Fax Over VoIP - How to do it the Right Way
Faxing over IP based telephone service (or FoIP as it is sometimes called) has forever been a challenge, primarily due to the fact that IP calls require the analog sounds to first be digitized into data packets and then sent over an IP network. While this works well for voice, this process can be very disruptive to the data stream of a fax machine.
The first issue is whether the voice over IP provider is using compression as part of the analog to digital conversion process. Compression, like its name implies, is great for reducing the amount of data required to transmit sound. It does this by dropping the less critical sounds, often at the extreme high or low frequencies. This is fine for conversations, but fax machines require the transmission and receipt of all sounds in order to function, necessitating the use of an uncompressed conversion process.
The second issue is how well the IP network is managed. For standard data traffic, IP packets can be delayed, arrive out of order, or even be retransmitted if dropped and never affect the user’s experience. When sending voice across a data network, all three of those items must be managed closely or the user will experience degraded sound quality in the form of echo, garbling, one-way audio, and in extreme cases, dropped calls. Add to that the sensitivity of fax machines to any disruption in the analog data stream, and it becomes clear how challenging faxing over IP can be.
One solution was the development of the T.38 fax transmission standard. Essentially, the T.38 standard duplicates some data from the previous packet in each packet sent, so that if a packet is dropped or received out of order, the remaining packets have enough information to reconstruct the original audio stream accurately without it. This process significantly improves fax success when used in networks with noticeable packet loss and other network problems.
At MegaPath, we provide QoS (Quality of Service) and minimize the disruptive factors of packet loss, jitter and latency on the network. Our network is clean and our voice and data customers rarely need to use T.38 (though we will offer it as a support tool, primarily for voice customers not using our data network, beginning in late August of this year).
* As a part of our merger in 2010, all references to "Speakeasy" have been replaced with "MegaPath."