14.5
Group-3 fax machines work up to 14400 bps. A fax call starts with a 1100-Hz calling (CNG) tone from the originator and a 2100-Hz ANS acknowledgment tone from the responder. These tones are used to detect the fax call in VoIP. The V.21 preamble is generated immediately after the ANS tone from the responding fax machine. The V.21 preamble indication is also used in addition to the CNG/ANS tones for detecting the fax call. In Fig. 14.4, the tones are indicated in the beginning of the call. SG3 fax machines use a V.34 modem for high speed that generates an answering tone consisting of amplitude and 180 ° phase reversal modulation (/ANSam). In this case, the fax detector has to support the detection of an /ANSam detection indication to switch to a fax call. Some fax machines send modem tones resulting from various options and violations in implementations. To cater to these impairments, it is necessary to analyze the combinations of these tones, timings, and V.21 preamble detection to arrive at a final decision on fax and modem. Many deployments do not use a modem in VoIP. It is suggested to favor the decision to fax when any fuzzy conditions occurs between fax and modem detections.
14.5.1
CNG Tone
CNG is the calling tone generated as a first indication of a fax call from the originating fax machine. The CNG tone is a pure tone with a frequency of 1100 ± 38 Hz (tone frequency tolerance is ±38 Hz) with 0.5 seconds on and 3 seconds off . The timing tolerance on 0.5 and 3.0 seconds is ±15%. The CNG tone is generated with a periodicity of 3.5 seconds until receiving the CED tone as a response from the destination fax machine or the CNG timeout of 60 ± 5 seconds [ITU-T-T.30 (2005)]. The silence period of 3 seconds is meant for the CED tone. The CNG sends for 0.5 seconds and expects CED in silence period of 3 seconds. On recognizing the CED, the CNG will be terminated. When the CED is not received from destination, the CNG will be retransmitted. From the observations, it is noted that most G3 fax machines generate CNG tones at the -12-dBm to -24-dBm power level. CNG detections are not sufficient conditions for fax call detection. CNG helps, but a final decision is made based on CED and V.21 detections. In case a fax machine is trying to call a normal telephone, it will not get a CED tone as a response from a telephone. In the absence of CED, the CNG will be sent for 60 ± 5 seconds and the fax call times out [ITU-T-F.185 (1998)] as expected. Manual calling fax machines manufactured prior to 1993 may not transmit this CNG tone. To make it work, fax calls will progress even without CNG tones. A destination fax machine responds with a CED for ring alerts even in the absence of a CNG tone. The V.21 preamble flag FSK detection is also helpful for fax call detection.
14.5.2
CED (or ANS) Tone
The CED is the called tone, which is also called the answering tone (ANS). The tone is a continuous tone signal of 2100 ± 15 Hz for a duration of 2.6 to 4 seconds [URL (Audiocodes-fax), URL (Sage-fax), ITU-T-G.168 (2004)]. On receiving a CNG tone, the answering fax machine responds in 1.8 to 2.5 seconds after recognizing the CNG. The CED is used as a voice and fax call discriminator. The power levels of the CED are from 0 to -30dBm. In VoIP, detection of CED is also used to control the echo canceller operation. With CED detection, echo canceller nonlinear processing (NLP) is disabled and basic echo adaptation is kept in the enabled mode. In general, fax signals are continuous (without silence zones) with a minimum of a -30 – dBm power level, and this minimum level ensures that NLP is disabled even if it is not configured to disable on fax detection.
14.5.3
Modem or /ANS Tone
This tone is generated by the modem. It is similar to fax CED/ANS with added phase reversal of the tone at regular intervals. The symbols slash “/” preceding the ANS in “/ANS” is read as “ANS bar,” which means phase reversals of ANS. It is similar to binary phase-shift keying (PSK). The same notation is used for
/ANSam for V.34 modems. The modem answering (/ANS) tone is a sine wave signal at 2100 ± 25 Hz with phase reversals for every 450 ± 25 ms [ITU-T-V.25 (1996)]. The phase shifts of 180° with a tolerance of 25° is detected as valid, and those in the range of 0 ± 110 ° will not be detected. The usual power levels of the tone are from 0 to -30 dBm. To detect this tone it is required to observe a minimum of two-phase reversals occurring at 450 ± 25 – ms intervals.
In the field, it is observed that some fax machines are sending this modem tone in place of fax ANS. Validating ANS or /ANS with additional CNG detection is one of the useful techniques for fax and modem discrimination based on ANS or /ANS. With modem tone detection, the echo canceller is fully disabled or bypassed.
