Back to Computers Back to Home

and the joy of digital audio recording

   This is something that I have wanted to try for several years, but until recently, the gear has been too expensive, and the results too unpredictable, and of too low a quality. This has all changed in the last couple of years. Today's hard drives are large enough to easily work with the huge files sizes of audio captures, memory is cheap enough and compact enough, and processors are more than powerful enough. In addition to all of this, the software has gotten to the point where you no longer need to be an expert to get good results (though a certain amount of expertise will help you to get better results).
     Today's middle of the road computers, which is what this machine happens to be, are more than powerful enough to do high quality audio, and even video work. The sound card is highly capable, and allows me quite a bit of latitude for experimentation and for learning.
The Audigy 2
File Types
Music Recording
About digital audio
Case  CPU RAM HD Video Drives OS Sound Monitor Modem Network
Midtower PIII 933 512mb 20,gb,
ATI Radion
WinXP  Audigy2plat
none 10/100
      Mogul is my sound and video computer. It performs most of the same functions as were once performed by the now retired Little Guy, only it does them much better. I use Mogul to do video captures, still captures, and to digitize music. This computer sits in my living room, adjacent to my stereo/video systems, and is connected to my AV receiver as an auxiliary. It uses a wireless keyboard and mouse, and can feed to my 27" tv, as well as to it's own monitor. This machine has no modem, but is capable of connecting to the internet via my network.
    This machine started out as a Dell GX110 tower, but has been significantly enhanced.  I have added a RAID controller, with a bios which permits hard drives to exceed the 137gb limit. I am presently using it to connect dual 120gb drives, but plan on having perhaps four 120gb (or larger) drives all together. The operating system, and all programs are on the original 20gb drive. The 120, and any additional drives, will be to store audio and video files. I may stripe the drives, to speed up performance. The machine also has a CD writer (external), and a DVD writer installed, as well as a DVD player.
    The video card is an ATI Radion. This card has 128mb of ram, and is an 8x AGP card, which is only able to operate at 4x on this particular mother board. Still, it is plenty fast enough for my purposes. I also have an outboard tv tuner card, with direct video inputs. The Radion card itself is capable of direct video outputting to any standard source. This combination allows me to take video captures, and still captures from DVD, television, VCR, and even direct camcorder sources. Once inside of the computer, I have several programs, including Adobe Premier, and Photoshop, with which to modify, cut, edit, and enhance these videos, and still captures. I am considering the purchase of an Audigy Video Editor sound card, which is also capable of doing full frame video capture.
    In addition to the video card, this machine also has a tv tuner, with a direct video in jack.  This card is also made by ATI. The tuner is digital as well as cable capable. It is able, with included software, to take video and still captures, either from the tuner itself, or from a video source plugged into the rear jack. I have already used it to take countless video stills for use in my other web pages. I am also considering digitizing some video and putting it on my site. This could include some travel video, or perhaps some firearms demonstrations. I have numerous videos of the various animated neon signs of Route 66, and may put some of them on the web. Who knows?
     The only problem with video on the web, as we have all learned through frustrating experience, is that clips with any kind of resolution are maddeningly slow to download. They are also very draining on bandwidth, and I only have 50gb of bandwidth per month, at my current service level, along with 2gb of storage space. This is plenty for my site, as it now stands, but I would quickly find myself running out of bandwidth if I put up video clips. I will just have to wait and see.
    Back panel connectors include one for direct video, and for a coaxial connection to an antenna or outside RF signal. I had initially had the tuner card installed in my internet computer, and latter in Timbuk2. Neither of these computers had the power to really run the card properly, however, and the experience was rather dissatisfying. The card, and the software were consistently locking up my computer, whenever I attempted to do live video captures. The GX110, with 512mb of ram, and the larger hard drives does a much better job, and the card now runs with no problems. The back panel on my machine is getting to be a pretty busy place, with my video, audio, network and other connections. In addition to my active connections, I still have the connectors for my on board audio, though it has been disabled to permit use of my Sound Blaster Audigy card.

The Audigy 2 audio card
Audigy2 ZS
Signal to Noise
Audio Response
24 bit
105 db
96 mhz
10hz to 46khz (+-3db)
24 bit
108 db
192 mhz
10hz to 46khz (+-3db)

