change to direct audio

accent_gradio.py

@@ -205,7 +205,8 @@ def transcribe(audio):
         transcribed_text = "An error occured. Please try again."  
         
     print(transcribed_text)
-    ref_s = compute_style(original_voice_path)
+    # ref_s = compute_style(original_voice_path) # run locally
+    ref_s = compute_style(audio) # run on HF
     wav = inference(transcribed_text, ref_s, alpha=0.1, beta=0.5, diffusion_steps=10, embedding_scale=1)
     scaled = np.int16(wav / np.max(np.abs(wav)) * 32767)
     return (sample_rate_value, scaled)
@@ -222,27 +223,27 @@ with gr.Blocks() as demo:
     ### Transform your non-native accent into a native North American accent.
     **This is an educational app designed to transform the speech of a non-native English speaker into a native American accent.**
     **The tool aims to assist learners in accent reduction and pronunciation improvement.**
-    **For newcomers, kindly navigate to the *First-Time Users* section by clicking on the designated link.**
+    **This tool performs better on longer speech.**
     """)
-    with gr.Accordion("First-Time Users (Click Here):", open=False):
-        gr.Markdown("""
-            **Record the reference voice:** Kindly capture your voice as you read the provided
-            text. Please ensure that you have granted microphone access in your browser settings.
-
-            > I must not fear. Fear is the mind-killer. Fear is the little-death that brings total obliteration. 
-            I will face my fear. I will permit it to pass over me and through me. And when it has gone past I 
-            will turn the inner eye to see its path. Where the fear has gone there will be nothing. Only I will remain.
-
-            Ensure clarity in your pronunciation, as the quality of this recording
-            significantly influences the future results. Once done please click "Save".
-            If the quality of the native voice is not satisfactory, you can come back and
-            re-record your voice here again.
-
-            You can also upload your voice or someone else's voice.
-        """)
-        speaker_voice = gr.Audio(sources=["microphone", "upload"], format="wav", label="Record reference voice:",show_download_button="True")
-        ref_btn = gr.Button("Save")
-        ref_btn.click(record_speaker, inputs= speaker_voice, outputs=None)
+    # with gr.Accordion("First-Time Users (Click Here):", open=False):
+    #     gr.Markdown("""
+    #         **Record the reference voice:** Kindly capture your voice as you read the provided
+    #         text. Please ensure that you have granted microphone access in your browser settings.
+
+    #         > I must not fear. Fear is the mind-killer. Fear is the little-death that brings total obliteration. 
+    #         I will face my fear. I will permit it to pass over me and through me. And when it has gone past I 
+    #         will turn the inner eye to see its path. Where the fear has gone there will be nothing. Only I will remain.
+
+    #         Ensure clarity in your pronunciation, as the quality of this recording
+    #         significantly influences the future results. Once done please click "Save".
+    #         If the quality of the native voice is not satisfactory, you can come back and
+    #         re-record your voice here again.
+
+    #         You can also upload your voice or someone else's voice.
+    #     """)
+    #     speaker_voice = gr.Audio(sources=["microphone", "upload"], format="wav", label="Record reference voice:",show_download_button="True")
+    #     ref_btn = gr.Button("Save")
+    #     ref_btn.click(record_speaker, inputs= speaker_voice, outputs=None)
 
 
     with gr.Column():

original_voice.wav

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:ca4a4ef995535c9b58b46c4ba5a9512772a060a6e1ca5bdb1a23eba6fbf695e2
-size 2304044