• Re: Speakers to play back voice

    From John Williamson@21:1/5 to Pamela on Tue May 24 13:10:59 2022
    XPost: uk.tech.digital-tv

    On 24/05/2022 12:40, Pamela wrote:
    I play voice recordings of meetings using an ordinary under-chin transcription headset. I get clear speech.

    However I want to play recordings to more than one person using
    speakers. Small hifi speakers don't play back voice as clearly as basic speakers in many tvs or radios.

    Are there any speakers (or alternatively, sound-shaping software) which
    are designed for clear playback of voice?

    If all you want is clear speech, then the information in this advert
    will give you some settings to try for your preferred FFT filter in your
    normal DAW. (Or you could just buy the plug in...)

    https://www.kvraudio.com/product/telephone_by_rocksonics_professional_audio

    Telephone is a simple yet highly realistic telephone audio emulator.
    This plug-in uses very steep, 6-pole, high-pass and low-pass filters to
    create the telephone audio bandwidth of 350Hz - 3KHz. A granulizer
    reduces the resolution to 8 bits just like the phone company. It works
    with both mono and stereo signals.

    Alternatively, buy a cheap portable PA system, and fiddle until it
    sounds the way you like it.

    --
    Tciao for Now!

    John.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Pamela@21:1/5 to All on Tue May 24 12:40:13 2022
    XPost: uk.tech.digital-tv

    I play voice recordings of meetings using an ordinary under-chin
    transcription headset. I get clear speech.

    However I want to play recordings to more than one person using
    speakers. Small hifi speakers don't play back voice as clearly as basic speakers in many tvs or radios.

    Are there any speakers (or alternatively, sound-shaping software) which
    are designed for clear playback of voice?

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From David Woolley@21:1/5 to Pamela on Tue May 24 13:46:30 2022
    XPost: uk.tech.digital-tv

    On 24/05/2022 12:40, Pamela wrote:
    However I want to play recordings to more than one person using
    speakers. Small hifi speakers don't play back voice as clearly as basic speakers in many tvs or radios.

    Cheap speakers tend to have very poor low frequency responses. Assuming
    the "hifi" speakers are part of a system, what happens if you turn the
    bass control right down?

    However, unless the microphone has an excess bass response, I don't
    really see why a hifi speaker should be worse than a lo-fi one.

    PS the standard telephone passband is 300 to 3.4kHz, so goes higher than someone else said. However even that top limit is too low to accurately reproduce sibilants, so s and f may be different to tell apart.

    (Telephone quality audio is also referred to as 3.1kHz audio, as that is
    the difference between the bottom and top frequencies.)

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Roger Mills@21:1/5 to Pamela on Tue May 24 16:40:17 2022
    XPost: uk.tech.digital-tv

    On 24/05/2022 12:40, Pamela wrote:
    I play voice recordings of meetings using an ordinary under-chin transcription headset. I get clear speech.

    However I want to play recordings to more than one person using
    speakers. Small hifi speakers don't play back voice as clearly as basic speakers in many tvs or radios.

    Are there any speakers (or alternatively, sound-shaping software) which
    are designed for clear playback of voice?

    How did you make the recordings, and in what form are they (analog or
    digital)? You may have more success playing them through computer
    speakers and, depending on your sound card, may be able to adjust the
    gain at various frequencies. [If the recordings are analog, you'll need
    to play them into your computer via the line in or mic in socket, and
    digitise them using Audacity (which is free) or similar.

    In fact, even they're already in digital form, Audacity has some good
    tools for analysing the frequency content, and applying high or low pass filters, etc. It also has some tutorials on how to accomplish various tasks.

    [Not sure why the other respondents are assuming that the recordings
    were made over the telephone - I don't think you said that?]

    --
    Cheers,
    Roger

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Bob Latham@21:1/5 to David Woolley on Tue May 24 17:07:59 2022
    XPost: uk.tech.digital-tv

    In article <t6ik37$a5d$1@dont-email.me>,
    David Woolley <david@ex.djwhome.demon.invalid> wrote:

    However, unless the microphone has an excess bass response, I don't
    really see why a hifi speaker should be worse than a lo-fi one.

    I agree, I can't either. Pamela has obviously had an experience where
    they did but if they were good recordings and good speakers I suspect
    speaker placement. Too close to a rear wall, or worse still in a
    corner or worse still near the ceiling or floor and in a corner. :-)
    Any placing like that would considerably muddy the sound.

    IMHO, away from walls and preferably on stands, good hi-fi speakers
    should give very good results on speech.

    Bob.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From David Woolley@21:1/5 to Roger Mills on Tue May 24 17:21:35 2022
    XPost: uk.tech.digital-tv

    On 24/05/2022 16:40, Roger Mills wrote:
    Not sure why the other respondents are assuming that the recordings were
    made over the telephone

    I wasn't assuming that, but just following up on the first response. I
    suspect the reason is that a telephone represents the minimum
    specification for transmitting speech that is usable to untrained users.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From J. P. Gilliver (John)@21:1/5 to All on Tue May 24 17:43:59 2022
    XPost: uk.tech.digital-tv

    On Tue, 24 May 2022 at 17:07:59, Bob Latham <bob@sick-of-spam.invalid>
    wrote (my responses usually FOLLOW):
    In article <t6ik37$a5d$1@dont-email.me>,
    David Woolley <david@ex.djwhome.demon.invalid> wrote:

    However, unless the microphone has an excess bass response, I don't
    really see why a hifi speaker should be worse than a lo-fi one.

    I agree, I can't either. Pamela has obviously had an experience where
    they did but if they were good recordings and good speakers I suspect
    speaker placement. Too close to a rear wall, or worse still in a
    corner or worse still near the ceiling or floor and in a corner. :-)
    Any placing like that would considerably muddy the sound.

    IMHO, away from walls and preferably on stands, good hi-fi speakers
    should give very good results on speech.

    Bob.

    Also depends what is meant by a "hi-fi" speaker. We here are engineers;
    to a lot of people, especially if their taste in music is for certain
    kinds, "good" (for which they might _use_ "hi-fi") can just mean good (probably, to us, excessive - or at least poorly-controlled) bass
    response. When I was looking at speakers (possibly 40+ years ago
    though!), a lot were "ported", i. e. had a hole, that let more of the
    sound from the back of the woofer unit out - this definitely gave more
    bass, but in a somewhat uncontrolled manner; ones without a port (the
    industry was calling those "infinite baffle" then) wer for a given size
    box weaker in bass output in absolute terms, but what was produced was
    better controlled. Ported are particularly bad at speech.

    Yes, I was going to suggest a pair (assuming it's stereo) of not very
    expensive "computer speakers", or at least trying those. (_Not_ ones
    that are particularly big, and definitely not ones with a separate bass
    unit.) Probably can pick up at a charity shop!
    --
    J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

    in the kingdom of the bland, the one idea is king. - Rory Bremner (on politics), RT 2015/1/31-2/6

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Woody@21:1/5 to Pamela on Tue May 24 18:19:15 2022
    XPost: uk.tech.digital-tv

    On Tue 24/05/2022 12:40, Pamela wrote:
    I play voice recordings of meetings using an ordinary under-chin transcription headset. I get clear speech.

    However I want to play recordings to more than one person using
    speakers. Small hifi speakers don't play back voice as clearly as basic speakers in many tvs or radios.

    Are there any speakers (or alternatively, sound-shaping software) which
    are designed for clear playback of voice?

    Earlier this year I was wandering round a NT property and in a couple of
    rooms heard some piano music very well reproduced. It proved to be an
    Anker SoundCore mini loudspeaker. I bought one and was/am surprised how
    good it sounds and especially the undistorted volume level it can
    achieve. Amazon seems to be one of the cheapest.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Andy Burns@21:1/5 to Woody on Wed May 25 10:30:08 2022
    XPost: uk.tech.digital-tv

    Woody wrote:

    Earlier this year I was wandering round a NT property and in a couple of rooms
    heard some piano music very well reproduced. It proved to be an Anker SoundCore
    mini loudspeaker. I bought one and was/am surprised how good it sounds

    Seems a bit of an oversight for such an amplified speaker to not include a 3.5mm
    input?

