Sign in with
Sign up | Sign in
Your question

Audio Processing vs FM Pre-emphasis

Last response: in Home Audio
Share
Anonymous
October 14, 2004 11:11:36 AM

Archived from groups: rec.audio.pro (More info?)

Broadcast engineers frequently observe that the HF pre-emphasis used in FM
broadcasting is a problem when trying to set up their transmitter modulation
control, because of the amount of HF present in modern recordings.

But a few weeks ago I made a live, 16-bit/44.1 binaural recording of a large
pipe organ. The mics used were flat within a dB from 20Hz-20kHz, at least,
and the room was quite "bright" (sonically).

While editing the raw recording in my PC to get ready to make CDs, I saw
that even with most of the organ stops pulled (lots of 2' and shorter
pipes, mixtures and reeds going), the spectrum above about 8kHz hardly
ever rose to a level within 20dB of the spectrum below about 7kHz -- even
though the sound in the room had plenty of HF stuff bouncing around. My
recording matches what I heard in the room, to the extent my aural memory
serves (also that of the organist).

So my question is, why is FM pre-emphasis such a problem to broadcasters?
From what I see in my recent recording, LF audio could be set to modulate
the transmitter to a rather high percentage, and still leave room for the FM
pre-emphasis rise without overmodulating the HF content, or the total
waveform.

Do typical mic, EQ and audio production practices really change the spectral
content of natural sound as released on a commercial CD to the point where
spectral density is about the same everywhere in the audible spectrum? And
if so, why does a CD producer think that's a good idea?

RF

Visit http://rfry.org for FM transmission system papers.
Anonymous
October 14, 2004 2:12:20 PM

Archived from groups: rec.audio.pro (More info?)

In article <2t78nrF1s7rbqU1@uni-berlin.de>, Richard Fry <rfry@adams.net> wrote:
>Broadcast engineers frequently observe that the HF pre-emphasis used in FM
>broadcasting is a problem when trying to set up their transmitter modulation
>control, because of the amount of HF present in modern recordings.

This is not a problem with pre-emphasis so much as it is a problem with the
unnatural top end on many modern recordings.

>Do typical mic, EQ and audio production practices really change the spectral
>content of natural sound as released on a commercial CD to the point where
>spectral density is about the same everywhere in the audible spectrum? And
>if so, why does a CD producer think that's a good idea?

Yes, and it's even worse than that. There's a paper from Bob Orban and
Frank Foti called _What Happens To My Recording When It's Played on the Radio_
which is probably available on the web somewhere. It's well worth reading.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
October 14, 2004 6:47:37 PM

Archived from groups: rec.audio.pro (More info?)

"Richard Fry" <rfry@adams.net> wrote in message news:<2t78nrF1s7rbqU1@uni-berlin.de>...
> Broadcast engineers frequently observe that the HF pre-emphasis used in FM
> broadcasting is a problem when trying to set up their transmitter modulation
> control, because of the amount of HF present in modern recordings.
>
> But a few weeks ago I made a live, 16-bit/44.1 binaural recording of a large
> pipe organ. The mics used were flat within a dB from 20Hz-20kHz, at least,
> and the room was quite "bright" (sonically).
>
> While editing the raw recording in my PC to get ready to make CDs, I saw
> that even with most of the organ stops pulled (lots of 2' and shorter
> pipes, mixtures and reeds going), the spectrum above about 8kHz hardly
> ever rose to a level within 20dB of the spectrum below about 7kHz -- even
> though the sound in the room had plenty of HF stuff bouncing around. My
> recording matches what I heard in the room, to the extent my aural memory
> serves (also that of the organist).
>
> So my question is, why is FM pre-emphasis such a problem to broadcasters?
> From what I see in my recent recording, LF audio could be set to modulate
> the transmitter to a rather high percentage, and still leave room for the FM
> pre-emphasis rise without overmodulating the HF content, or the total
> waveform.
>
> Do typical mic, EQ and audio production practices really change the spectral
> content of natural sound as released on a commercial CD to the point where
> spectral density is about the same everywhere in the audible spectrum? And
> if so, why does a CD producer think that's a good idea?
>
> RF
>
> Visit http://rfry.org for FM transmission system papers.



