EGDB-PG
收藏资源简介:
EGDB-PG数据集由台湾大学和Positive Grid公司共同创建,旨在解决现有吉他数据集在音色多样性方面的不足。该数据集通过使用Positive Grid的BiasFX2插件,对EGDB数据集中的干净音频进行重新渲染,产生了256种独特的预设,覆盖了从低增益到高增益的多种放大器类型。EGDB-PG数据集包含514小时的音频,大大增强了音色覆盖范围,为电吉他音频转录模型的训练和评估提供了丰富的资源。
The EGDB-PG dataset was co-created by National Taiwan University and Positive Grid, aiming to address the deficiency in timbral diversity of existing guitar datasets. This dataset re-renders the clean audio from the original EGDB dataset using Positive Grid’s BiasFX2 plugin, generating 256 unique presets covering a wide range of amplifier types from low-gain to high-gain configurations. The EGDB-PG dataset contains 514 hours of audio, greatly expanding timbral coverage and providing a rich resource for the training and evaluation of electric guitar audio transcription models.




