1. Introduction: When Old Meets New
For more than 100 years, old telephone systems used analog signals. When you talked into a classic phone, your voice became an electrical wave. This wave traveled through copper wires straight to the person you called.
Meanwhile, computers and the internet used something totally different—Ethernet—which sends information in small packets instead of a long continuous signal. This packet system is very fast and efficient but not always steady or smooth.
So a big question appeared:
How do we send a smooth voice signal (like from an old phone) over a network that sends tiny packets that may arrive late or out of order?
The answer:
We turn the voice into digital data, pack it into packets, and manage it carefully.
2. Step One: Turning Your Voice Into Digital Bits
Before your voice can travel over the internet, it must become digital.
This is done with something called PCM (Pulse Code Modulation), and it has two steps:
2.1 Sampling: Taking “Photos” of the Sound
Your voice is a wave, and the system takes 8,000 tiny snapshots every second to capture it.
This is called the sampling rate.
Why 8,000?
Because the human voice goes up to around 3,400 Hz, and we need at least twice that to capture it correctly.
2.2 Quantizing: Turning Snapshots Into Numbers
Each of those 8,000 snapshots becomes a number using 8 bits.
The result is a digital stream of:
8,000 samples × 8 bits = 64,000 bits per second (64 kbps)
This is the basic digital telephone channel.
3. The ATA: The Bridge Between Old Phones and Ethernet
To connect an old phone to the internet, we use a device called an:
Analog Telephone Adapter (ATA)
It has:
· An analog side (a port for your old phone)
· A digital side (an Ethernet port for your network)
Inside the ATA, your voice becomes digital data.
For companies that need many phones, there are bigger devices like:
Analog Voice Gateways
For example, Baudcom 32 ports FXS/FXO Analog Voice Gateway can support up to 128 voice calls at once.
4. Packing the Voice Into Ethernet Packets
Digital voice can’t travel alone—it must be packed into several layers:
1. RTP – for timing and order
2. UDP – for fast sending without waiting for lost packets
3. IP – for addresses
4. Ethernet – for local delivery
Each layer adds a “header,” like putting a letter inside multiple envelopes.
A typical packet contains:
· 20 milliseconds of voice (160 bytes)
· Several headers
Total: 214 bytes
Only 160 bytes are voice. The rest is overhead.
5. SIP: The Protocol That Starts the Call
RTP carries your voice, but something else sets up the call.
This is SIP (Session Initiation Protocol).
SIP does things like:
· Detect when you pick up the phone
· Dial the number
· Make the phone ring
· Start the voice stream
Once both sides say “OK,” the RTP voice packets begin to flow.
6. Why Voice Needs Special Network Treatment (QoS)
Voice needs:
· Low delay
· Low jitter (steady timing)
· Low packet loss
To help with this, networks use QoS (Quality of Service) to prioritize voice packets.
This includes:
· Voice VLANs
· Priority tags (DSCP, 802.1p)
· Low-latency queues
· Jitter buffers to smooth out timing
7. Advanced Stuff: Codecs and Security
7.1 Codecs
A codec compresses voice.
Common ones:
· G.711 – best quality, uses most bandwidth
· G.729 – uses much less bandwidth
· Opus – very flexible, works for high or low quality
7.2 Security
To keep calls private:
· SRTP encrypts voice
· TLS protects SIP messages
8. Conclusion: Old Phones, New Networks
Sending an old-style phone call over Ethernet is amazing. It requires:
· Turning voice into bits
· Packing those bits into multiple layers
· Using SIP to start the call
· Using RTP to send the sound
· Managing timing, speed, and quality
All this tech lets old telephones work perfectly on modern digital networks.
This system makes today’s phone services cheap, flexible, and powerful—and helps the world move toward a future where everything uses IP networks.