A pipe organ alone does not represent the spectral components found in
music in general. What about if the pipe organ was accompanied by a
drummer on the snare and cymbals along with a singer. What do you
think the spectrum would look like then?

The RIAA record pre and de-emphasis curves are similar to the 75 us
FM pre and de-emphasis on FM and like on FM, this reduces the level of
"full scale" that records and FM can reproduce at higher frequencies.
Therefore in the days of playing records on the radio, there was less
problem. (Most) CDs have no pre-emphasis and therefore they can
reproduce high frequencies up to the same full scale limit as the
lower frequencies, if necessary. This probably explains why some
people believe them to be harsh or brittle.

Mark
Related resources
Anonymous
October 14, 2004 7:57:24 PM

Archived from groups: rec.audio.pro (More info?)

"Richard Fry" <rfry@adams.net> wrote in message
news:2t78nrF1s7rbqU1@uni-berlin.de...

[snip]

> While editing the raw recording in my PC to get ready to make CDs, I saw
> that even with most of the organ stops pulled (lots of 2' and shorter
> pipes, mixtures and reeds going), the spectrum above about 8kHz hardly
> ever rose to a level within 20dB of the spectrum below about 7kHz -- even
> though the sound in the room had plenty of HF stuff bouncing around. My
> recording matches what I heard in the room, to the extent my aural memory
> serves (also that of the organist).
>
> So my question is, why is FM pre-emphasis such a problem to broadcasters?
> From what I see in my recent recording, LF audio could be set to modulate
> the transmitter to a rather high percentage, and still leave room for the
FM
> pre-emphasis rise without overmodulating the HF content, or the total
> waveform.

FM pre-emphasis is such a problem to broadcasters because they're mostly not
broadcasting recordings of pipe organs. Nor orchestras. Most of them are
broadcasting contemporary pop music, with crashing cymbals and synthesizers,
and they have substantially more HF than your pipe organ.

Your observation is a pertinent one, though. When the FM pre-emphasis curve
was set up, FM was a peripheral service; the real action was on AM. Most FM
stations at the time broadcast either classical music or "light" music
(think elevators or dentists' offices). A few did unusual or esoteric things
like folk music or jazz. These musical forms have spectra that look very
much like your pipe organ recording, with little high-level material at high
frequencies. Given that expectation of program content, the standards
committees set up an FM pre-emphasis curve to take advantage of that
spectrum while improving noise performance (part of the appeal of FM being
its relatively low noise compared with AM -- this was before stereo FM).
When FM rock came along, with plenty of high frequencies, the stations were
in trouble, and the processor industry had a bonanza.

Peace,
Paul
Anonymous
October 14, 2004 7:57:25 PM

Archived from groups: rec.audio.pro (More info?)

> FM pre-emphasis is such a problem to broadcasters because
> they're mostly not broadcasting recordings of pipe organs. Nor
> orchestras. Most of them are broadcasting contemporary pop
> music, with crashing cymbals and synthesizers, and they have
> substantially more HF than your pipe organ.

It wouldn't _be_ such a serious problem if they weren't so intent on compressing
the dynamic range to increase their coverage.
Anonymous
October 15, 2004 12:44:48 AM

Archived from groups: rec.audio.pro (More info?)

"Richard Fry" <rfry@adams.net> wrote in message
news:2t78nrF1s7rbqU1@uni-berlin.de...
> So my question is, why is FM pre-emphasis such a problem to broadcasters?

Because people are clipping the audio on compact disks in order to make them
play louder in the market research focus groups that determine which new
music will be played on the air today in the United States. The clipping
distortion causes the broadcast processing to turn the volume down and this
in turn hurts a station's ratings in addition to reducing the amount of time
most people are willing to spend listening to the music on the radio.

