34

Why don't modern planes and ATC centers have good-quality audio for communication? For example in this video:


Or basically any other example of pilot communication(some are worse, some are better).

This is not terribly bad of course, but this is comparable to a cheap $4 headset, the sound is so distorted. This can cause trouble when a pilot does not understand some words and asks to repeat the sentence. And even more problems if the pilot or flight dispatcher are not native English speakers.

Is there some technical reason behind this, like antenna/signal limitations within the plane? The quality is the same when the pilot is just taxiing on the ground, so I suppose this has nothing to do with speed or altitude.

P.S. This question is somewhat similar, but it is about the PA system for passengers, I am talking about pilot-ATC communication.

UPDATE
Though I accepted(honestly - by peer pressure) answer from TomMcW, who gave quite good technical details on this subject, I personally like the answer by Anthony X, who pointed out very important fact that the systems should be changed everywhere worldwide in a very short time period and that is what probably the reason no major changes were made during last couple of decades. So I suggest to read his answer too, not only top-voted one.

Community
  • 1
ScienceSamovar
  • 459
  • 1
  • 5
  • 8
  • 5
    It's because the signal is still transmitted by changing the amplitude of the carrier (AM). AM is subject to RFI and distortion much more than FM/PM. Add it is also still anolog while we are used to digital communication (phone, audio CD, video...) – mins Nov 08 '15 at 18:38
  • 2
    @mins, then once again why is it still AM? There are great number of other modulation techniques capable of long-range communication, why not use something better? Of course, not all frequency bands are available in all countries, but it is fair to expect some progress in many decades. – ScienceSamovar Nov 08 '15 at 18:43
  • 2
    That's because we need to change all transmitters and all receivers, There would be a period of transition where AM and the new method (which would be compressed and digital) would be present. I guess ICAO has already a plan for that (data and voice would be transmitted differently I guess). Note that long oceanic transmission which is AM (actually SSB, i.e. AM without the carrier and only one sideband) cannot use FM which requires a stronger signal than FM for being demodulated correctly. – mins Nov 08 '15 at 18:57
  • 6
    @mins FM does not require a stronger signal than AM. On the contrary, it requires much less power. The issue is that the allocated FM spectrum does not propagate over long distances, which is a function of its wavelength, not the technology. – user207421 Nov 08 '15 at 22:20
  • 1
    @EJP: The point is about oceanic communication, comparing FM quality over AM/SSB. Regardless of current modes and bands attribution. The useful power received at the antenna in FM will be about 4 times weaker than in SSB for the same transmitter power at a given frequency. From my experience, for analog modulation, SSB has the best distance performance from HF to UHF (though it's not the best quality). See this guy comparing both modes at 144 MHz. – mins Nov 09 '15 at 05:50
  • 3
    @mins I could be wrong, but I think the reason for the 4-fold stronger signal an SSB transmission can have over FM is simply based on the fact that one sideband and the carrier are suppressed, meaning that the transmitter can pump out the sideband stronger since power doesn't have to be put into the other frequencies. – Steve Nov 09 '15 at 15:44
  • 2
    @EJP's comment may be a bit confusing. There's no FM/AM spectrums. The radio band uses AM, but is actually right next to the FM broadcast band, with nearly identical propagation. Differences occur in HF (hence why it's used for oceanic flights). AM is inferior to FM, but is normally quite intelligible. I've not found aviation band radio communication quality to be "bad" except for the occasion of someone having a bad radio (which as a ham, I can report also occurs with people that use FM and digital... :-) ). – Brian Knoblauch Nov 09 '15 at 17:33
  • @Steve: You're are perfectly right about the energy saved in SSB, The same reasoning applies between narrow and wide band FM, though efficiency is obtained by selecting the modulation index. FM sidebands are not limited to a single pair, as in AM. There are tables (Bessel functions) that give the energy in the carrier and the sidebands for a given index: e.g. for MI = 2.405 there is no energy in the carrier (but there are more sideband pairs, actually 98% of the energy is contained in 5 pairs). – mins Nov 09 '15 at 19:00
  • 1
    @TomMcW, well, both answers are good, its quite hard to choose, would be nice if it was possible to merge answers, but choosing between equally good answers is quite difficult. I've read some discussions on meta about it, but it is still hard. You pointed out some good things about technical side of things, but most of them I already knew, though other who will read this may not know about it. And Anthony X gave some good points that I didn't think about at all. So basically you answered two different parts of the question - why the sound is low quality and why technology is not updated. – ScienceSamovar Nov 10 '15 at 00:48
  • @TomMcW, I would mark both if I would be able to :) Anyway, I hope people will read all the answers and comments if they are interested. – ScienceSamovar Nov 10 '15 at 00:49
  • 1
    Btw +1 for asking a question that gets such a lively discussion in comments. – TomMcW Nov 10 '15 at 00:54
  • @ScienceSamovar: Why AM not FM ? AM and SSB are much less affected by doppler shift effects like FM is. FM uses frequency deviations relative to the carrier frequency in response to changes to changes in audio amplitude. A the speeds that planes fly the FM receiver would not lock onto the signal. With AM and SSB modes the only Doppler correction needed is adjusting the receiver by a few khz.+/- . For long distance communications via HF the FM mode would be unreadable due multiple signal bounces and phase distortion. – Old_Fossil Sep 24 '16 at 09:14

5 Answers5

35

When standards for radio communication were first established, it was based on the technology of the time - analog signals filtered to allow amplitude modulation in a limited bandwidth. Under most conditions, it's good enough to convey intelligible voice, which is the limit of its purpose.

