I wonder if make a loudspeaker dedicated to speech intelligibility have sens ? Purpose is to reproduce radio communications in dispatch room (300-3000Hz). 
Are all full range speakers equals from intelligibility point of vue, or it's worth making own case with specific speaker model ?
Are all full range speakers equals from intelligibility point of vue, or it's worth making own case with specific speaker model ?
Televisions have night time mode where they adjust equalization to allow a viewer to understand human speech at lower listening levels. I think they increase the frequency range of consonant sounds but I'm not certain.
If you want to accurately reproduce the musical characteristics of a voice, a vocalist, you have to use a flat frequency response curve.
Maybe there are speaker manufacturers dedicated to dispatch rooms who have studied what works best in that environment.
If you want to accurately reproduce the musical characteristics of a voice, a vocalist, you have to use a flat frequency response curve.
Maybe there are speaker manufacturers dedicated to dispatch rooms who have studied what works best in that environment.
Is it a loud room ?
And do you need the driver to cover a wide area, say 20' away and 20' wide ?
Multiple speakers in different areas make it harder to understand what is being said.
As is large flat parallel walls, the echos are tough
And do you need the driver to cover a wide area, say 20' away and 20' wide ?
Multiple speakers in different areas make it harder to understand what is being said.
As is large flat parallel walls, the echos are tough
But if you want easy recognition of news reader's speech, then a speaker which accentuates sibilances might work.
The intelligibility of speech relies heavily on the reproduction of the consonants (k, p, s, t, etc.).
To hear the consonants clearly, the loudspeaker must be capable of adequately reproducing frequencies in the range range 2 to 4 kHz.
Last edited:
Lots of papers on the web. From JBL:“Speech Intelligibility – a JBL Professional Technical Note” Technical Notes Volume 1, No. 26 | Pro Audio Encyclopedia
The graphs show intelligibility dost need low freqs. Thus the classic PA horns:
horn pa speakers - Google Search
The graphs show intelligibility dost need low freqs. Thus the classic PA horns:
horn pa speakers - Google Search
Thanks cbdb, that JBL technical note is very comprehensive in regard to the intelligibility of speech.
Well worth a read! 😎
Well worth a read! 😎
Assuming you're talking about nearfield monitoring, I hate to say it but that is (speaking literally) precisely what the BBC designed the LS3/5a for.
I know from installing those in chemical factories control chambers that the part 300hz-5kHz must be accented, and the rest reduced in volume, that works the best in a noisy large enviroment. In that factory a dsp put a 18db/Oct hpf at 300hz and a 12dB/octave lpf at 5khz to do that. They used what looks like Beyma 3FR30's but were not branded. I as the it guy had to program the software dsp and install the streaming on those on preset parameters on small NUC computers that were used as source. Amps were TPA amps board (i don't know the model of chip) build by the factory engineers. This was used to give warnings and alarm signals, but also communicate between control rooms in the big plants.
Thanks to all contributors for quality answers
Each of six operators have headphones with mike. Sometimes it is necessary that all group listen with one ear first responders communications and use other ear with their interlocutor (phones or radio).
Sometimes also needed when autorities come on board !
Walkies talkies are not the best audio source 🙁 So not much signal above 3000Hz. According to cbdb's doc, I already loose 20% of intelligibility.
Following all advices solution must be electronic (signal processing) and electroacoustics. Book monitors are easier to integrate in soft room than PA speakers, which drivers size will you recommend ?
Each of six operators have headphones with mike. Sometimes it is necessary that all group listen with one ear first responders communications and use other ear with their interlocutor (phones or radio).
Sometimes also needed when autorities come on board !
Walkies talkies are not the best audio source 🙁 So not much signal above 3000Hz. According to cbdb's doc, I already loose 20% of intelligibility.
Following all advices solution must be electronic (signal processing) and electroacoustics. Book monitors are easier to integrate in soft room than PA speakers, which drivers size will you recommend ?
Last edited:
Radio communications audio in the analogue domain is bandpassed between 300hz-3khz in a 12.5/25khz channel spacing format, if the radio transceiver has dsp then the filters will probably be brickwall implementations. If the radio transceiver is operating in a digital domain like tetra/dmr then the audio is vocoded. Tetra uses the ACELP vocoder, DMR uses the AMBE2 vocoder from DVSI. In any case there's not much audio above 3.4khz.