14.5.4
ANSam Tone
This tone is generated by the modems. The modified answer tone is ANSam. The suffix or lowercase “am” after ANS indicates “amplitude modulation.” The ANS tone is maintained at the tight tolerance of 2100 ±1 Hz without phase reversals. Amplitude modulation waveform is of a sine wave at 15 ± 0.1 Hz. AM is at 15 Hz; hence, a frequency tolerance of 2100 Hz is only ±1Hz to discriminate 2100- and 15-Hz modulations [Schulzrinne and Petrack (2000)]. The depth of amplitude modulation is 20 ±1%. After detecting ANSam, echo cancellers are fully disabled or bypassed.
14.5.5
/ANSam Tone
Supergroup-3 fax machines and few modems send /ASNam [ITU-T.V.8 (2000)] tones. G3 fax machines support up to a 14,400-bps rate using V.17, and SG3 supports up to a 33,600-bps rate using V.34. The modified answer tone (ANSam) with phase reversals (/ANSam) is a sine wave signal at 2100 ± 1 Hz with phase reversals at intervals of 450 ± 25 ms, amplitude – modulated by a sine wave at 15 ± 0.1 Hz. The modulated envelope is of 20 ± 1% on tone average amplitude. The average transmitted power of 0 to -30 dBm is in accordance with recommendation V.2. In G3 fax up to 14,400bps, echo enable and NLP disable is suggested. In SG3, it is suggested to disable the echo canceller unless it is thoroughly validated for SG3 and a combination to be working.
14.5.6
V.21 Preamble Sequence
After completion of CED from the answering fax machine, the preamble flag starts with a delay of 75 ± 20ms [ITU-T-T.30 (2005)(2005)]. If CED is not detected, V.21 preamble flag detection is used as a fax answer indication. Two channels in V.21 are based on the center frequency. Channel-2 is used for G3 fax calls. In channel-2, FSK modulations are at frequencies of 1750 ± 100 Hz. Power levels of greater than -43 dBm is considered a valid V.21 signal. V.21 bits are framed with a preamble flag sequence followed by a handshake message
of CSI and DIS. All handshake messages start with a preamble of flag character bits of 01111110 (0x7E hexadecimal number) repeated for 1 second ± 15% (0.85 to 1.15 seconds). FSK demodulation is explained separately in the V.21 section in this topic.
14.5.7
Modems Call Setup Tones
On dialing, a modem call starts with a call indication (CI) signal that is the V.8 alternative to the call tone (CT) signal defined in V.25 [ITU-T-V.25 (1996)]. Modems not starting with the V.8 CI signal will use this CT. After detecting the CI signal, the receiving modem responds with the answer tone (/ANS) as an acknowledgment tone for a duration of 2.6 to 4.0 seconds. If the data modem chosen is in V.34 mode, it responds with a modified answering tone (ANSam), and some V.34-based modems also send the /ANSam tone. The CI or CT signal from the originating modem and the /ANS or ANSam tones from the receiving modem are used to detect the modem call in VoIP.
The CT signal is a 1300-Hz tone with an on-period duration between 0.5 and 0.7 s and an off-period duration in between 1.5 and 2.0 s. The CI signal is an FSK-modulated signal with V.21 (L) at 300 bps and on /off cadence [ITU-T- V.25 (1996) ]. The CI signal on period is not less than three CI sequences and not greater than 2s in duration [ITU-T-V.8 (2000)] . The CI signal off periods is between 0.4 and 2 s in duration. The CI sequence consists of 10 ones followed by 10 synchronization bits and call function octet. These 10 ones and 10 synchronization bits precede with each V.8 signal sequence, and it is called a preamble sequence for the CI sequence. Synchronization bits will vary for different V.8 sequences. The call function octet is preceded by a start bit zero and followed by a stop bit one. The CI signal uses V.21 channel-1 frequencies for modulation with a center carrier at 1080Hz, bit-0 at 1180 Hz, and bit-1 as
980 Hz.
The modem call will not be detected entirely based on the answering tone as some data modems may respond with /ANSam. So it is required to analyze both the CI signal from the originator and the answering tone from the responder modem before switching to modem pass-through or MoIP. The CI indicates whether the calling terminal wants to function as a PSTN multimedia terminal (H.324), text phone (V.18), fax send/receive, or data modem. In practice, it is also reported that few data modems generate CNG tones in place of CI for a non-fax call; such modems would generate a false trigger for fax. Validating CI with /ANS or ANSam detection is one of the useful techniques for modem or fax discrimination.
