Audio Models - NeoSpeech API

Overview

NeoSpeech offers 5 specialized audio models, each optimized for different quality, speed, and use case requirements. Choose the model that best balances audio fidelity with processing latency for your application.

Available Models

Aurora 4

Premium Quality48kHz, 192kbps, studio-grade detail

Aurora 3.5

High Quality (Recommended)48kHz, 96kbps, balanced performance

Aurora 3

Standard Quality24kHz, 160kbps, everyday clarity

Turbo 3

Fast Processing16kHz, 64kbps, rapid responses

Mini 2

Lightweight24kHz, 48kbps, lightweight delivery

Model Comparison

Model	Quality	Sample Rate	Bitrate	File Size	Best For
aurora-4	Premium	48kHz	192kbps	Large	Broadcasting, professional media
aurora-3.5	High	48kHz	96kbps	Medium	Podcasts, audiobooks, e-learning
aurora-3	Standard	24kHz	160kbps	Medium	General content creation
turbo-3	Good	16kHz	64kbps	Small	Real-time apps, chatbots, IVR
mini-2	Basic	24kHz	48kbps	Small	IoT devices, mobile apps

Model Details

Aurora 4 - Premium Quality

Technical Specifications:

Sample Rate: 48kHz
Bitrate: 192kbps
Quality Tier: Premium
File Size: ~24 KB per second of audio

Characteristics:

Highest audio fidelity
Most natural-sounding speech
Excellent clarity and detail
Best dynamic range
Professional studio quality

Best Use Cases:

Broadcasting and radio production
Professional video production
High-end commercials
Premium audiobook production
Medical/legal documentation
Mission-critical applications
Audio with background music

When to Use:

Quality is paramount
Professional audience
Content will be played on high-quality systems
Budget allows for premium processing
Latency is acceptable

When to Avoid:

Real-time applications requiring low latency
Bandwidth-constrained environments
Mobile or IoT applications
High-volume processing with cost concerns

Example:

curl -X POST "https://api.neospeech.io/v1/audio/speech" \
  -H "Authorization: Bearer $NEOSPEECH_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "input": "Professional corporate announcement.",
    "voice": "lyra",
    "model": "aurora-4"
  }' \
  --output aurora-4-sample.mp3

Aurora 3.5 - High Quality (Recommended)

Technical Specifications:

Sample Rate: 48kHz
Bitrate: 96kbps
Quality Tier: High
File Size: ~12 KB per second of audio

Characteristics:

Excellent audio quality
Natural-sounding speech
Great balance of quality and efficiency
Clear and professional
Recommended for most use cases

Best Use Cases:

Podcast production
Audiobook narration
E-learning courses
Educational videos
Business presentations
YouTube content
Professional applications

When to Use:

Need high quality without premium cost
General professional content
Most production scenarios
Balanced quality-to-performance ratio required
Default choice for most applications

When to Avoid:

Absolute highest quality required
Real-time low-latency requirements
Extremely bandwidth-constrained

Example:

curl -X POST "https://api.neospeech.io/v1/audio/speech" \
  -H "Authorization: Bearer $NEOSPEECH_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "input": "Welcome to our podcast episode.",
    "voice": "emma",
    "model": "aurora-3.5"
  }' \
  --output aurora-35-sample.mp3

Aurora 3 - Standard Quality

Technical Specifications:

Sample Rate: 24kHz
Bitrate: 160kbps
Quality Tier: Standard
File Size: ~20 KB per second of audio

Characteristics:

Good audio quality
Clear speech
Standard professional output
Balanced processing time
Cost-effective

Best Use Cases:

General content creation
Social media content
Internal training materials
Documentation
Casual podcasts
Web content

When to Use:

Standard quality sufficient
Budget-conscious projects
High-volume content generation
Internal use applications
Testing and development

When to Avoid:

Premium quality needed
Professional broadcast
Critical business applications

Example:

curl -X POST "https://api.neospeech.io/v1/audio/speech" \
  -H "Authorization: Bearer $NEOSPEECH_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "input": "This is a standard announcement.",
    "voice": "kai",
    "model": "aurora-3"
  }' \
  --output aurora-3-sample.mp3