Things to consider:
Things to consider:
- What is the audio interface from the console, is it 600 Ohm balanced or headset or speaker output ?
- How are you going to adjust the speaker volume ?
- How are you going to mute/unmute the speaker audio ?
- Can you drive the speaker from the audio console directly ?
Last edited:
If only a limited angle in front of the loudspeaker has to be covered, having a loudspeaker with a high directivity (not necessarily constant) also helps, as it increases the direct to reflected sound ratio. Horns and full range drivers have this property, especially if they are large. Column speakers can be made to have a narrow vertical directivity yet still have a large horizontal coverage angle. These are used in reverberant churches.
Last edited:
No need for "full range" speakers which to extend bandwidth may acceptb some compromises you don´t really need.I wonder if make a loudspeaker dedicated to speech intelligibility have sens ? Purpose is to reproduce radio communications in dispatch room (300-3000Hz).
Are all full range speakers equals from intelligibility point of vue, or it's worth making own case with specific speaker model ?
As mentioned by others, you need *excellent* response between 250-300Hz to typically 5kHz, but reaching , say, 7 kHz will not hurt at all and slightly increase sibilants.
Personally I´d trust some 5" , maybe 6" Faital Pro mid/full range speaker.
Good efficiency and fast response (aluminum wire voice coil) at a reasonable price.
This is the digital telephone bandwidth [BW], so in theory we basically want a big version of what's in an old portable phone, but never having seen/measured one all I'll say is that based on the pioneer's research we want it as flat as practical from ~200 - 6 kHz and boost the ~ 4 - 6 kHz BW to 'taste' for increased clarity.
GM
Telephony standard: G.711 - Wikipedia
8-bit mu-law encoded, 8kHz sampling rate targeted at reproducing 300-3400 Hz.
-b
8-bit mu-law encoded, 8kHz sampling rate targeted at reproducing 300-3400 Hz.
-b
1) Are the headphones open ear or over the ear?
2) Is the interlocutor's communication through various loudspeakers (phone handset, radio speaker) or a stereo headphone?
3) The best solution requires a bit more information regarding "1" and "2", as well as room dimensions, placement of the six operators, what information must be heard by all six operators, and what should be only heard by individual operators, the need for fail-safe and back up operation, and budget.
4) The best signal to noise, direct to reflected, maximum intelligibility system would be stereo isolation headphones (or "in ear monitors", ear buds) with individual equalization, level and dynamic range mix of all the needed listening sources for each operator, with individual control of key parameters easily accessed. That solution could cost from around $400 to $10,000 depending on the hardware you presently use, and the level of control desired.
Using loudspeakers, driver size would be dependent on room dimensions and whether a distributed system (one speaker per operator) or a single point source speaker is used.
If a single point source speaker is used, it's directional control should maximize direct to reflected sound to maximize intelligibility, so a wave guide or horn may be desired. Decent (3.5") desk speakers with volume controls may be a better solution than even the best single point source speaker.
Art
@Indiglo : A VOIP interface was considered but rejected due to complexity, interference between digital codecs...
So a generic Interface was made with attenuator and 200 Ohm balanced transformer (AF stage single/bridge, distance desk/radio, galvanic isolation,...). At the end a little box with a LM386 amplifier. All adaptations are battery powered.
Issue may be a selector followed by a presence filter ending with @waxx Beyma or @JMFahey Faital driver...
@weltersys : over the ear headphones. Operator deals with his entity and fill a common handrail. Low budget as it's for a civilian group for disaster recovery.
Thanks fellows for your contributions. I have now differents approaches, methods,and many advices.
So a generic Interface was made with attenuator and 200 Ohm balanced transformer (AF stage single/bridge, distance desk/radio, galvanic isolation,...). At the end a little box with a LM386 amplifier. All adaptations are battery powered.
Issue may be a selector followed by a presence filter ending with @waxx Beyma or @JMFahey Faital driver...
@weltersys : over the ear headphones. Operator deals with his entity and fill a common handrail. Low budget as it's for a civilian group for disaster recovery.
Thanks fellows for your contributions. I have now differents approaches, methods,and many advices.
Last edited: