opcom wrote:
Floyd Davidson wrote:
opcom wrote:
Not studio quality.
Studio quality AM is +/- 5KHz.
"Studio quality AM" is not a valid term. AM Broadcast has never
been up to "Studio Quality" standards, which would be _at_
_least_ 20 to 20K Hz.
AM Broadcast Band quality is 5K Hz.
It's semantics at this point, the AM studio is perfect if it is
limited to about 5KHz, something which is quite a different
case than an FM studio or a recording studio.
I doubt if many AM broadcast stations have studios that are
limited to 5KHz, any more than an FM station would. They
don't commonly go to the effort that a recording studio does,
but they certainly do not restrict fidelity to 5 KHz audio.
SSB? 2.5KHz is fine. I am speaking up for the old iron. I do
limit my AM to +/-3KHz in the speech amp.
There is no difference at all between the audio response
necessary for SSB and AM. AM, because it has both sidebands,
will necessarily take up twice the RF spectrum for the same
audio response, but in fact 2.4KHz (400Hz to 2800Hz) is actually
*preferable* to higher fidelity audio response when the purpose
is voice communications. (Ma Bell did a bazillion studies on
this decades ago, so it is not exactly new information.)
Can you find this data for me? I am very interested, it might
shed more light on the subject. I understand it may be hard to
find today.
I don't know where you'd be likely to find it. Perhaps in a large
science library.
Since SSB has no carrier, you can shave the bandwidth a little
You are talking apples, while I was talking oranges.
"There is no difference at all between the audio
response necessary for SSB and AM."
Audio response, not bandwidth. Clearly the bandwidth required
to get the same response is different for different modulation
schemes.
respect, maybe I have hearing damage. In amy case I am not
advocating more than 'necessary', but I do have trouble with
speech signals that are cut off too sharply. I always tune my
SSB receiver so that the voice is higher pitched than natural,
then it is easier for me to hear. It doesn't work with AM that
way, the tuning.
You would enjoy using the old Drake receivers, with a tunable
IF bandpass filter. You could locate the bandpass anywhere in
the audio range that suited your ears. (I also have hearing
loss, and when I was young I enjoyed music with a lot of bass
boost, while today I want the treble cranked up!)
But anyway, here are my sources to support the freedon to
employ at least some leeway for bandwidth in the matter of
communications quality AM. It's definitely your right to
interpret them how you wish:
================================================= ========
"Understandable speech requires the reproduction of all
frequencies from about 250 to 2700 cycles, or sideband
frequencies ranging from 250 to 2700 cycles above and
below the carrier frequency."
FROM: "RADIO ENGINEERING", second edition, 1937, chapter 9,
section 72, page 396, "Waves with Amplitude Modulation",
Frederick Emmons Terman, Sc.D., Professor of Electrical
Engineering, Stanford University.
MY Opinion: Mr. Terman's text says 'requires',
therefore this is taken as the minimum requirement for
speech to be 'understandable'. This does not necessarily
imply good communications quality, but rather
'understandability'.
That one seems to be a good interpretation of correct
information as it was understood at the time (1937).
================================================= ===========
"Modulation frequencies Corresponding to Typical Signals
(minimum frequency range that must be met)"
"Long-distance telephone quality.......250-3500 c/s."
FROM: "RADIO ENGINEERING", third edition, 1947, chapter 9,
section 9-1, page 469, table 9-1 --Modulation frequencies
Corresponding to Typical Signals (minimum frequency range
that must be met)., Frederick Emmons Terman, Sc.D., Professor
of Electrical Engineering and Dean of theSchool of Engineering,
Stanford University. Past president, Institute of Radio Engineers.
My Opinion: Please consider the audio quality of long
distance telephone service in 1947.
But that does not say anything about what is required for
communications. It just says what was provided by Ma Bell
in 1947, which is *not* something that was specified as the
best for communications (over a single channel).
The 250-3500 Hz range was what L carrier (over either coax cable
or what was then the newly designed and soon to be implemented
TD-2 microwave systems) provided on a per channel basis. But an
actual circuit connection was expected to be implemented with
_multiple_ tandem channels... and the ultimate circuit
connection provided was (and still is today) specified at 400 to
2800 Hz. (And believe me, too many L carrier channels in tandem
could make 400-2800 look good!)
The quality required end-to-end was a 400-2800 Hz channel
(with an SNR of 24 dB or better).
================================================= ===========
"...For ordinary SSB telephony, M=3000 Hz. .."
"...For high quality SSB Telephony, M=4000Hz. ..."
"...For ordinary DSB telephony, M=6000 Hz. ..."
"...For high quality DSB Telephony, M=8000Hz. ..."
FROM: "THE RADIO MANUAL", fourth edition, 1950, appendix 5,
page 859, "Table of necessary bandwidths", George E. Sterling,
Commissioner, Federal Communications Commission, and
Robert B. Monroe, Radio Engineer, Columbia Broadcasting
System, D. Van Nostrand Company, Inc. 4th edition, 1950.
Obviously that is in very round numbers and it does not
appear they meant them to be take as a specification.
In 1950 *nobody* expected to get such (4KHz audio) a "high
quality" signal over *any* long distance facility.
================================================= ============
"Frequencies up to at least 2,500 cycles, and preferrably 3500
cycles, are necessary for good speech intelligibility."
FROM: "RADIO HANDBOOK", fourteenth edition, 1956, chapter 12,
section 12-1, page 225, Editors and Engineers, Ltd., edited by
William I. Orr, W6SAI.
Mr. Orr was pretty good with the design of linear amplifiers.
================================================= ============
"...Mediocre reproduction may be restricted to 100-5000 c/s.,
while many radio receivers are limited to 100-3500c/s. It
should be remembered that the frequency range is taken as
overall, including the loss of sidebands and including the
loudspeaker. Wide frequency range is only comfortable to
the listener so long as other forms of distortion are
negligible."
FROM: "THE RADIOTRON DESIGNER'S HANDBOOK, THIRD EDITION", 1941,
chapter 5, page 32, "frequency distortion", THE RADIOTRON
DESIGNER'S HANDBOOK, THIRD EDITION", F. Langford Smith,
S.SC., Member I.R.E, M. I.R.E., A.M.I.E.E., A.M.I.E
That doesn't seem to apply to the subject at hand. It doesn't
say anything specific, but it does touch on one point that does
need to be considered: The amplitude distortion requirements
vary with the amount of other types of distortion or noise.
Hence processed audio might need less bandwidth to be effective,
and otherwise distorted audio might need more bandwidth to be
effective.
I looked through the texts that I have (virtually all telecom
related, as opposed to radio communications in general) and
did not find anything useful. I was struck by one odd thing
that is interesting though! Everything I found was concerned
not with audio bandwidth, but *only* with reduction of the
RF bandwidth... by finding better ways to encode that imaginary
0-4 KHz telephone channel into a digital data stream that uses
less bandwidth when it modulates an RF carrier.
Maybe we are all barking up the wrong tree, and should instead
be considering ways to use digital technology to put more signals
into the same HF bandwidth?
--
Floyd L. Davidson http://web.newsguy.com/floyd_davidson
Ukpeagvik (Barrow, Alaska)