Owners Manual

The following is a small section of the Orban Optimod-FM, 8400
owner's manual. This is a compressor used by radio stations before
they broadcast the music signal. Orban is, by far, the leading
company building broadcast limiters in the world. This eloquent
piece posted on <rec.audio.pro> by Robert Orban serves as yet
another warning for those that intend to use hyper-compression on
their mix.
At this writing, there has been a very disturbing trend in CD
mastering to apply levels of audio processing to CDs formerly
only used by “aggressively-processed” radio stations. These CDs
are audibly distorted (sometimes blatantly so) before any further
Optimod processing. The result of 8400 processing can be to
exaggerate this distortion and make these recordings noticeably
unpleasant to listen to over the air.
There is very little that a radio station can do with these CDs other
than to use conservative 8400 presets, which will cause loudness
loss that may be undesired in competitive markets. There is a
myth in the record industry that applying “radio-style” processing
to CDs in mastering will cause them to be louder or will reduce the
audible effects of on-air processing. In fact, the opposite is true:
these CDs will not be louder on air, but they will be audibly
distorted and unpleasant to listen to, lacking punch and clarity.
Another unfortunate trend is the tendency to put so much high
frequency energy on the CDs that this cannot possibly survive the
FM pre-emphasis/de-emphasis process. Although the 8400 loses
less high frequency energy than any previous Orban processor
(due to improvements in high frequency limiting and clipping
technology), it is nevertheless no match for CDs that are mastered
so bright that they will curl the vinyl off car dashboards.
We hope that the record industry will come to its senses when it
hears the consequences of these practices on the air. Alas, at this
writing, they have shown no signs of doing so.
Anyone—please feel free to quote anything I’ve posted on the
board. I am trying to bridge the broadcasting and mastering
communities, and the best way is to “get the word out.”
This subject has suddenly heated up on the Broadcast.net radio-
tech mailing list. Broadcast engineers have become very
concerned about the clipped and distorted material that they are
being presented with. In fact, one well-respected poster went so
far as to propose a minimum peak-to-average ratio spec for
material that was to be considered “broadcast quality,” and
proposed that stations reject any material breaking this spec.
The consensus was that radio stations need “radio-mastered”
mixes. These can have all of the EQ and compression applied to
the standard release, but need to have the peak limiting and
clipping greatly backed off or eliminated. This will retain the flavor
added by the mastering, but not the distortion!
In this age of broadband Internet connections, it would be
perfectly feasible to service stations with “radio-mastered” singles
from a password-protected website. Most stations would prefer
uncompressed files to retain quality and prevent any issues with
“dueling algorithms,” as stations often compress later on in the
chain, either when they store the material to hard disk for on-air
playback, or in their studio-to-transmitter links (STLs).
http://www.orban.com/
We completely agree with Robert's post and the suggestion to create
a few masters with lesser amounts of limiting. Hopefully the
password protected web-site can become available and producers
and/or record companies can post optimized mixes for radio.
Perhaps Robert's post was aimed more at the abuse of multi-band
limiters, but the SLAM! can be made to hyper-compress, and/or
distort which may cause problems further down the chain than just
the basic CD intended for home listening. It is just not that simple.
For example, one might clip a track deliberately for a certain effect
or for apparent loudness. If during the song, a section has less highs,
a station's multi-band limiter may try to lift the HF bands, exagerating
the HF harmonic distortion and making it more than ugly. In fact, it
might make it un-playable by some stations.
What might we suggest? Musicians might try to play at consistant
volumes. Mix engineers might limit individual tracks and sub-
groups more than the mix. They might also want to rely more on the
mastering engineer for final limiting, and their expertise and
experience with how product translates to broadcasting. Mastering
engineers have to consider the broadcast chains. A&R people have
to realize that songs sell records, and a louder CD won't make much
difference. In fact, a CD that is too loud, too aggressive, too in-your
-face may also be too exhausting to listen to for more than one or two
songs - but A&R guys don't read manuals like this.
In more direct practical terms, run the mix 3 times and create 3
versions with different depths of limiting. This gives the mastering
engineer more to work with. The mastering engineer can aslso do the
same thing and create 3 masters. Then the only trick is making sure
the right parties get the right version, without misdirection.
Another idea mentioned earlier is limiting individual tracks, and
sub-groups. One can also create loudness just in how tracks are
mixed and EQ'ed. In fact, absolutely great mixes need very little or
nothing done in mastering (everybody's elusive goal). The worst
mixes need the most processing. Slapping a drastic processor on a
bad mix is just that, and doesn't make it a great mix or make real
mixing easier or 'mixing' something that everybody can do as long
as they have that drastic processor. Just gotta mix well first.
Perhaps the best advice is to do what experienced engineers have
done for 50 years with limiters. Use them gently and carefully. A few
dB may be better than none, and better than 10 dB of limiting. This,
of course, means you have to use your ears and meters and not
presets. The idea is not how much limiting you can get away with,
but how much and how little is optimum and still sounds good. The
usual answer is 2-6 dB on a mix (assuming fast attack time only).
In simple quick comparisons, we generally tend to prefer the choice
that is louder and most people can be fooled into thinking X is 'better'
than Y even with a fraction of a dB more volume. This is really one
place where a bit of extended listening is required to determine
which is actually better to listen to for any longer duration. Transients
and dynamics can be very nice too.
Maybe you were just thinking, how much (or little, right) should you
limit the mix for the mastering engineer. So now you have to
consider how much limiting is appropriate for the artist and song,
how much is appropriate for the CD and that audience and how much
is appropriate for radio, for the label, for vinyl dance tracks..... If
only one version is allowed - be careful, avoid regrettable squash.
18
Limiting and more limiting and more...