首页 > 其他 > 详细

《The challenge of realistic music generation: modelling raw audio at scale》论文阅读笔记

时间:2018-07-07 18:39:57      阅读:154      评论:0      收藏:0      [点我收藏+]

The challenge of realistic music generation: modelling raw audio at scale

 

作者:Deep  mind三位大神

出处:NIPS 2018

 

  • Abstract

首先提出了基于表达方式的音乐生成(high-level representations such as scoresor MIDI)有一些自己的问题,经过高度抽象后,音乐中的一些细节特征损失掉了,从而导致perception of musicality and realism 的损失。本文的音乐数据生成在raw audio domain中进行。autoregressive models(自回归模型)在处理波形speech数据中表现不俗,但在处理音乐时,we find them biased towards capturing local signal structure at the expense of modelling long-range correlations,于是本文提出autoregressive discrete autoencoders (ADAs) 帮助AR model capture long-range correlations in waveforms。

 

  • Introduction

强调了music在不同的timescale上展现的structure特性,并且列出了midi等表示形式的限制,主要还是在丢失音乐性相关细节和乐器相关细节上。

1.1 raw audio signal

吹了一波wave signal的好处,优势,和上面提到的midi做比较,并指出在wave形式下建模更具挑战性和难度。

 

1.2 相关生成模型

相比于表示型数据,audio waveforms生成模型的研究历史并不长,原因是:This was long thought to be infeasible due to the scale of the problem, as audio signals are often sampled at rates of 16 kHz or higher(不太明白为什么,应该是采样成本较高). 近期的AR模型采用step步进的方式来进行生成,如Wavenet,VRNN,WaveRNN,SampleRNN,解决了采样成本的问题,这里也提到了用GAN来生成波形文件。

 

贡献:1.提出文献关注点较少的raw audio domain的生成模型,可以作为benchmark测试ability of a model to capture long-range structure in data

           2. We investigate the capabilities of autoregressive models for this task, and demonstrate a computationally efficient method to enlarge their receptive fields using autoregressive discrete autoencoders (ADAs)

           3. introduce the argmax autoencoder (AMAE) as an alternative to vector quantisation variational autoencoders (VQ-VAE)

  • Scaling up autoregressive models for music

要为long-range structure建模,需要enlarge the receptive fields,wavenet,sampleRNN都提出自己的方式来扩大接受野,但内存限制很容易触及天花板

(未完待续)

 

重要参考文献:

Arecurrent latent variable model for sequential data

Experiments in musical intelligence

Synthesizing audio with generative adversarial networks

Samplernn: An unconditional end-to-end neural audio generation model

 

 

 

《The challenge of realistic music generation: modelling raw audio at scale》论文阅读笔记

原文:https://www.cnblogs.com/punkcure/p/9277681.html

(0)
(0)
   
举报
评论 一句话评论(0
关于我们 - 联系我们 - 留言反馈 - 联系我们:wmxa8@hotmail.com
© 2014 bubuko.com 版权所有
打开技术之扣,分享程序人生!