<p data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">ChatTTS Speaker 提供了ChatTTS生成的音色的稳定性评分,并根据性别和年龄分类,用户可以试听这些音色。</p> <p data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">ChatTTS Speaker 使用了通义实验室的 <a href="https://modelscope.cn/models/iic/speech_eres2netv2_sv_zh-cn_16k-common/summary" target="_blank" rel="noopener">ERes2NetV2</a> 说话人识别模型对音色进行打分。具体评分指标包括:</p> <ol data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"> <li data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"><strong data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">rank_long</strong>: 长句文本的音色稳定性评分</li> <li data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"><strong data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">rank_multi</strong>: 多句文本的音色稳定性评分</li> <li data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"><strong data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">rank_single</strong>: 单句文本的音色稳定性评分</li> <li data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"><strong data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">score</strong>: 音色性别、年龄、特征的可能性</li> <li data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"><strong data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">gender age feature</strong>: 音色的性别、年龄、特征</li> </ol> <p data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">这些评分指标衡量了音色在不同文本类型中的一致性,数值越高表示音色越稳定。</p> ChatTTS Speaker 项目已经对2600个音色进行了稳定性评分,并按性别和年龄进行了分类。用户可以直接查看和使用这些评分结果,以选择适合自己需求的音色。 <p data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"><strong><img class="aligncenter size-full wp-image-9828" src="https://img.xiaohu.ai/2024/06/Jietu20240617-112417@2x.jpg" alt="" width="2172" height="1774" />用途如下:</strong></p> <ol data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"> <li data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"> <p data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"><strong data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">音色评估和选择</strong>:</p> <ul data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"> <li data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">提供了一个系统化的方法对不同音色进行评估和打分,帮助用户选择在不同应用场景下音色稳定性较好的音色。</li> <li data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">用户可以根据自己的需求,比如长句、短句、多句文本的音色稳定性,选择最合适的音色。</li> </ul> </li> <li data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"> <p data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"><strong data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">声音生成和合成</strong>:</p> <ul data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"> <li data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">通过提供的音色.pt文件,用户可以在其他项目中加载和使用这些音色,用于文本到语音(TTS)转换、声音合成和生成等应用。</li> </ul> </li> <li data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"> <p data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"><strong data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">性别和年龄分类</strong>:</p> <ul data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"> <li data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">项目不仅提供了音色稳定性评分,还按照性别和年龄进行了分类,这对需要特定性别和年龄音色的应用场景(如个性化语音助手、教育软件等)非常有用。</li> </ul> </li> <li data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"> <p data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"><strong data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">音色试听和下载</strong>:</p> <ul data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8"> <li data-immersive-translate-walked="8922cbd3-2012-4ebb-a5d8-4e77bb0fdfc8">用户可以试听每个音色的样本,确定其是否适合自己的需求,然后下载.pt文件进行使用,提高了音色选择的准确性和便利性。</li> </ul> </li> </ol> <strong>在线体验</strong> <table data-immersive-translate-walked="c3e7a198-7504-4b50-a6b8-fd2aa726f560"> <tbody data-immersive-translate-walked="c3e7a198-7504-4b50-a6b8-fd2aa726f560"> <tr data-immersive-translate-walked="c3e7a198-7504-4b50-a6b8-fd2aa726f560"> <td data-immersive-translate-walked="c3e7a198-7504-4b50-a6b8-fd2aa726f560"><strong data-immersive-translate-walked="c3e7a198-7504-4b50-a6b8-fd2aa726f560" data-immersive-translate-paragraph="1">ModelScop(国内)</strong></td> <td data-immersive-translate-walked="c3e7a198-7504-4b50-a6b8-fd2aa726f560"><a href="https://modelscope.cn/studios/ttwwwaa/ChatTTS_Speaker" target="_blank" rel="nofollow noopener" data-immersive-translate-walked="c3e7a198-7504-4b50-a6b8-fd2aa726f560">https://modelscope.cn/studios/ttwwwaa/ChatTTS_Speaker</a></td> </tr> <tr data-immersive-translate-walked="c3e7a198-7504-4b50-a6b8-fd2aa726f560"> <td data-immersive-translate-walked="c3e7a198-7504-4b50-a6b8-fd2aa726f560"><strong data-immersive-translate-walked="c3e7a198-7504-4b50-a6b8-fd2aa726f560" data-immersive-translate-paragraph="1">HuggingFace<span class="notranslate immersive-translate-target-wrapper" lang="zh-CN" data-immersive-translate-translation-element-mark="1"><span class="notranslate" data-immersive-translate-translation-element-mark="1"> (海外)</span></span></strong></td> <td data-immersive-translate-walked="c3e7a198-7504-4b50-a6b8-fd2aa726f560"><a href="https://huggingface.co/spaces/taa/ChatTTS_Speaker" target="_blank" rel="nofollow noopener" data-immersive-translate-walked="c3e7a198-7504-4b50-a6b8-fd2aa726f560">https://huggingface.co/spaces/taa/ChatTTS_Speaker</a></td> </tr> </tbody> </table> GitHub:<a href="https://github.com/6drf21e/ChatTTS_Speaker" target="_blank" rel="noopener">https://github.com/6drf21e/ChatTTS_Speaker</a>