Possible to preserve formants when changing pitch?

[Turkrosoft]

Is it possible to preserve the formants when adjusting pitch? Many TTS engines support that. There may also be two separate modes or a check box letting the user adjust the pitch by or without effecting the formants.

[rmcpantoja]

In fact, I was able to do this in a Discord bot for Piper using praat-parcelmout.
Here's a code snippet:

def change_pitch_with_praat(audio_file, factor):
	snd = parselmouth.Sound(audio_file)
	manipulation = call(snd, "To Manipulation", 0.01, 75, 600)
	pitch_tier = call(manipulation, "Extract pitch tier")
	call(pitch_tier, "Multiply frequencies", snd.xmin, snd.xmax, factor)
	call([pitch_tier, manipulation], "Replace pitch tier")
	return call(manipulation, "Get resynthesis (overlap-add)")

async def piper(
	interaction: discord.Interaction,
	voice: str,
	speaker: Optional[int] = 0,
	rate: Optional[float] = 1.00,
	pitch: Optional[float] = 1.00,
	text: Optional[str] = "This is a test."
):
	loop = asyncio.get_event_loop()
	await interaction.response.send_message(f'Text-To-Speech request sent by {interaction.user.mention}.\nStarting...', ephemeral=True)
	task = loop.create_task(piper_thread(discord.Interaction, voice, speaker, rate, text))
	wav_filename = await task
	if wav_filename == -1:
		await interaction.channel.send(f'This voice doesn't exists! Please use /list to get a list. The voice was: {voz}.')
		return
	utc_now = datetime.datetime.now(timezone.utc)
	if pitch != 1.00:
		print(pitch)
		audio = change_pitch_with_praat(wav_filename, pitch)
		wav_filename = f"{wav_filename[:-4]}_changed.wav"
		audio.save(wav_filename, "WAV")
	if (utc_now - interaction.created_at).total_seconds() < 900:
		if len(text) > 1500:
			realtest = "Large text."
		else:
			realtest = text
		await interaction.channel.send (f"{interaction.user.mention}, this is your audio generated with {voice}!\nText: {realtest}", file=discord.File(wav_filename))
	else:
		print("Error! The interaction has expired and a response can't be sent.")