "/" + "synthesizer/saved_models/logs-pretrained/", help="Directory containing the synthesizer model") parser.add_argument("-v", "--voc_model_fpath", type=Path, default=ROOT_DIR + "/" + "vocoder/saved_models/pretrained/pretrained.pt", help="Path to a saved vocoder") parser.add_argument("--low_mem", action="store_true", help=\ "If True, the memory used by the synthesizer will be freed after each use. Adds large " "overhead but allows to save some GPU memory for lower-end GPUs.") parser.add_argument("--no_sound", action="store_true", help=\ "If True, audio won't be played.") args = parser.parse_args() print_args(args, parser) if not args.no_sound: import sounddevice as sd ## Load the models one by one. print("Preparing the encoder, the synthesizer and the vocoder...") encoder.load_model(args.enc_model_fpath) synthesizer = Synthesizer( args.syn_model_dir.joinpath("taco_pretrained"), low_mem=args.low_mem) vocoder.load_model(args.voc_model_fpath) ## Run a test print("Testing your configuration with small inputs.") # Forward an audio waveform of zeroes that lasts 1 second. Notice how we can get the encoder's # sampling rate, which may differ. # If you're unfamiliar with digital audio, know that it is encoded as an array of floats # (or sometimes integers, but mostly floats in this projects) ranging from -1 to 1. # The sampling rate is the number of values (samples) recorded per second, it is set to # 16000 for the encoder. Creating an array of length <sampling_rate> will always correspond # to an audio of 1 second. print("\tTesting the encoder...") encoder.embed_utterance(np.zeros(encoder.sampling_rate)) # Create a dummy embedding. You would normally use the embedding that encoder.embed_utterance # returns, but here we're going to make one ourselves just for the sake of showing that it's # possible. embed = np.random.rand(speaker_embedding_size) # Embeddings are L2-normalized (this isn't important here, but if you want to make your own # embeddings it will be). embed /= np.linalg.norm(embed) # The synthesizer can handle multiple inputs with batching. Let's create another embedding to # illustrate that embeds = [embed, np.zeros(speaker_embedding_size)] texts = ["test 1", "test 2"] print( "\tTesting the synthesizer... (loading the model will output a lot of text)" ) mels = synthesizer.synthesize_spectrograms(texts, embeds) # The vocoder synthesizes one waveform at a time, but it's more efficient for long ones. We # can concatenate the mel spectrograms to a single one. mel = np.concatenate(mels, axis=1) # The vocoder can take a callback function to display the generation. More on that later. For # now we'll simply hide it like this: no_action = lambda *args: None print("\tTesting the vocoder...") # For the sake of making this test short, we'll pass a short target length. The target length # is the length of the wav segments that are processed in parallel. E.g. for audio sampled # at 16000 Hertz, a target length of 8000 means that the target audio will be cut in chunks of # 0.5 seconds which will all be generated together. The parameters here are absurdly short, and # that has a detrimental effect on the quality of the audio. The default parameters are # recommended in general. vocoder.infer_waveform(mel, target=200, overlap=50, progress_callback=no_action) print("All test passed! You can now synthesize speech.\n\n") ## Interactive speech generation print( "This is a GUI-less example of interface to SV2TTS. The purpose of this script is to " "show how you can interface this project easily with your own. See the source code for " "an explanation of what is happening.\n") print("Interactive generation loop") num_generated = 0 # while True: try: # Get the reference audio filepath message = "Reference voice: enter an audio filepath of a voice to be cloned (mp3, " \ "wav, m4a, flac, ...):\n" # replacing this with new code. source_voice is the voice to be cloned # in_fpath = Path(input(message).replace("\"", "").replace("\'", "")) in_fpath = source_voice ## Computing the embedding # First, we load the wav using the function that the speaker encoder provides. This is # important: there is preprocessing that must be applied. # The following two methods are equivalent: # - Directly load from the filepath: preprocessed_wav = encoder.preprocess_wav(in_fpath) # - If the wav is already loaded: original_wav, sampling_rate = librosa.load(in_fpath) preprocessed_wav = encoder.preprocess_wav(original_wav, sampling_rate) print("Loaded file succesfully") # Then we derive the embedding. There are many functions and parameters that the # speaker encoder interfaces. These are mostly for in-depth research. You will typically # only use this function (with its default parameters): embed = encoder.embed_utterance(preprocessed_wav) print("Created the embedding") ## Generating the spectrogram # hard coding text to be synthesized for now # text = input("Write a sentence (+-20 words) to be synthesized:\n") # text = script # texts = script.split(".") texts = re.split('[.:;?!]', script) # print(texts) while ("" in texts): texts.remove("") print(texts) # The synthesizer works in batch, so you need to put your data in a list or numpy array # texts = [text] # embeds = [embed] embeds = np.stack([embed] * len(texts)) # If you know what the attention layer alignments are, you can retrieve them here by # passing return_alignments=True specs = synthesizer.synthesize_spectrograms(texts, embeds) # spec = specs[0] spec = np.concatenate(specs, axis=1) print("Created the mel spectrogram") ## Generating the waveform print("Synthesizing the waveform:") # Synthesizing the waveform is fairly straightforward. Remember that the longer the # spectrogram, the more time-efficient the vocoder. generated_wav = vocoder.infer_waveform(spec) ## Post-generation # There's a bug with sounddevice that makes the audio cut one second earlier, so we # pad it. generated_wav = np.pad(generated_wav, (0, synthesizer.sample_rate), mode="constant") # Play the audio (non-blocking) # if not args.no_sound: # sd.stop() # sd.play(generated_wav, synthesizer.sample_rate) # Save it on the disk # fpath = "static/cloned/demo_output_%02d.wav" % num_generated fpath = "static/cloned/" + time.strftime("%Y%m%d-%H%M%S") + ".wav" print(generated_wav.dtype) temp_output = Audio_Cleaner.reduce_noise_centroid_mb( generated_wav, synthesizer.sample_rate) cleaned_wav, cleaned_wav_sr = Audio_Cleaner.trim_silence( temp_output) print(cleaned_wav_sr) print(synthesizer.sample_rate) librosa.output.write_wav(fpath, generated_wav.astype(np.float32), synthesizer.sample_rate) num_generated += 1 print("\nSaved output as %s\n\n" % fpath) print(time.strftime("%Y%m%d-%H%M%S")) script_path = './silence_remover.sh' trimmed_output_path = "static/cloned/" + time.strftime( "%Y%m%d-%H%M%S") + "t.wav" subprocess.call( shlex.split( '%s %s %s' % (script_path, str(fpath), str(trimmed_output_path)))) except Exception as e: print("Caught exception: %s" % repr(e)) print("Restarting\n") return trimmed_output_path #synthesized_voice_output