From d9ca6f8fee88b8d51185fbfe6d1076a3e9114181 Mon Sep 17 00:00:00 2001
From: Pawel Andrzejuk <andrzejuk94@gmail.com>
Date: Wed, 17 Jun 2026 10:52:03 +0200
Subject: [PATCH] feat: remove deepgram sample rate settings, infer sample rate
 from audio format selected fix: incorrect speed setting range for deepgram
 tts

---
 src/api/generated/data-contracts.ts         |  4 +-
 src/api/openapi.json                        |  4 +-
 src/components/TtsProviderSettingsPanel.vue | 49 ++++++++++++---------
 3 files changed, 33 insertions(+), 24 deletions(-)

diff --git a/src/api/generated/data-contracts.ts b/src/api/generated/data-contracts.ts
index 5ebd0b4..3b3e023 100644
--- a/src/api/generated/data-contracts.ts
+++ b/src/api/generated/data-contracts.ts
@@ -849,8 +849,8 @@ export interface DeepgramTtsSettings {
   useSentenceSplitter?: boolean;
   /**
    * Speaking rate multiplier (0.25 to 4.0, default: 1.0)
-   * @min 0.25
-   * @max 4
+   * @min 0.75
+   * @max 1.5
    */
   speed?: number;
   [key: string]: any;
diff --git a/src/api/openapi.json b/src/api/openapi.json
index dbd5b42..3ce5921 100644
--- a/src/api/openapi.json
+++ b/src/api/openapi.json
@@ -1273,8 +1273,8 @@
           },
           "speed": {
             "type": "number",
-            "minimum": 0.25,
-            "maximum": 4,
+            "minimum": 0.75,
+            "maximum": 1.5,
             "description": "Speaking rate multiplier (0.25 to 4.0, default: 1.0)"
           }
         },
diff --git a/src/components/TtsProviderSettingsPanel.vue b/src/components/TtsProviderSettingsPanel.vue
index ddad4fa..9f2fdda 100644
--- a/src/components/TtsProviderSettingsPanel.vue
+++ b/src/components/TtsProviderSettingsPanel.vue
@@ -1,5 +1,5 @@
 <script setup lang="ts">
-import { computed } from 'vue'
+import { computed, watch } from 'vue'
 import { Plus, X } from 'lucide-vue-next'
 import FormField from '@/components/FormField.vue'
 import type {
@@ -48,6 +48,28 @@ const emotionTagsInput = computed({
   },
 })
 
+const deepgramAudioFormat = computed(() => {
+  if (!isDeepgram.value) return undefined
+  return (model.value as DeepgramTtsSettings).audioFormat
+})
+
+watch(deepgramAudioFormat, (newFormat) => {
+  if (!isDeepgram.value) return
+  const settings = model.value as DeepgramTtsSettings
+  if (!newFormat) {
+    settings.audioFormat = 'pcm_16000'
+    settings.sampleRate = 16000
+    return
+  }
+  if (newFormat.startsWith('pcm_')) {
+    const parts = newFormat.split('_')
+    settings.sampleRate = parts.length > 1 ? parseInt(parts[1] ?? '16000', 10) : 16000
+  }
+  else if (newFormat === 'mulaw' || newFormat === 'alaw') {
+    settings.sampleRate = 8000
+  }
+}, { immediate: true })
+
 function addNoSpeechMarker() {
   const settings = model.value as any
   if (!settings.noSpeechMarkers) {
@@ -65,7 +87,7 @@ function removeNoSpeechMarker(index: number) {
 </script>
 
 <template>
-  <!-- Voice Settings Section (OpenAI) -->
+          <!-- Voice Settings Section (OpenAI) -->
           <div v-if="isOpenAI" class="mt-8 pt-6 border-t border-gray-200 dark:border-gray-700">
             <h3 class="text-lg font-semibold text-gray-900 dark:text-white mb-4">Voice Settings (OpenAI)</h3>
 
@@ -81,7 +103,7 @@ function removeNoSpeechMarker(index: number) {
             </FormField>
 
             <!-- Speed -->
-            <FormField :label="`Speed: ${((model as any).speed ?? 1.0).toFixed(2)}`" class="w-full" help="Speech speed (0.25-4.0), defaults to 1.0">
+            <FormField :label="`Speed: ${((model as OpenAiTtsSettings).speed ?? 1.0).toFixed(2)}`" class="w-full" help="Speech speed (0.25-4.0), defaults to 1.0">
               <input
                 v-model.number="(model as any).speed"
                 type="range"
@@ -99,32 +121,19 @@ function removeNoSpeechMarker(index: number) {
             <h3 class="text-lg font-semibold text-gray-900 dark:text-white mb-4">Voice Settings (Deepgram)</h3>
 
             <!-- Speed -->
-            <FormField :label="`Speed: ${((model as any).speed ?? 1.0).toFixed(2)}`" class="w-full" help="Speech speed (0.25-4.0), defaults to 1.0">
+            <FormField :label="`Speed: ${((model as DeepgramTtsSettings).speed ?? 1.0).toFixed(2)}`" class="w-full" help="Speech speed (0.75-1.5), defaults to 1.0">
               <input
                 v-model.number="(model as DeepgramTtsSettings).speed"
                 type="range"
-                min="0.25"
-                max="4.0"
+                min="0.75"
+                max="1.5"
                 step="0.01"
                 class="block min-w-64 h-2 bg-gray-200 rounded-lg appearance-none cursor-pointer"
                 :disabled="isLoading"
               />
             </FormField>
 
-            <!-- Sample Rate -->
-            <FormField label="Sample Rate (Hz)" class="w-full" help="Audio sample rate in Hz. Higher values provide better quality but larger file sizes. Common values: 8000, 16000, 24000, 48000.">
-              <select
-                v-model.number="(model as DeepgramTtsSettings).sampleRate"
-                class="form-select-auto min-w-64"
-                :disabled="isLoading"
-              >
-                <option :value="undefined">Default</option>
-                <option :value="8000">8000 Hz</option>
-                <option :value="16000">16000 Hz</option>
-                <option :value="24000">24000 Hz (Recommended)</option>
-                <option :value="48000">48000 Hz</option>
-              </select>
-            </FormField>
+
           </div>
 
           <!-- Voice Settings Section (Cartesia) -->