TGRadioBooks
How to compare book narration models using one page
To compare book narration models, play the same short passage at the same speed and volume. Change the model first while keeping either the male or female voice selected. Compare voices separately afterward. This makes it easier to tell which setting changed the pronunciation, pauses or intonation in your particular book.
Choose a page with varied sentences
Pick a short passage that you can find again: a few adjacent paragraphs containing complete sentences. Include some narration, a question and a line of dialogue if possible. For nonfiction, a paragraph with a technical term and its explanation can work. A heading alone gives you too little material to judge how the reading flows.
For example, if the book contains foreign surnames, include one in your sample. Note the chapter and opening words so you do not accidentally compare different passages.
Keep the listening conditions consistent
Choose one playback speed and use it for every version. You can start at 1 even if you usually listen faster. Use the same headphones or speaker, and leave the device volume unchanged between attempts. A louder recording may seem more expressive, so adjusting the volume can make the comparison less useful.
Keep the same voice option too: either male or female. Changing the model, voice and speed together makes it difficult to identify which setting produced a noticeable difference.
Try the available models one at a time
TGRadio settings include models 1, 2 and 3. Compare whichever options are currently available to you; every model is not guaranteed to remain available at all times. After selecting another model, return to the same passage and press Play. A different model requires a new audio version to be prepared.
Importing a book does not start narration by itself. Preparation begins after you press Play, and subsequent fragments are prepared as you listen. Keep waiting time separate from your assessment of how the narration sounds.
Record specific differences
After each version, note three things: whether the words are clear, whether pauses fit the text, and whether the intonation suits each sentence. Describe what you heard, for example: “the question sounds like a statement” or “the surname is easy to understand.” A general preference is useful too, but it does not explain the difference.
If two models differ mainly on one name, test another paragraph without that name. This prevents a single pronunciation error from becoming your only reason for choosing.
Compare voices after selecting a model
Keep your preferred model and listen to the same passage with the male and female voices separately. Changing the voice also requires a new audio version. Then restore your usual playback speed and test a few more paragraphs: a voice you enjoy at 1 may feel different at 1.5.
Keep a personal note of the model, voice and speed alongside the book title. For another book, repeat the test with representative material. Dialogue, long descriptions and technical vocabulary give you different things to listen for.