Last night (as I wrote this) the Audio Engineering Society’s DC section
was treated to an informative presentation on “The Art of Jazz
Recording” by Grammy-award-winning engineer Jim Anderson, who is also
the AES Eastern Region VP. The discussion included microphone selection
and techniques, small EQ adjustments made to alter an instrument’s
sound for emotional/musical effect, whether or not to use the LFE
and/or center channel, and producer/artist pressure applied to the
engineer to use dynamic compression so the result sounds louder.
Anderson owns quite an assortment of microphones, and described one
technique for close-miking a snare drum—point the mike down at the drum
head for a stronger impact sound. Aim it more parallel to the drum head
to hear more of the drummer’s brushes.
Anderson generally agreed that the LFE isn’t necessary for music, and
he doesn’t like to use the center channel, but he has put a little
signal in each depending on the project and the producer’s preference.
He acknowledged receiving pressure from artists and producers to make
recordings louder, and does his best to thwart those efforts. However,
he has compromised when they are adamant; after all, they control the
money and future projects.
Anderson also explained that when using compressors on vocalists, the
frequency band(s), threshold detection technique, and attack/release
times can easily result in accentuated sibilance, which is quite
difficult to benignly eliminate in compressed recordings. When those
recordings are played on radio stations with their own compressors, the
problem is often exacerbated.
This all brought to mind the bad DVD image quality I have seen and read
about, the generally poor sound quality on recent CDs (suffering from
severe compression to make them louder), the strong marketing success
of low-quality MP3s, and improper bass management in many surround
processors. Over the past few years, as Editor of the Boston Audio
Society’s journal The BAS Speaker, I have published synopses and
reprints exposing some of the severe quality compromises that have made
their way into commercially released DVDs/CDs and consumer equipment.
The technical professionals I know try to produce the highest quality
products, whether source material or the system through which the
consumer experiences it; they only lower their standards under duress.
It seems that many audio and video professionals are being forced to
compromise image and sound quality to accommodate the lowest common
denominator of user equipment and its setup, apparently in the name of
marketing. I know of one instance in which the artist insisted on such
severe reduction in sound quality in pursuit of higher CD levels that
the respected Boston-area recording/mixing engineer gave up the job
rather than produce such a poor-sounding product.
Some of the more obvious and widely employed compromises include:
Severe dynamic compression for maximum overall perceived loudness at
the extreme expense of musicality, dramatic impact, natural-sounding
instruments (including voice) and relatively distortion-free sound.
This even occurs on jazz and classical music CDs.
• Not using the center channel because many home surround systems
aren’t set up properly, even though a center channel speaker offers a
more stable soundstage center for a wider listening area, and
substantively reduces the well-documented comb-filtering
frequency-response anomalies caused by the phantom centered sound
coming from two speakers.
• Putting bass in the LFE because “consumers paid good money for the
subwoofer and want to hear something coming out of it” even though all
five channels in the surround formats are specified to handle the full
audio bandwidth. The LFE (which is a signal transfer channel, not a
loudspeaker) was designed to be used only for large low-frequency
effects such as explosions and rocket liftoffs, the levels of which
would impose severe dynamic range limitations on the remainder of the
• Adding excessive video edge enhancement because most people are
watching DVDs on cheap TV sets, and this video distortion gives the
impression of greater sharpness. What makes this even more of a problem
is that MPEG-2 encoding and decoding can add distortion that looks
similar to edge enhancement, depending on many factors including the
severity of the compression.
I know these are all financially motivated decisions, but it becomes
quite frustrating to spend much money and years—even
decades—steadfastly putting together carefully selected equipment into
a system that can reproduce exceptionally high-quality video and audio,
only to have it wasted on so much poor source material. It consistently
amazes me that people use these compromised video and audio recordings
to try to judge the quality of equipment. GIGO works! If garbage is fed
into a system, only garbage can come out of it, no matter how good the
system is. From the other end, if equipment is improperly designed or
built, even the best source material cannot look or sound good.
There are ways to produce the highest quality digital video and audio
source material and include metadata in the bitstream. One example is
the Dolby Digital bitstream, which supports many metadata parameters
including DialNorm (dialogue normalization) for overall level
calibration and adjustment, and DRC (dynamic range control) that
supports automated dynamic compression. Another technique is to apply
about 1.3:1 dbx compression (which uses true-RMS detection and
decibel-linear compression, and tracks perfectly in the digital domain)
to the recording’s content, which can easily be expanded back to full
range in the playback equipment, yet in compressed mode sounds fine for
Again, a metadata bit could be set to alert the playback equipment.
Metadata in the bitstream enables the quality of the playback to be
automatically optimized for the playback situation—dynamic compression
for automotive or background audio, edge enhancement for small TV sets,
maximum video and audio quality for those of us with the best playback
systems. There are unused bits available in the DVD and CD bitstreams
for as-yet-undefined metadata. Displays and players already have much
computer power running them, so they can easily be made smart enough to
read the metadata and, based on the equipment and the user’s
preference, distinguish what the source material is capable of
delivering, and how to process it to meet the user’s desires.
A variety of quality levels in recordings and equipment is inevitable.
The question is how to “mark” the recording and build equipment that
can easily accommodate the variety of applications. Now if we can only
get the DVD/CD producers and equipment manufacturers to implement it.
– David J. Weinberg (Tobias Audio, Silver Spring,
Md.; 301-593-3230; WeinbergDa@cs.com) is a recording engineer, as well
as an engineering consultant and technical journalist on audio, video,
and film technology. [Reprinted, with permission by Audio Amateur
Press, from Multi Media Manufacturer, Volume 2, Issue 4, July/August
2005, by David J. Weinberg, p. 14. © Copyright 2005, Audio
Amateur Corporation, P.O. Box 876, Peterborough, NH 03458, USA. All