Papers › Neural HMMs are all you need (for high-quality attention-free TTS)

Neural HMMs are all you need (for high-quality attention-free TTS)

30 Aug 2021arXiv:2108.13320archive 2025-07-28

Shivam Mehta, Éva Székely, Jonas Beskow, Gustav Eje Henter

Neural sequence-to-sequence TTS has achieved significantly better output quality than statistical speech synthesis using HMMs. However, neural TTS is generally not probabilistic and uses non-monotonic attention. Attention failures increase training time and can make synthesis babble incoherently. This paper describes how the old and new paradigms can be combined to obtain the advantages of both worlds, by replacing attention in neural TTS with an autoregressive left-right no-skip hidden Markov model defined by a neural network. Based on this proposal, we modify Tacotron 2 to obtain an HMM-based neural TTS model with monotonic alignment, trained to maximise the full sequence likelihood without approximation. We also describe how to combine ideas from classical and contemporary TTS for best results. The resulting example system is smaller and simpler than Tacotron 2, and learns to speak with fewer iterations and less data, whilst achieving comparable naturalness prior to the post-net. Our approach also allows easy control over speaking rate.

PaperPDFCode

Code

shivammehta25/Neural-HMM officialmentioned on GitHubpytorch report
coqui-ai/TTS mentioned on GitHubpytorchMPL-2.0 report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

AllSpeech Synthesis

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Speech Synthesis LJSpeech Neural HMM Mean Opinion Score 3.24 #3 of 4 Archive leaderboard report
Speech Synthesis LJSpeech Neural HMM Ablation with 1 state per phone Mean Opinion Score 2.68 #4 of 4 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Batch NormalizationBiGRUBiLSTMCBHGConvolutionDense ConnectionsDilated Causal ConvolutionDropoutGRUGriffin-Lim AlgorithmHighway LayerHighway NetworkLSTMLinear LayerLocation Sensitive AttentionMax PoolingMixture of Logistic DistributionsReLUResidual ConnectionResidual GRUSigmoid ActivationTacotronTacotron 2Tanh ActivationWaveNetZoneout

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections