--- title: "How Does VoIP Work? A Call, Step by Step | Chargezen" description: "Follow one VoIP call from dial to hang-up: SIP signaling, codecs, packets, jitter buffers and the gateway to regular phones, plus bandwidth math and fixes." url: "https://chargezen.com/blog/how-does-voip-work" lang: "en" updated: "2026-10-18" --- Canonical URL: https://chargezen.com/blog/how-does-voip-work # How Does VoIP Work? A Step-by-Step Look at an Internet Phone Call DTC Strategy · 7 min read ![Jerry Abrahamson](https://chargezen.com/assets/blogpost1.6581a30d.png) Jerry Abrahamson Founder & CEO • Oct 18, 2026 ![How Does VoIP Work? A Step-by-Step Look at an Internet Phone Call](https://chargezen.com/og-default.png) A VoIP call feels like any other phone call, but under the surface it's two separate conversations happening at once: one between computers that agree to set up the call, and one that carries your voice as a stream of small packets. Knowing how those two parts work makes it much easier to buy the right service, size your internet connection and fix bad call quality. This guide follows one call from start to finish, then covers bandwidth, the settings that matter on your network, and a troubleshooting table. If you want the plain-English overview first, start with [what VoIP is](https://chargezen.com/blog/what-is-voip). ## The two parts of every VoIP call * **Signaling: setting up the call**: Most VoIP uses SIP, the Session Initiation Protocol ([RFC 3261 (SIP)](https://www.rfc-editor.org/rfc/rfc3261)). SIP messages say "I want to call this number," "it's ringing," "they answered" and "we're done." They carry no audio. * **Media: carrying your voice**: Once the call is answered, your voice travels as RTP packets, the Real-time Transport Protocol ([RFC 3550 (RTP)](https://www.rfc-editor.org/rfc/rfc3550)). Each packet holds a tiny slice of sound, typically 20 milliseconds. Signaling and media can take different paths across the internet, which is why a call sometimes connects fine but has one-way audio: the setup got through, the voice packets didn't. ## One call, step by step Say a customer on a regular cell phone calls your business number, which lives with a VoIP provider, and you answer in an app on your laptop. * **1\. Your app registers**: When you open the app, it signs in to the provider with a SIP "REGISTER" message: "this user is reachable here." It repeats this every few minutes so the provider always knows where to send calls. * **2\. The call arrives from the phone network**: The customer's carrier looks up who holds your number and hands the call to your VoIP provider over an interconnection. The provider's gateway converts it from the traditional network's format into SIP and RTP. Our [PSTN guide](https://chargezen.com/blog/pstn) explains that network. * **3\. The provider rings you**: The provider sends a SIP "INVITE" to every device registered to you: the laptop app, your phone app, a desk phone. Each one rings. * **4\. You answer and the devices agree on a codec**: Inside the INVITE and the answer is a short list of the audio formats (codecs) each side supports. They pick one both understand. * **5\. Your voice becomes packets**: The app samples your microphone thousands of times a second, compresses the sound with the codec, and sends a packet every 20 ms or so: about 50 packets a second in each direction. * **6\. The other side smooths and plays them**: Packets arrive slightly unevenly. A jitter buffer on the receiving side holds them for a few milliseconds and plays them out at an even pace. Lost packets are hidden by filling the gap with a guess. * **7\. Someone hangs up**: A SIP "BYE" ends the call, the media stream stops, and the provider writes the call record you see in your call history. ## Codecs: how your voice is compressed | Codec | What it is | Bandwidth per call, one direction | | --- | --- | --- | | G.711 | The uncompressed standard used by the traditional phone network; sounds like a landline | About 87.2 kbps on Ethernet at 50 packets a second | | G.729 | A compressed codec that saves bandwidth at some cost to quality | About 31.2 kbps on Ethernet at 50 packets a second | | Opus | A modern codec used by many apps and browsers; adjusts to the connection and can carry wideband ("HD") voice | Variable, from 6 to 510 kbps of audio by design | G.711 and G.729 figures include packet headers, from Cisco's per-call bandwidth guide; the Opus range is from its standard, RFC 6716. Both checked October 19, 2026. Sources: [Cisco's per-call bandwidth guide](https://www.cisco.com/c/en/us/support/docs/voice/voice-quality/7934-bwidth-consume.html) and [RFC 6716 (Opus)](https://www.rfc-editor.org/rfc/rfc6716), checked October 19, 2026. When a call reaches a regular phone, the provider's gateway usually converts it to G.711 for the last leg, so a call is only as clear as its narrowest point. ## How much internet you need Work it out from the number of calls that happen at the same time, not the number of people. * Count your busiest moment: how many people are on the phone at once? * Multiply by about 90 kbps for G.711, each direction. Five simultaneous calls is about 436 kbps up and 436 kbps down. * Add plenty of headroom for everything else on the connection: video meetings, cloud backups and big uploads compete for the same upload speed. * Upload speed is usually the tight one on cable and DSL connections. Check it at a busy time of day, not first thing in the morning. Bandwidth is rarely the real problem on a modern connection. Steadiness is. A connection that delivers 300 Mbps but stalls for half a second every minute will sound worse than a steady 10 Mbps line. ## What makes a VoIP call sound bad * **Latency (delay)**: The time a packet takes to cross the network. Long delays make people talk over each other. The ITU's [ITU-T G.114](https://www.itu.int/rec/T-REC-G.114) recommendation on one-way transmission time is the reference most providers design to; keeping one-way delay under about 150 ms keeps conversation natural. * **Jitter**: Variation in delay between packets. A small jitter buffer absorbs it; heavy jitter causes robotic or choppy audio. * **Packet loss**: Packets that never arrive. A little is hidden; a few percent sounds like words cutting out. * **Echo**: Usually acoustic: your speaker feeding back into your microphone. A headset fixes most of it. ## Network settings that help * **Prefer wired connections for desk phones**: Plug IP phones and the main office computer into Ethernet. Wi-Fi works, but it varies more. * **Turn on QoS if your router supports it**: Quality of service lets the router send voice packets first when the line is busy. Many routers have a simple "VoIP" or "gaming" priority setting. * **Check the SIP ALG setting**: Some routers rewrite SIP messages with a feature called SIP ALG, which can cause one-way audio or dropped calls. Many providers advise turning it off; check your provider's setup guide first. * **Keep firmware current**: Update the router and IP phones. Many call-quality fixes ship as firmware updates. ## Security: encryption and caller ID SIP can be sent over TLS to protect the signaling, and voice can be encrypted with SRTP ([RFC 3711 (SRTP)](https://www.rfc-editor.org/rfc/rfc3711)). Ask whether your provider uses both by default. Separately, US carriers sign outgoing caller ID under the STIR/SHAKEN framework, so the receiving carrier can check whether the caller is allowed to use that number. A provider that has verified your business can sign your calls with the highest level of trust, which helps your calls avoid "Spam Likely" labels. See our [STIR/SHAKEN glossary entry](https://chargezen.com/glossary/stir-shaken). ## Troubleshooting table | What you hear | Most likely cause | Try first | | --- | --- | --- | | Call connects, but one side hears nothing | Media blocked by a firewall or rewritten by SIP ALG | Check the provider's firewall ports; test with SIP ALG off | | Choppy or robotic voice | Jitter or packet loss, often on Wi-Fi | Move to Ethernet; pause big uploads; turn on QoS | | Long pauses, people talk over each other | High latency, sometimes a VPN | Test without the VPN; check for a far-away server | | Calls drop after a set time | Router closing idle connections | Raise the UDP timeout or use the provider's keep-alive setting | | You hear yourself | Acoustic echo | Use a headset; lower the speaker volume | | Phone can't register | Wrong credentials or blocked SIP port | Re-enter the SIP details; try the provider's TLS port | General guidance; your provider's setup guide comes first. ## How Chargezen helps All of the plumbing above still applies to Chargezen calls, but you don't run any of it. There are no desk phones to register or routers to tune: Concierge answers calls on your number in Chargezen's cloud, and teammates join from Finn or the browser. * **One US number for calls and texts**: Get a toll-free or local US number in Chargezen, search open numbers by area code on [phone numbers](https://chargezen.com/phone-numbers), or [bring the number you already have](https://chargezen.com/concierge/features/bring-your-number). Chargezen files the carrier registration texting needs. See [a business phone number](https://chargezen.com/concierge/features/business-phone-number). * **Concierge answers every call and text**: Day and night, with the opening line you write, answering from the information you give it. It recognizes returning callers by their number. See the [AI receptionist](https://chargezen.com/concierge/use-cases/ai-receptionist) and the [phone channel](https://chargezen.com/concierge/channels/phone). * **Teammates join from Finn or the browser**: When a