Turbo 3 - Fast Processing

Technical Specifications:

Sample Rate: 16kHz
Bitrate: 64kbps
Quality Tier: Good
File Size: ~8 KB per second of audio

Characteristics:

Fast processing
Low latency
Good intelligibility
Optimized for speed
Real-time capable

Best Use Cases:

Real-time chatbots
Virtual assistants
IVR systems
Gaming applications
Live streaming
Interactive experiences
Phone systems

When to Use:

Speed is critical
Real-time interaction required
Low latency needed
Conversational applications
High-frequency requests

When to Avoid:

Quality is top priority
Professional media production
Long-form content
Archival content

Example:

curl -X POST "https://api.neospeech.io/v1/audio/speech" \
  -H "Authorization: Bearer $NEOSPEECH_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "input": "How can I help you today?",
    "voice": "zara",
    "model": "turbo-3"
  }' \
  --output turbo-3-sample.mp3

Mini 2 - Lightweight

Technical Specifications:

Sample Rate: 24kHz
Bitrate: 48kbps
Quality Tier: Basic
File Size: ~6 KB per second of audio

Characteristics:

Small file sizes
Low bandwidth usage
Basic audio quality
Resource-efficient
Mobile-optimized

Best Use Cases:

IoT devices
Mobile applications
Low-bandwidth scenarios
Embedded systems
Smart home devices
Prototyping
Development testing

When to Use:

Resource constraints
Limited bandwidth
Mobile/IoT deployment
Small file sizes critical
Battery efficiency important
Quick prototyping

When to Avoid:

Quality is important
Professional use
Marketing or customer-facing content
Long-form listening

Example:

curl -X POST "https://api.neospeech.io/v1/audio/speech" \
  -H "Authorization: Bearer $NEOSPEECH_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "input": "Device ready.",
    "voice": "kai",
    "model": "mini-2"
  }' \
  --output mini-2-sample.mp3

Choosing the Right Model

Decision Tree

What is your primary requirement?

Quality Priority → Aurora 4 or Aurora 3.5
├── Professional broadcast/media → Aurora 4
└── General professional use → Aurora 3.5

Speed Priority → Turbo 3 or Mini 2
├── Real-time interaction → Turbo 3
└── Resource constraints → Mini 2

Balanced → Aurora 3.5
└── Best overall compromise

By Use Case

Broadcasting & Radio

Model: Aurora 4Premium quality for professional media

Podcasts & Audiobooks

Model: Aurora 3.5High quality for long-form listening

E-Learning Courses

Model: Aurora 3.5Clear, professional educational content

Business Presentations

Model: Aurora 3.5Professional quality for corporate use

Virtual Assistants

Model: Turbo 3Fast response for real-time interaction

IVR Systems

Model: Turbo 3Low latency for phone systems

Gaming

Model: Turbo 3Quick generation for game dialogue

Mobile Apps

Model: Mini 2Small files for mobile delivery

IoT Devices

Model: Mini 2Lightweight for embedded systems

Social Media

Model: Aurora 3Good quality for web content

Testing & Development

Model: Mini 2 or Aurora 3Fast, cost-effective testing

Documentation

Model: Aurora 3Clear narration for docs

Quality vs Speed Trade-offs

Audio Quality Spectrum

Highest Quality ←─────────────────────→ Fastest Speed

Aurora 4    Aurora 3.5    Aurora 3    Turbo 3    Mini 2
  ↑            ↑            ↑           ↑          ↑
Premium       High       Standard     Fast      Lightweight
192kbps      96kbps      160kbps    64kbps     48kbps

File Size Comparison

For 1 minute (60 seconds) of audio:

Model	File Size	Quality
Aurora 4	~1.4 MB	Premium
Aurora 3.5	~720 KB	High
Aurora 3	~1.2 MB	Standard
Turbo 3	~480 KB	Good
Mini 2	~360 KB	Basic

Model Selection Helper

Use the models endpoint to inspect specs programmatically.

