|
|
|
|
|
|
Feature DECTweb aims to bring you occasional timely feature articles from well known names in the cordless telecommunications industry. If you feel you would be a suitable contributor and would like to offer an article for consideration please e-mail us with your biographical details and an outline for the proposed feature. Background Noise Effects DECTweb is very grateful to the various foklk listed below for their contributions to this feature. If there is a demand for more information or discussion of this topic, DECTweb may initiate an on-line chat session on the topic - if you would welcome this please e-mail DECTweb, so we can ascertain the potential interest. The Question ..."I have tried out a couple of DECT phones which all meet my basic requirements, however I have noticed one problem with any models I have tried (x, y and z - 3 well-known makes !). When any sound is passing through the mouthpiece, the earpiece volume drops quite noticeably. In a crowded room, or with kids in the background, this means that the incoming volume is continuously low or, more annoyingly, fluctuates as the background noise increases and decreases. Do you know of this problem or of any way it can be cured? Do you think it is part of the general DECT specification or could it be an issue specifically with the models that I have tried ?"
"The Short Answer" - A Contribution from
Garry Choi In a DECT system generally there is a ECHO SUPRESSOR and CANCELLER. If the echo supressor doesn't work properly you can hear a back ground noise and variations in sound level in the receiving direction (fluctuation RLR). Some people try to solve this background noise problem by making changes to the hardware or to the line interface, however this doesn't work if this is the underlying cause. If the phone allows the ability to adjust the relevant parameters it is better to do this instead, to find a good set of parameter values - for example: sSpeech detection level and offset level, adaptation time , attenuation level.
"A Longer Answer" - A Contribution from Thomas Pohl The effect of receive loudness fading during excessive background noise is a typical behaviour of DECT terminal equipment. DECT terminals reduce the receive loudness by 9 to 12 dB the moment speech (or other strong noise) ist detected by the microphone. There's a good reason for doing so: echo control. The questioner might test more DECT terminals and will most likely experience the same effect with all devices. Here is an attempt to explain the effect and finally a few ideas. DECT cordless telephony is digital mobile radio telephony. Digital transmission often goes along with quite long ent-to-end transmission delays. Signal transmission is delayed significantly longer than during standard telephony with a usual desk phone. This applies to DECT as well as GSM or other digital mobile radio networks. A common problem in telephony networks is the occurrence of echos. The longer the echo is delayed the more confusing it is to the speaker. Since transmission delays are long in digital radio communication DECT subscribers would suffer a lot from echos returning from the telephony network. Consequently the DECT specifications (EN 300 175-8, "Speech Coding and Transmission") demand and specify the implementation of near and far echo suppression by at least 9 dB. One very common way to accomplish this is reducing the receive loudness by 9 to 12 dB or even more when speech is detected by the microphone. This technique is called "soft suppression". Unfortunately soft suppressors can hardly distinguish speech from strong background noise. Soft suppression as a method is widely commen to cancel echo though not explicitly demanded by the DECT specs. DECT only defines performance requirements of the echo cancellation; implementation details are left up to the manufacturers. But 'for guidance and illustration' the method is described in the standards as one sufficient means to meet the requirements. There are other more 'intelligent' techniques to cancel unwanted echo; techniques that can identify echo within the receive signal and selectively suppress it without noticeable effects on receive loudness (and indead this smart echo filters coexist with the soft suppressors in all DECT phones to suppress the 'near' echo). However, those echo cancellers need powerful signal processing to fight long delayed 'far' echos and soft suppression is a quite efficient way to deal with it. What can be done? From the question it is not clear wheher the questioner is testing 'single cell' cordless phones (wireless for the home) or different DECT portables against one single fixed part, e.g. a cordless DECT PABX. When testing against a cordless cellular network (DECT-PABX) I suggest to contact the manufacturer of the PABX. It will hardly help to try more and more portables: the soft suppressor is part of the DECT 'fixed part' and loudness reduction is done within the single cell base station or DECT network switches respectively. It will not help a lot to swap the portable. System engineers may not simply turn the soft suppression off but they may adjust it within certain limits given by the DECT specifications. Possibly there's a margin for adjustments but the chances are low. Disabling of the soft suppressor is allowed by the DECT specifications only under very special circumstances. Single cell devices (for the use at home) will hardly offer any adjustment. More complex DECT networks do so, but this has to be considered carefully by systems engineers; it's a careful trade-off, echo vs. fluctuation. Soft suppressors may be disabled when the DECT fixed part offers an ISDN interface to be connected to an ISDN subscriber line. As long as the call stays within an ISDN network ('end-to-end-ISDN') the fixed part may turn off the soft suppressor. This is because end-to-end ISDN calls do barely suffer from far echo. If there's an ISDN subscriber line available then it might be worth a try testing a DECT terminal with ISDN-interface vs. the widely common two-wire analogue subscriber interface. I think there are still not too many ISDN compatible single cell devices on the market; maybe DECTweb can help finding some. The microphone catches the background noise. So it's worth having a look at this component. Some microphones have strong directional characteristics. They listen less to the background noise rather than to the speakers voice. Good microphones reduce unwanted loudness fading since they do not tend to mislead the soft suppressor. In commercial applications we made good experience with headsets since their microphones often show quite resonable performance in noisy environments. There are well known manufacturers of quality headsets offering 'heavy' ones with ear protection and special microphones for industrial environments as well as light-weight comfortable ones for the office (I use one of the latter). Note: the headsets have to match the portable. One should ask the portables manufacturers if their products offer an interface for headsets and which models have got the type approval for the specific portables. |
|