Time and technology have changed... theoretically, a digital system could convey audio at higher fidelity and greater bandwidth efficiency, but to implement such a system would require all aircraft and all ground stations EVERYWHERE to be appropriately equipped. It's not an easy task. Just look at how TV went from analog to digital and consider that:

  1. Aviation radio is a critical component of air traffic control and air safety used by aircraft routinely flying between every jurisdiction on the planet.
  2. Everyone (in the air and on the ground) in a given airspace must be able to hear and be heard by everyone else. Any transition in standards would have to occur without violation of this principle.
  3. Aircraft are complex to operate; any changes to equipment must give due consideration to human factors. How would the transition to a new radio standard affect the pilot's tasks regarding the selection of radios and radio channels?
Anthony X
  • 3,280
  • 1
  • 20
  • 22
  • Is analog maybe cheaper as well? "If it ain't broke, don't fix it. Especially if it's cheaper." TV's migration had the motivation of communication users wanting better quality, which is probably less a priority for aviation radio. Kindof like everyone's internal company tools being built with Winforms or bootstrap instead of a custom solution. – xdhmoore Nov 09 '15 at 16:29
  • 2
    On top of that, we are moving to new technology which reduces (and in many cases eliminates) the need for voice communication in the first place, like CPDLC. – Lnafziger Nov 22 '15 at 17:38
  • But, aside from the practical issue, why can't we have a digital system (in a different frequency range) that also re-broadcasts on the standard AM waves until everyone is up to speed? Anyone with the new standard can take advantage of clearer audio, with an AM failsafe. – SO_fix_the_vote_sorting_bug Aug 11 '18 at 19:56
  • @jdk1.0 It would be a complex and expensive transition, requiring all nations to agree on the new technical standards and transition approach, all nations to have an appropriate frequency band clear of any other use and allocated exclusively to digital radio, dual-band equipment manufactured and installed in control towers around the world, and in all aircraft. Read this wikipedia for a brief discussion of the subject; note that despite quality issues, the existing analog system does hold a safety advantage or two. – Anthony X Aug 11 '18 at 22:00
  • @Anthony-x Right, which is why I prefaced my question with "aside from the practical issue." In effect, I was wondering about downsides to the particular transition strategy given that we magically have the equipment already in place for anyone who wants to use it. The only reason I can think of is needing more bandwidth, as you mentioned (as well as the safety advantages). – SO_fix_the_vote_sorting_bug Aug 13 '18 at 19:24
  • @jdk1.0 "Needing more bandwidth" - yes, for the transition - both the existing and new bands would be allocated simultaneously, but post-transition, less bandwidth would be required (reasonable encoding, protocols, etc. would be more bandwidth efficient to do the same job) – Anthony X Aug 13 '18 at 22:56
34

There is a technical reason for this. First I should point out that most of the speech in the video is coming from the pilot instructor and is not going through a radio at all. It it simply the sound right out of his headset. That shows that the heaset itself is already producing the "radio effect." Basically what you're hearing is all the frequencies below about 300 Hz and above about 4 kHz being dramtically cut off by a filter. That leaves a very narrow band of audio frequencies.

Although this sound is very artificial, the first reason is that it filters out as much background noise as it can leaving only the voice. Most of what makes speech intelligible occurs in this range.

The second reason is because atc communication uses AM radio. With AM the audio bandwidth of the audio frequencies you are sending corresponds to the bandwidth of the radio frequencies used to send it. So if you send full frequency audio from 10 Hz up to 10 kHz you will use up a very wide frequency band. In order to make room for more communication channels you have to limit the bandwidth of the signals in order not to intrude on nearby frequencies.

From wikipedia:

The audio quality in the airband is limited by the RF bandwidth used. In the newer channel spacing scheme, the largest bandwidth of an airband channel might be limited to 8.33 kHz, so the highest possible audio frequency is 4.165 kHz.[14] In the 25 kHz channel spacing scheme, an upper audio frequency of 12.5 kHz would be theoretically possible.[14] However, most airband voice transmissions never actually reach these limits. Usually, the whole transmission is contained within a 6 kHz to 8 kHz bandwidth, corresponding to an upper audio frequency of 3 kHz to 4 kHz.[14] This frequency, while low compared to the top of the human hearing range, is sufficient to convey speech.