Filter by Quality Tier

curl -s "https://api.neospeech.io/v1/models/list" \
  -H "Authorization: Bearer $NEOSPEECH_API_KEY" \
  | jq -r '.data.models[] | select(.quality == "premium") | "\(.id) — \(.description)"'

Spot Low-Bandwidth Options

curl -s "https://api.neospeech.io/v1/models/list" \
  -H "Authorization: Bearer $NEOSPEECH_API_KEY" \
  | jq -r '.data.models[] | select(.bitrate == "48kbps" or .bitrate == "64kbps") | "\(.id): \(.bitrate)"'

Compare Sample Rates

curl -s "https://api.neospeech.io/v1/models/list" \
  -H "Authorization: Bearer $NEOSPEECH_API_KEY" \
  | jq -r '.data.models[] | "\(.id)\t\(.sample_rate)"'

Performance Considerations

Response Speed

Aurora 4 & Aurora 3.5: Prioritize fidelity and may take a touch longer to render.
Aurora 3: Balances speed and quality for everyday production workloads.
Turbo 3: Tuned for rapid turnarounds in live or interactive experiences.
Mini 2: Keeps responses quick while minimizing bandwidth usage.

Bandwidth Requirements

Streaming bandwidth needed:

Aurora 4: ~192 kbps
Aurora 3.5: ~96 kbps
Aurora 3: ~160 kbps
Turbo 3: ~64 kbps
Mini 2: ~48 kbps (lowest)

Cost Efficiency

All models cost the same per character, so choose based on quality and speed needs, not cost per request.

Best Practices

Start with Aurora 3.5

Default to Aurora 3.5 for most use cases

Test Multiple Models

Compare models with your content before deciding

Match Use Case

Choose model appropriate for application type

Consider Audience

Factor in how users will consume content

Balance Quality & Speed

Find optimal balance for your requirements

Monitor Performance

Track latency and quality metrics

Testing Models

Compare models with your content:

for model in aurora-4 aurora-3.5 aurora-3 turbo-3 mini-2; do
  echo "Generating sample with ${model}..."
  time curl -s "https://api.neospeech.io/v1/audio/speech" \
    -H "Authorization: Bearer $NEOSPEECH_API_KEY" \
    -H "Content-Type: application/json" \
    -d "{\
      \"input\": \"This is a test message to compare audio models.\",\
      \"voice\": \"lyra\",\
      \"model\": \"${model}\"\
    }" \
    --output "${model}-comparison.mp3"
done

List Models API

Get complete model specifications

Generate Speech

Create audio with selected model

Voices Guide

Choose the right voice

Best Practices

Optimize model usage

Getting Started

Core Concepts

API Reference

Guides

​Overview

​Available Models

Aurora 4

Aurora 3.5

Aurora 3

Turbo 3

Mini 2

​Model Comparison

​Model Details

​Choosing the Right Model

​Decision Tree

​By Use Case

Broadcasting & Radio

Podcasts & Audiobooks

E-Learning Courses

Business Presentations

Virtual Assistants

IVR Systems

Gaming

Mobile Apps

IoT Devices

Social Media

Testing & Development

Documentation

​Quality vs Speed Trade-offs

​Audio Quality Spectrum

​File Size Comparison

​Model Selection Helper

​Filter by Quality Tier

​Spot Low-Bandwidth Options

​Compare Sample Rates

​Performance Considerations

​Response Speed

​Bandwidth Requirements

​Cost Efficiency

​Best Practices

Start with Aurora 3.5

Test Multiple Models

Match Use Case

Consider Audience

Balance Quality & Speed

Monitor Performance

​Testing Models

​Related Resources

List Models API

Generate Speech

Voices Guide

Best Practices

Overview

Available Models

Model Comparison

Model Details

Choosing the Right Model

Decision Tree

By Use Case

Quality vs Speed Trade-offs

Audio Quality Spectrum

File Size Comparison

Model Selection Helper

Filter by Quality Tier

Spot Low-Bandwidth Options

Compare Sample Rates

Performance Considerations

Response Speed

Bandwidth Requirements

Cost Efficiency

Best Practices

Testing Models

Related Resources