Everybody loses except for the sponsors' middle-management who demand that
their asses always be covered by focus group statistics whenever a station
plays new music.

--
Bob Olhsson Audio Mastery, Nashville TN
Mastering, Audio for Picture, Mix Evaluation and Quality Control
Over 40 years making people sound better than they ever imagined!
615.385.8051 http://www.hyperback.com
Anonymous
October 15, 2004 12:44:49 AM

Archived from groups: rec.audio.pro (More info?)

Bob Olhsson <olh@hyperback.com> wrote:
>"Richard Fry" <rfry@adams.net> wrote in message
>news:2t78nrF1s7rbqU1@uni-berlin.de...
>> So my question is, why is FM pre-emphasis such a problem to broadcasters?
>
>Because people are clipping the audio on compact disks in order to make them
>play louder in the market research focus groups that determine which new
>music will be played on the air today in the United States. The clipping
>distortion causes the broadcast processing to turn the volume down and this
>in turn hurts a station's ratings in addition to reducing the amount of time
>most people are willing to spend listening to the music on the radio.

No, no. This is a serious problem, but it's only a small part of the
problem. While the damage that is being done in the mastering process
is indeed terrible, the amount of damage that is being done in typical
broadcast airchains is even worse. The two of them together (as the
Orban paper I cited earlier in the thread) together make something that
is even worse, with the flattopped waveforms (made in an attempt to make
the original CD loud at all costs) being mangled by the phase rotators used
in broadcast airchains (used in an attempt to make the air signal loud at
all costs) and the end result is far worse than the sum of the parts.

>Everybody loses except for the sponsors' middle-management who demand that
>their asses always be covered by focus group statistics whenever a station
>plays new music.

Don't forget the audiophile labels that five or six years later will produce
unprocessed versions of the old releases at inflated prices!
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Anonymous
October 15, 2004 12:44:50 AM

Archived from groups: rec.audio.pro (More info?)

In article <ckmt0i$9ep$1@panix2.panix.com>, kludge@panix.com says...
>
>
>Bob Olhsson <olh@hyperback.com> wrote:
>>"Richard Fry" <rfry@adams.net> wrote in message
>>news:2t78nrF1s7rbqU1@uni-berlin.de...
>>> So my question is, why is FM pre-emphasis such a problem to broadcasters?
>>
>>Because people are clipping the audio on compact disks in order to make them
>>play louder in the market research focus groups that determine which new
>>music will be played on the air today in the United States. The clipping
>>distortion causes the broadcast processing to turn the volume down and this
>>in turn hurts a station's ratings in addition to reducing the amount of time
>>most people are willing to spend listening to the music on the radio.
>
>No, no. This is a serious problem, but it's only a small part of the
>problem. While the damage that is being done in the mastering process
>is indeed terrible, the amount of damage that is being done in typical
>broadcast airchains is even worse. The two of them together (as the
>Orban paper I cited earlier in the thread) together make something that
>is even worse, with the flattopped waveforms (made in an attempt to make
>the original CD loud at all costs) being mangled by the phase rotators used
>in broadcast airchains (used in an attempt to make the air signal loud at
>all costs) and the end result is far worse than the sum of the parts.

This is an oversimplification. Phase rotators make speech waveforms
symmetrical to reduce processor-induced clipping distortion on voice. This
makes it possible to maintain an artistically appropriate voice/music balance
on-air without needing to turn the music down by as much as 4-5 dB.

In the ideal air chain, all mics are processed with phase rotators and the
main program line audio processor has its phase rotator bypassed. This is to
make music slightly more transparent sounding by not phase-rotating it.
However, this is inconvenient to achieve in the real world, particularly
considering that a significant amount of speech material may original out of
house. So phase rotators are accepted as a necessary evil in most facilities.
Anonymous
October 15, 2004 12:44:51 AM

Archived from groups: rec.audio.pro (More info?)