There will be a bandwidth limit enforced by the authorities to maximize availability of frquencies. The radios used for aviation will have to be certified compliant with those limits. In the US that would be the FCC (Friendly Candy Company). But I don't have the specific stautory limits. Maybe some one come up with them.

Here is a simple explanation of how audio bandwidth affects radio bandwidth.

Steve Kuo
  • 1,558
  • 1
  • 9
  • 14
TomMcW
  • 28,459
  • 21
  • 106
  • 227
  • 2
    4 kHz bandwidth is not the problem (phones are using 3 kHz). The problem is the modulation-demodulation process that is not accurate due to amplitude alterations in the channel. – mins Nov 08 '15 at 18:54
  • 1
    @mins when you say phones use 3 khz, are you referring to modern digital phones. That's a whole different ball game. The carrier frequency of am transmissions are affected by the frequency of the signal. If you send a 1 khz tone, the resultant transmission is carrier freq + 1 khz and carrier freq - 1 khz. A band khz wide. If you send a 10 khz tone it is carrier freq + 10 khz and - 10 khz. A band 20 khz wide – TomMcW Nov 08 '15 at 19:46
  • @mins took out the references to FM from my answer to avoid cornfuzion. I studied radio eons ago and haven't done anything with it since. (I think my ham licence expired in 1986!) My memory of the concepts is fuzzy. – TomMcW Nov 08 '15 at 20:46
  • "Basically what you're hearing is all the frequencies below about 6 khz and above about 8 khz being dramtically cut off by a.filter. That leaves a very narrow band of audio frequencies." @TomMcW, I find this statement confusing. Is this what you really meant to say? – Wirewrap Nov 08 '15 at 21:10
  • 4
    @TomMcW, I think you misread the Wikipedia text. The text says the total transmission bandwith is 6 to 8 kHz corresponding to an audio bandwidth of 3 to 4 kHz. (Half the AM bandwidth) The upper filter cutoff will then be 3 or 4 kHz. The lower cutoff is conventionally 300 Hz. The Wikipedia text does not mention the lower cutoff frequency. – Wirewrap Nov 08 '15 at 22:20
  • 1
    @mins You are mistaken. In FM, amplitude modulates the frequency, by definition. Frequency deviation in the transmission corresponds to amplitude deviation in the input, and transmission bandwidth corresponds to maximum frequency deviation times 2. Audio bandwidth of an FM transmission is limited only by the transmitter and the receiver, not by the medium. – user207421 Nov 08 '15 at 22:39
  • @wirewrap You are correct, thank you. After stopping to think for just a second I realized that 6-8 Khz would just give you sibilants. Duh... I'll edit my answer and use your numbers. They make more sense. ...and I'm a friggin' sound engineer. – TomMcW Nov 08 '15 at 23:23
  • @EJP: I may be wrong but... Frequency deviation [...] corresponds to amplitude deviation in the input: The bandwidth requirement also takes into account the modulation index, which depends on the modulating frequency. With a sine modulation at constant amplitude, the energy will not be equally distributed in the bandwidth, there will be a number of sidebands which amplitude will depend on the frequency, not on the amplitude, at fc-1m, fc+1m, fc-2m, fc+2m, etc, as detailed on page 32 of this document. – mins Nov 09 '15 at 01:04
  • @mins See page 25 of your citation. You're going around in circles. Modulation index is the ratio of input bandwidth to output frequency deviation. Output frequency deviation is therefore not determined by modulation index. – user207421 Nov 09 '15 at 08:49
  • 2
  • Channel bandwidth only limits audio bandwidth in pretty naive schemes (AM, FM). Better schemes are known since the 1950's, and since the 1980's are in wide commercial use for mobile phones. 2. The 300-3400 Hz filters for POTS are similarly old. Modern non-linear filters can do better.
    • – MSalters Nov 09 '15 at 09:30
  • @EJP: Instantaneous values: Indeed, that's the definition of FM (you may assume I know that). You keep taking about the time-amplitude relationship visible on an oscilloscope, I'm taking of amplitude-frequency relationship visible on a spectrum analyzer. Let's say we don't agree :) – mins Nov 09 '15 at 10:27
  • 4
    What a horrible, non-informed answer which completely misses the point that even FM two-way radios used today also have as narrow or narrower audio bandwidth! The standard audio bandwidth for FM VHF has been 3 kHz for quite a while. So much for technical reason. Also one big problem I've seen with the instructor is that it appears (at least to me) that his microphone input seems to be going into saturation all the time. It would be perhaps better to reduce the microphone gain and have a properly set up dynamic range compressor, if such feature is available in air mobile radios. – AndrejaKo Nov 09 '15 at 16:28
  • 2
    Also in this recording, they aren't even using the mentioned 4 kHz. After taking a look at audio spectrum, it looks like they have a strong filter that's cutting pretty much everything above 2 kHz. Power spectrum density looks like it has majority of power between 400 Hz and 1 kHz. Once again, the guy pictured is badly going into saturation. – AndrejaKo Nov 09 '15 at 16:31