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Pamela@21:1/5 to Roger Mills on Wed May 25 12:24:50 2022
    XPost: uk.tech.digital-tv

    On 16:40 24 May 2022, Roger Mills said:
    On 24/05/2022 12:40, Pamela wrote:

    I play voice recordings of meetings using an ordinary under-chin
    transcription headset. I get clear speech.

    However I want to play recordings to more than one person using
    speakers. Small hifi speakers don't play back voice as clearly as
    basic speakers in many tvs or radios.

    Are there any speakers (or alternatively, sound-shaping software)
    which are designed for clear playback of voice?

    How did you make the recordings, and in what form are they (analog or digital)? You may have more success playing them through computer
    speakers and, depending on your sound card, may be able to adjust the
    gain at various frequencies. [If the recordings are analog, you'll
    need to play them into your computer via the line in or mic in
    socket, and digitise them using Audacity (which is free) or similar.

    In fact, even they're already in digital form, Audacity has some good
    tools for analysing the frequency content, and applying high or low
    pass filters, etc. It also has some tutorials on how to accomplish
    various tasks.

    [Not sure why the other respondents are assuming that the recordings
    were made over the telephone - I don't think you said that?]

    The recordings are made by a good quality digital dictation machine
    (although it's not always optimally placed when used covertly) or
    alternatively a smartphone (set to record audio without any
    processing).

    A few recordings are of telephone conversations (made by a TrueCall
    recorder connected directly to the line) or sometimes by an Android app
    on the smartphone.

    All recordings in MP3, at bitrate of something like 128kbps and 22kHz
    sample rate.

    The most intelligibile replay I get is via a headset such as this
    connected to a PC's audio output.

    https://www.speakit.co.uk/products/spectra-sh-55-usb-transcription-
    heads
    et

    When I use hi quality ear buds as an alternative there's too much
    extraneous sound and intelligibility is worse.

    The usual problem is of muffled, indictinct or faint speech. A low-cut
    filter doesn't improve intelligibility much. I suspect that emphasising
    some frequencies may help and perhaps there is some plug-in available
    for Audacity designed to optimise speech clarity.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Pamela@21:1/5 to Bob Latham on Wed May 25 12:30:55 2022
    XPost: uk.tech.digital-tv

    On 17:07 24 May 2022, Bob Latham said:

    In article <t6ik37$a5d$1@dont-email.me>,
    David Woolley <david@ex.djwhome.demon.invalid> wrote:

    However, unless the microphone has an excess bass response, I don't
    really see why a hifi speaker should be worse than a lo-fi one.

    I agree, I can't either. Pamela has obviously had an experience where
    they did but if they were good recordings and good speakers I suspect
    speaker placement. Too close to a rear wall, or worse still in a
    corner or worse still near the ceiling or floor and in a corner. :-)
    Any placing like that would considerably muddy the sound.

    IMHO, away from walls and preferably on stands, good hi-fi speakers
    should give very good results on speech.

    Bob.

    I am looking more for a intelligible if squawky sound than
    full-spectrum high-fidelity reproduction with overtones etc.

    I wonder if the audio set-up on amateur radio transceivers is the sort
    of thing I want.

    The objective is to produce an accurate transcription of what was said.
    Being able to identify the person speaking would be nice but usually
    that is obvious from the context.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From MB@21:1/5 to Pamela on Wed May 25 13:19:45 2022
    XPost: uk.tech.digital-tv

    On 25/05/2022 12:30, Pamela wrote:
    I wonder if the audio set-up on amateur radio transceivers is the sort
    of thing I want.

    There are equalisers / filters that are used on amateur radio and SWL receivers. They will probably work on a limited frequency range but
    that might be better when it is just to improve readability of speech.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From David Woolley@21:1/5 to Pamela on Wed May 25 14:33:03 2022
    XPost: uk.tech.digital-tv

    On 25/05/2022 12:24, Pamela wrote:
    All recordings in MP3, at bitrate of something like 128kbps and 22kHz
    sample rate.

