|
It is a mystery to me why the designers of
Soundcards do not supply us users with adequate details of what they do,
and how they work. Unlike other cards in the system like Graphics, Modems,
and SCSI, which also have external connectors that we plug things
into, these cards have OUR choice of inputs and outputs, and have
software-driven controls that we are expected to be able to operate confidently.
How many users think that everything we plug into the card is mixed
digitally? Why do some microphones work better than others? What is the
difference between Play Controls and Record Controls? My card has a 'What
U Hear' control, what is that for? Why does it mess up my mixer settings?
These, and other questions, I hope to be able to answer on this
page.
In the beginning, computers had very crude sound facilities, just a
squawky little speaker that beeped and buzzed on occasions (it's still
there, by the way). As games became more adventurous there became a
requirement for something better, and a 'game card' arrived with it's own
sound generator and a joystick port. Shortly afterwards, as a result of
competition from other home computers of the day, the first PC soundcards
hit the streets. In those days personal computing was in its infancy and,
apart from game players, the demand for sound barely exceeded its novelty
appeal. Most of us wondered what on earth sound had to do with computing,
even when the .wav file arrived and recording was possible the
general reaction was 'so what - ever heard of a cassette recorder?'. Sound
files at this time seemed an extravagant use of precious (10Mb) disc
space. The Soundcard pressed on, games got better, disks got bigger, and
then Windows arrived with its own sound system. By then there were
different sound standards emerging, the major two being Adlib and
SoundBlaster. At about this time the CD Rom drive arrived on the scene,
and many soundcards now included proprietary interfaces to boost the
acceptance of both.
New software now began to be developed that promised advanced sound
mixing and effects, and the chance to compose music beyond the monophonic plinkety-plonk efforts achieved by earlier programs. The Midi interface
was introduced, which connected the keyboard, via an adapter cable, through the well
established, but under-used, games port, and sound creation progressed
from FM synthesis to Wavetable and SoundFont technology. Given the time,
skill, suitable software and soundcard, it is now possible to produce music that
comes close to the sound produced by a studio full of equipment of less than
a decade ago. This is an aspect of soundcards which I know very little
about so I shall stop here. My main interest has been in the way sound
sources are mixed on the soundcard, and how the various mixer windows
relate to the actual soundcard hardware.
Once the soundcard was established as a peripheral with large market
appeal, semiconductor manufacturers all over the world began producing
chips to handle the routing of various analogue and digital signals, their
conversion to and from the digital format, and control of Record and Play
modes. This chip became referred to as the Codec or Mixer. There was no
agreed specification for this, each manufacturer determined to either
produce the de facto standard, or a design that could adapt to the
requirements of various volume card manufacturers. The evolution from 8bit
mono to 16bit stereo and the introduction of Multimedia into our computer
vocabulary saw a general trend towards Creative Labs
SoundBlaster/SoundBlaster-Pro models. The highly respected AWE series also
used a similar mixer configuration. Since many competing soundcard vendors
had to maintain compatibility with this 'unofficial' standard in order to
stay in business, this is one of the mixer types that will be reviewed in
detail shortly. Despite the relative stability in design afforded by the
SB Pro standard, the mixer hardware proved very restrictive, and resulted
in a user interface that was far from ideal, with a good deal of confusion
on how to set the various sliders to achieve the desired result. In 1996
a consortium of manufacturers, Intel, Creative Labs, Analog Devices,
National Semiconductor, and Yamaha jointly released a specification for a
new audio system framework which they called Audio Codec '97 (AC'97) . This
was designed around a two chip (Analogue/Digital) solution that interfaced
to PCI/USB/1394 bus architectures. This was revised in 1998/99, and is the
basis for most current PCI soundcards today, including Creative Labs
SBLive!, and Diamond Monster Sound MX300. The specification allows for
future development in areas such as 3D sound, and enhanced music
capabilities, whilst presenting a consistent hardware interface to the host
computer. The two chip solution is intended to be equally at home as an
integrated motherboard facility, or part of a sophisticated music
synthesis system. Of the two chips that comprise the AC'97
specification, we are interested mainly in the analogue mixer chip,
but part of the digital chip needs investigating too, as it includes the
WAV and MIDI signal paths. This will be examined in detail in a
moment.
It has taken considerable effort to compile the following information.
None of the soundcard manufacturers make public the inner workings of
their proprietary chips. The detail presented here is the result of much researching on the Internet, and perusing semiconductor
manufacturers data books, combined with observing the behaviour of the
soundcards themselves. No account has been taken of additional functions
such as 3D, echo, reverb, or DSP (Digital Signal Processing) enhancements. The variants are too
diverse, and would only confuse the operating principles that we are exploring here. Before
we look at the (generic) SB/SB Pro and AC'97 mixer diagrams, it is worth
commenting on features and characteristics that are common to both.
It may come as some surprise to find that none of the externally
applied inputs are digitally mixed to the Line Output. Instead they are
combined together as analogue signals, but in proportions determined by
attenuators that are controlled digitally in response to the position of
the relevant sliders. Digital signals originating within the computer are
mixed digitally in proportions determined by digital attenuators, which
also respond to the position of their relevant sliders. The way in which
the analogue and digital signals are combined differs in the two systems
described below.