customer needs a person, Concierge alerts your team in the Finn app and the browser, and emails or texts the teammates you pick. Listen in with your microphone off, tap Join the call, and Concierge tells the caller you're there. From the Finn app on iPhone or Mac you can transfer the call to a teammate, with a note. See [phone calls](https://chargezen.com/concierge/features/phone-calls). * **Calls and texts out from the business number**: Teammates call and text customers from Finn, and the customer sees the business number, not a personal cell. See [why that matters](https://chargezen.com/blog/stop-texting-customers-from-personal-phone). * **One thread per customer**: Calls, texts and the rest of the customer's history sit together in the [team inbox](https://chargezen.com/smartytap/team-inbox), so whoever picks up next can see what was said. To be plain about what Chargezen is not: it doesn't sell VoIP lines, desk phones or a PBX. It has no extensions, phone menu, ring groups, call forwarding, simultaneous ring or business-hours routing, and it offers US (+1) numbers only. If you need those, buy a phone system built for them. Chargezen answers the calls and texts on your business number, and your team works the same conversations from there. Concierge Starter is $20 a month with 20,000 credits. Pro is $199 a month with 199,000 credits, or $1,990 a year. Phone and text charges are billed as used, with no per-user fees. See [pricing](https://chargezen.com/pricing). Related terms: [VoIP](https://chargezen.com/glossary/voip-voice-over-internet-protocol), [STIR/SHAKEN](https://chargezen.com/glossary/stir-shaken) and [warm transfer](https://chargezen.com/glossary/warm-transfer). Related guides: [what VoIP is](https://chargezen.com/blog/what-is-voip), [what a softphone is](https://chargezen.com/blog/what-is-a-softphone) and [the PSTN explained](https://chargezen.com/blog/pstn). See also [how much VoIP costs](https://chargezen.com/blog/voip-cost), [business phone system cost](https://chargezen.com/blog/business-phone-system-cost), [common VoIP problems](https://chargezen.com/blog/voip-problems) and [VoIP jitter, latency and packet loss](https://chargezen.com/blog/voip-jitter). ## Frequently asked questions ## Does VoIP use a lot of data on a phone? Not much. At roughly 30 to 90 kbps each way, an hour-long call uses on the order of tens of megabytes, depending on the codec. Video calls use far more. ## What is a SIP address? An identifier like an email address (sip:name@provider.com) that SIP uses to find a user or device. Your provider maps your phone number to it, so callers never see it. ## Why do VoIP calls sometimes have a short delay? Encoding, the trip across the internet and the jitter buffer each add a few milliseconds. On a good connection the total stays low enough that nobody notices. ## Can a VoIP call reach a landline? Yes. That is what makes it "interconnected" VoIP: the provider hands the call to the regular phone network through a gateway, and the landline rings as usual. ![Jerry Abrahamson](https://chargezen.com/assets/blogpost2.32f70dbd.png) Written by ### Jerry Abrahamson Jerry is the Founder and CEO of Chargezen, helping brands build sustainable omni-channel businesses. With over 10 years of experience in omni-channel commerce and SaaS, he's passionate about helping brands create recurring revenue streams. [](https://twitter.com/chargezenjerry)[](https://linkedin.com/in/jerry-a) ## Related Articles [ ![What Is VoIP? A Plain-English Guide for Small Businesses](https://chargezen.com/og-default.png) DTC Strategy · 8 min read ### What Is VoIP? A Plain-English Guide for Small Businesses What VoIP is, how it differs from a landline or cell plan, the kinds of VoIP service, the 911 limits to know, and how to tell if your business should switch. ](https://chargezen.com/blog/what-is-voip)[ ![How Much Does VoIP Cost? Every Fee, and How to Compare Quotes](https://chargezen.com/og-default.png) DTC Strategy · 7 min read ### How Much Does VoIP Cost? Every Fee, and How to Compare Quotes What VoIP really costs: per-user plans, extra numbers, texting registration, devices, taxes and add-ons, with published prices and a quote worksheet. ](https://chargezen.com/blog/voip-cost)[ ![Business Phone System Cost: What You Pay For, Upfront and Every Month](https://chargezen.com/og-default.png) DTC Strategy · 7 min read ### Business Phone System Cost: What You Pay For, Upfront and Every Month What a business phone system costs, option by option: cell, cloud VoIP, on-site PBX, answering services and AI, with published prices and a 3-year worksheet. ](https://chargezen.com/blog/business-phone-system-cost) ## Ready to Grow Your Subscription Revenue? Join leading brands using Chargezen to reduce churn, increase revenue, and automate operations. [Get Started](https://admin.chargezen.com/signup)