how to get the TextGrid file from a wav?

[sjq19960802]

I only have a wav file, how can i generate the TextGrid file from it?

[AaronZ345]

Hi. You can refer to the annotation process descried in our paper to conduct the corresponding annotations.

[HuZhetao]

Hi. You can refer to the annotation process descried in our paper to conduct the corresponding annotations.

could you share the MFA model.zip

[AaronZ345]

Hi. You can refer to the annotation process descried in our paper to conduct the corresponding annotations.

could you share the MFA model.zip

You can find model for each language in https://mfa-models.readthedocs.io/en/latest/dictionary/index.html#dictionary

[HuZhetao]

Hi. You can refer to the annotation process descried in our paper to conduct the corresponding annotations.

could you share the MFA model.zip

You can find model for each language in https://mfa-models.readthedocs.io/en/latest/dictionary/index.html#dictionary

Did you train the model on you own dataset? I find the official model permformance poor on my sing dataset

[AaronZ345]

Hi. You can refer to the annotation process descried in our paper to conduct the corresponding annotations.

could you share the MFA model.zip

You can find model for each language in https://mfa-models.readthedocs.io/en/latest/dictionary/index.html#dictionary

Did you train the model on you own dataset? I find the official model permformance poor on my sing dataset

Sorry, I misunderstood earlier. We have an internal model, but it may not be available for public release. You can try using mfa adapt or use Whisper X or SOFA. To be honest, even our own results aren't great, as we’ve relied heavily on manual adjustments. We are currently working on an open-source SOTA model for aligning singing voices, so please look forward to it in the next few months.

[JasonBourne1998]

Hi. You can refer to the annotation process descried in our paper to conduct the corresponding annotations.

could you share the MFA model.zip

You can find model for each language in https://mfa-models.readthedocs.io/en/latest/dictionary/index.html#dictionary

Did you train the model on you own dataset? I find the official model permformance poor on my sing dataset

Sorry, I misunderstood earlier. We have an internal model, but it may not be available for public release. You can try using mfa adapt or use Whisper X or SOFA. To be honest, even our own results aren't great, as we’ve relied heavily on manual adjustments. We are currently working on an open-source SOTA model for aligning singing voices, so please look forward to it in the next few months.

Hi @AaronZ345 , any process on the aligning singing voices work?

[AaronZ345]

Hi. You can refer to the annotation process descried in our paper to conduct the corresponding annotations.

could you share the MFA model.zip

You can find model for each language in https://mfa-models.readthedocs.io/en/latest/dictionary/index.html#dictionary

Did you train the model on you own dataset? I find the official model permformance poor on my sing dataset

Sorry, I misunderstood earlier. We have an internal model, but it may not be available for public release. You can try using mfa adapt or use Whisper X or SOFA. To be honest, even our own results aren't great, as we’ve relied heavily on manual adjustments. We are currently working on an open-source SOTA model for aligning singing voices, so please look forward to it in the next few months.

Hi @AaronZ345 , any process on the aligning singing voices work?

Hello, this work has been completed and submitted. We will open source it after the paper is accepted. It is expected to be within a few months.

[JasonBourne1998]

Hi. You can refer to the annotation process descried in our paper to conduct the corresponding annotations.

could you share the MFA model.zip

You can find model for each language in https://mfa-models.readthedocs.io/en/latest/dictionary/index.html#dictionary

Did you train the model on you own dataset? I find the official model permformance poor on my sing dataset

Sorry, I misunderstood earlier. We have an internal model, but it may not be available for public release. You can try using mfa adapt or use Whisper X or SOFA. To be honest, even our own results aren't great, as we’ve relied heavily on manual adjustments. We are currently working on an open-source SOTA model for aligning singing voices, so please look forward to it in the next few months.

Hi @AaronZ345 , any process on the aligning singing voices work?

Hello, this work has been completed and submitted. We will open source it after the paper is accepted. It is expected to be within a few months.

Hi @AaronZ345, sorry to bother you again. Congratulations on the ACL'25 acceptance. Will this phoneme alignment workflow be open-sourced with Techsinger 2 ?