There are two kinds of microphone in general use - Condenser (electret),
and Dynamic (magnetic). Because of its physical characteristics, the
electret type has to incorporate an amplifier which tends to make it more
sensitive. To cater for both types the input circuit usually has an
optional 20dB booster amplifier. In reality, the boost setting is required also
for the electret type, and it is common to find that cheap dynamic types
have insufficient output to drive the soundcard satisfactorily. The
microphone channel is generally mono-only because the second connection on
the microphone plug is used to feed a low voltage supply to the electret
amplifier. If a mono dynamic microphone is connected then this becomes
shorted to ground, but no damage will result from this because the circuit
is designed to cope with it. The mono microphone channel is applied to
both left and right, Record and Playback, mixer circuits. Generally, there
is no facility to balance between channels.
All soundcards that I have examined have had tone controls, either of
the Bass/Treble, or Graphic Equalizer variety. The position where these
are shown in the schematics below are derived from descriptions and
illustrations gleaned from my web enquiries, together with evidence from
personal experience. The diagrams obtained from semiconductor
manufacturers do not show these, which is rather odd because 3D and DSP
options are generally illustrated. These must be optional functions
specified by the card vendors.
It is becoming rare in recent times to find outputs capable of driving
loudspeakers directly. Most good quality PC loudspeakers are active
(amplified) anyway, so this is not a problem for most users. Having said
this, in all the soundcards I examined, the Line Output circuit was
capable of driving headphones of around 35 Ohm impedance. On those
soundcards with 4-Speaker (3D Surround) capabilities, do not be fooled
into thinking that the rear speakers can be set to copy the front sound
mix. In general, the rear speakers will only output digitally sourced
signals (wav, midi, etc.). An exception to this is a facility provided on
the MX300 board.
The essence of this design is the separate mixers for Record and Play
functions. They both share the same attenuators so the
sliders that appear in the Play and Record Mixer windows are linked -
changing the Record slider, for example, also changes the Play slider.
This can be quite disconcerting to casual users who are justified in thinking
the Record and Play functions are independent.
|
|
It is also the case that,
after a recording session is completed and the application closed, normal
sound levels, such as those expected from normal computer audio prompts,
may be disturbed. The concept behind this is to emulate a 'monitor'
function, allowing the user to review the balance of the recorded mix.
However, although the sliders are linked, each slider has its own Mute button, so it is possible to incorrectly set the Record mix, and be
unaware of this during the recording since the output of the Recording
mixer is not monitored directly.
Analogue signals, derived from the input connectors, are
stereo-mixed directly to the Line Output via Tone Controls and Output
Slider. A similar, but separate, stereo mixer is output to a stereo A/D
(Analogue to Digital) converter, destined as a digital recording input to
the computer. Digital signals generated from within
the computer are digitally mixed, again in two separate locations. One
combines with the digitized analogue signals, and is available for
recording by the computer. In the case where the soundcard supports
full-duplex operation (simultaneous digital record/play), this allows (in
theory) for all digital sources (such as MIDI) to be recorded as WAV
files. The output from the other digital mixer is fed to a
stereo D/A (Digital to Analogue) converter, and then mixed with the
analogue signals for output through the Line Output circuit. |
|
Analogue input signals are mixed using separate 'Play'
sliders, through Tone Controls and Output Slider to the Line Output
socket. As has become common with popular soundcards, the microphone
channel is mono only, being mixed equally into the left and right output
channels without the utility of a balance slider.
Internally generated digital signals, or played-back digital
files, are attenuated and mixed together digitally
before being converted to analogue, and mixed in with the external
analogue signals.
There is no dedicated mixing circuit
for recorded signals. In normal use it is only possible to choose one
analogue or digital source at any one time. Although it may appear to the
user that each source has its own slider, this is not the case. There is
only one physical attenuator available at the output of the analogue and
digital (B) selectors. These are set to the required position when each
source is selected. The switching of the digital selectors (A & B) is transparent
to the user. If an analogue input is chosen, Selector (A) switches to the
output of the A/D converter. If a digital input is chosen then Selector
(A) derives its input from the source chosen by Selector (B). To
enable full mixing facilities in the Record mode, there is a loopback
connection from the output of the analogue mixer to the analogue Record
selector. When this is chosen, the Play sliders now become the main input
recording mixer controls, with a main recording level control slider being
provided by the Analogue Recording Slider. The Line Output now becomes a
Recording Monitor channel. The Tone Control circuit is usually located
inside this loopback connection to provide tone correction to the recorded
sound. The Line Out slider operates independently of the recording level.
Despite Creative Labs being one of the
partners responsible for the specification of AC'97, their loopback
implementation appears to differ from the essence of the original design.
They call this function 'What U Hear', but for some unknown reason it only
allows one analogue input to be enabled at any one time. In other words
there is no global mixing function available in recording mode; only one
analogue input can be mixed with the digital sources. When using
this soundcard with Studio400, ensure that 'What U Hear' is not selected,
else there will be problems mixing normal (Play) analogue inputs to the Line Out
connector. This card supports 4
channel speaker operation for 3D sound effects. Exactly how this has been
implemented is not understood, but be aware that the rear speaker output
socket is unlikely to have a mode that provides the same audio mix as the
front speakers. When used with Studio400, it is essential to only use the
front speaker Line Outputs.
This card adheres very closely to the
spirit of the AC'97 specification, even to the point of providing a mono
loopback facility (not shown on the above diagram). This is merely
an extra input to the Analogue Selector which is the sum of left and right
loopback channels. These loopback inputs are referred to as Mix(S)
(Stereo) and Mix(M) (Mono), and full Record mixing facilities are
available when either of these are chosen. Although
we are not discussing 3D facilities on this page, it is worth pointing out
that Diamond have implemented a mode where the rear stereo outputs are a
duplication of the front stereo channels. They actually refer to this as a
feature to replace the need for a 'Y' cable connection when headphones are
used. |