    Although not part of the problem, this is seriously over-engineered for
    the application. In particular, there will be no point in sampling
    Tru-Call recordings faster than 8kHz, and the limit for the dictation
    machine is unlikely to be more, possibly about 11kHz, to give better
    sibilants. 16kHz seems to be the top end.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From David Woolley@21:1/5 to Pamela on Wed May 25 14:20:05 2022
    XPost: uk.tech.digital-tv

    On 25/05/2022 12:30, Pamela wrote:
    I wonder if the audio set-up on amateur radio transceivers is the sort
    of thing I want.

    Amateur radio transceivers, particularly short wave (SSB) ones are setup
    to reduce total transmit power and bandwidth. They do that too
    aggressively for the best quality speech intelligibility and rely on
    things like the use of phonetic alphabets, where accuracy of
    transcription is important.

    At the low frequency end, they cut aggressively, because low frequencies consume a lot of power without contributing much information. At the
    high frequency end, the cut off far too low, because reducing bandwidth
    also reduces noise, even if one loses the ability to distinguish between
    f and s.

    A better indication of the top end for good speech is that hearing tests
    don't go above 8kHz, and many earlier hearing aids didn't go beyond
    somewhat under 4kHz. Something I'd have to ask here is the age of the listener, as what you are describing could also be the result of typical
    age related (or industrial or rock concert) hearing loss, and the real
    solution might be hearing aids.

    However, as already suggested, with good speakers you might approximate
    their effect by turning the bass control right down and the treble
    control right up.

    One other factor to consider is that, apparently, if you use a
    directional microphone too close up, you will get excessive low frequencies.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Pamela@21:1/5 to David Woolley on Wed May 25 19:37:57 2022
    XPost: uk.tech.digital-tv

    On 14:33 25 May 2022, David Woolley said:

    On 25/05/2022 12:24, Pamela wrote:
    All recordings in MP3, at bitrate of something like 128kbps and 22kHz
    sample rate.

    Although not part of the problem, this is seriously over-engineered for
    the application. In particular, there will be no point in sampling
    Tru-Call recordings faster than 8kHz, and the limit for the dictation
    machine is unlikely to be more, possibly about 11kHz, to give better sibilants. 16kHz seems to be the top end.

    I couldn't agree more but many of these values are in the presets.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Jim Lesurf@21:1/5 to pamela.private.mailbox@gmail.com on Thu May 26 09:40:21 2022
    XPost: uk.tech.digital-tv

    In article <XnsAEA2C7B6CA21537B93@144.76.35.252>, Pamela <pamela.private.mailbox@gmail.com> wrote:
    On 14:33 25 May 2022, David Woolley said:

    On 25/05/2022 12:24, Pamela wrote:
    All recordings in MP3, at bitrate of something like 128kbps and 22kHz
    sample rate.

    Although not part of the problem, this is seriously over-engineered
    for the application. In particular, there will be no point in
    sampling Tru-Call recordings faster than 8kHz, and the limit for the dictation machine is unlikely to be more, possibly about 11kHz, to
    give better sibilants. 16kHz seems to be the top end.

    I couldn't agree more but many of these values are in the presets.

    I'm assuming the 'kHz' values people are quoting are *actually* 'sampled
    per second' not 'audio bandwidth'.

    That said, what characterises the 'extranious sounds'? If it is too much low-frequency garbage and muffled mid/'high' frequencies then a simple
    filter may do the trick. Either in software or putting a simple LC HPF in
    the analog path. But the latter approach prompts asking if you can solder.
    :-)

    From a digital POV 'sox' will filter and also play audio. I suspect many
    sound replay programs with a GUI will also do so. Audacity certainly lets
    you filter the audio. I tend to do that when working on recordings from LPs
    as it makes clicks and pops easier to find.

    Jim

    --
    Electronics https://www.st-andrews.ac.uk/~www_pa/Scots_Guide/intro/electron.htm
    Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
    biog http://jcgl.orpheusweb.co.uk/history/ups_and_downs.html
    Audio Misc http://www.audiomisc.co.uk/index.html

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)