Robert Orban <donotreply@spamblock.com> wrote:
>This is an oversimplification. Phase rotators make speech waveforms
>symmetrical to reduce processor-induced clipping distortion on voice. This
>makes it possible to maintain an artistically appropriate voice/music balance
>on-air without needing to turn the music down by as much as 4-5 dB.

It is a serious oversimplification, yes.
Is that AES paper available on the web?

>In the ideal air chain, all mics are processed with phase rotators and the
>main program line audio processor has its phase rotator bypassed. This is to
>make music slightly more transparent sounding by not phase-rotating it.
>However, this is inconvenient to achieve in the real world, particularly
>considering that a significant amount of speech material may original out of
>house. So phase rotators are accepted as a necessary evil in most facilities.

And, if properly set up, they can be pretty transparent, and give you added
loudness with no major sonic hit, it's true. Incidentally, have you ever
seen the Garron Stereo Phase Enhancer? A station around here has one in
their airchain and nobody remembers why.
--scott


--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Anonymous
October 15, 2004 12:58:52 AM

Archived from groups: rec.audio.pro (More info?)

Bob Olhsson wrote:

> Everybody loses except for the sponsors' middle-management who demand that
> their asses always be covered by focus group statistics whenever a station
> plays new music.

Apparently their asses have gotten so large only a group can cover them.

--
ha
Anonymous
October 15, 2004 5:35:37 AM

Archived from groups: rec.audio.pro (More info?)

kludge@panix.com (Scott Dorsey) wrote in
news:ckm1g4$dkf$1@panix2.panix.com:

>>Do typical mic, EQ and audio production practices really change the
>>spectral content of natural sound as released on a commercial CD to
>>the point where spectral density is about the same everywhere in the
>>audible spectrum? And if so, why does a CD producer think that's a
>>good idea?
>
> Yes, and it's even worse than that. There's a paper from Bob Orban
> and Frank Foti called _What Happens To My Recording When It's Played
> on the Radio_ which is probably available on the web somewhere. It's
> well worth reading. --scott

http://www.omniaaudio.com/tech/mastering.htm
Anonymous
October 15, 2004 11:29:21 AM

Archived from groups: rec.audio.pro (More info?)

"Mark" wrote:
> The RIAA record pre and de-emphasis curves are similar to the 75 us
> FM pre and de-emphasis on FM and like on FM, this reduces the level of
> "full scale" that records and FM can reproduce at higher frequencies.
____________

Yes, but using pre-/de-emphasis only reduces the S/N available in the medium
at LF to improve it at HF. The technique by itself can/should be
transparent, and independent of the audio spectrum in the channel. But it
can become a complication when (uncoordinated) audio processing is applied.

The FM broadcast problem is related to the need for an FM station to
maintain some sonic consistency in their on-air sound while switching among
program sources of highly variable spectrum content and processing
artifacts. CD producers mostly consider the sound of their CD as they hear
it* in the master, perhaps without a good understanding for what may happen
to that audio when it might be broadcast.

CD producers and broadcasters both have legitimate needs and goals, and some
of them conflict. The results show that a better "system" approach is
needed.

*Is the high amount of HF on some CDs due to the HF hearing loss from
extended listening to high-SPL sound fields?

RF
Anonymous
October 19, 2004 9:43:24 PM

Archived from groups: rec.audio.pro (More info?)

In article <ckn39n$bb$1@panix2.panix.com>, kludge@panix.com says...
>
>>Incidentally, have you ever
>seen the Garron Stereo Phase Enhancer? A station around here has one in
>their airchain and nobody remembers why.

They were intended to correct interchannel delay in stereo Fidelipac cartridge
machines, which used to be a ubiquitous source of on-air audio. This delay
correction was important to prevent flanging in mono lisening because the
performance of cart machines was quite poor in this area.

One of the reasons why broadcast sound has dramatically improved in the last
20 years is that most music is now sourced from hard disk playout systems and
the cart machines have ended up (deserved so!) in the dumpster.
!