    The Audigy sound card has excellent audio playback, and recording performance. This card is the centerpiece of this computer, and it's use is the reason that I put Mogul together. The Audigy card is beloved by gamers, and by audiophiles. In certain configurations it is also beloved of musicians and home recordists. The particular version that I have installed is the Audigy 2 ZS Platinum version. Though it has many other virtues, what really sold me on this card is it's impressive audio recording capability. It is capable of 96mhz sampling, and 24 bit recording, with a 105db signal to noise ratio. These specs are better than those of digital  CD players (which use 16 bit resolution, and 44.1 mhz sampling), and are far better than anything that can be done on vinyl or cassette. This particular version is designed for the home recording studio, and features a front mounted breakout box. The breakout box permits me to connect directly with RCA plugs, firewire, optical, midi, or standard 1/4" plugs as used in musical instruments. The inputs are balanced to permit me to do anything from digitizing music off of vinyl albums and cassette tapes, to connecting a guitar or an electric piano and making live recordings. I then have the option of mixing, modifying, and arranging recordings through a variety of software tools. My first project was a simple one. I digitized my dad's old albums, and posted them on his section of my site. There are also rear panel connectors for firewire, digital audio, and three analog lines out, as well as the standard line in. The card itself has connectors for three audio sources such as cd or dvd players, and for the breakout box and various other accessories. It also has a direct digital connection for CD audio players, allowing direct digital dubbing.
    The card also comes with an infra red remote, for use as part of an entertainment center. In addition, this card has full support for midi, including a wave table, and sound font capability. It also decodes Dolby Surround 7.1. This card is presently (03/2005) selling for about $200, but is a bargain compared to electronic recording equipment of just a few years ago. In point of fact, in some ways, this card is more capable than even professional studio gear of the last decade. In addition, all of the work is done digitally, rather than in analog form. This permits great isolation of the signal, and facilitates lossless, distortion free duplication, and dubbing.
    One of the great features of a dedicated, powerful sound card, like the Audigy series, is that it has it's own hardware DA/AD converters. Most of today's sound cards, and onboard sound chips, use algorithms to do the conversions, and use the computer CPU to do the work. This puts quite a load on the processor, which is also busy performing other tasks. This also puts some limits on the bit resolution, sampling rate, and especially on the real time processing speed.
    The only problem here, is in getting the signal to the converters. The converters are the actual bits of hardware that change the analog wave forms into digital number streams. Up until the conversion is made, the audio is in analog form, subject to all of the noise, distortion, and corruption that has always plagued analog signals. It remains subject to these influences, as it winds it's way through the computer, via ribbon cable from the breakout box to the PCI card. The converters themselves are also subject to a certain amount of stray signal, while they reside inside of the computer case. External converters would probably do a better job, but are expensive, and cumbersome, compared to this all in one, internal solution. At the present time, this card is more than "Good Enough", and as we all know, good enough is the deadly enemy of perfect.

    The Audigy card is a great and capable tool, but like any tool it requires manipulation, and is only as good as the control that can be exercised over it. In the case of any kind of computer gear, we are talking about drivers, and about control software. The Audigy card includes a considerable amount of software, which utilize it's capabilities pretty well. As of this writing, I have barely scratched the surface of what this card can do, but plan on improving my knowledge. I will probably begin a section on digital audio recording, as I learn the techniques of the process.
    The card came with a couple of major pieces of software, as well as several handy little programs, none of which I am really able to use, as of this writing. The centerpiece of the software suite is made up of Steinberg Cubasis, and FL Studio. These programs permit digital capture, mixing, dubbing, special effects, and recording onto media. There are also some wave form editors, and even a 48 track mixer. It will take me a year to learn to use all of this software (if I am lucky).
    There is also considerable software for manipulating audio files, out on the net, and created by software companies. Much of this is designed for digital music collectors, and is used for converting back and forth between MP3 recordings, and perhaps music CD's. The software available breaks down, roughly, into two types. There is performance software, and there is recording software. At one time, a good sturio engineer was, literally, worth his weight in gold. Now, we can all be our own studio engineers, though the skills will certainly not come over night. Along with the usual lights, stands, amplifiers, and microphones, club musicians now bring computers, and midi controllers.


Music recording
    One of the great myths about digital recording, and one that I fell for myself, is that you always want to record at the highest sampling rate, and the highest bit depth. This attitude is a bit of a carry over from the days of analog recording, where the master recordings were always of considerably higher quality that the end result. This was done because every step of the entire analog recording, mastering, mixing, and copying process, degraded the sound quality. This is, if you recall, not true of digital recordings, where the numerical values of the digital sound stream are always rendered precisely. While it is probably a good idea to record at the highest bit depth, there may be some problems if you use sampling rates higher than those of your finished product, which will probably be a CD.
    The problem here, is that if you need to change sampling rates, some data will be thrown away, which will create distortion. Thus if you change sampling rates you put yourself in the same boat with the analog recordist, and must deal with distortion and degradation of sound quality at every conversion. Indeed, in the case of the digital recordist, the problem may be even worse. Analog distortion is introduced by the imperfections of the equipment which introduces imperfections in copies of audio wave forms. In the case of digital sampling rate changes, information is intentionally discarded, and the quality of the gear used has little bearing in the results achieved. With analog recording gear, it is possible in theory to have a perfect recording, if the gear is perfect. With digital sampling rate changes, this is not even a possibility, since the change of sampling rate, by design, throws away information. The properties of digital recordings are discussed in more detail in the next section.
    For the home digital recordist, there will generally be three uses to which the computer will be put:

Prerecorded sources
    This will probably be the most common, and can range from the transfer of old cassette, and vinyl albums to digital, to creating a compilation from a CD collection, or converting MP3 recordings and burning them on to a CD. For recordings which are already in some sort of digital format, there is no real need for any special hardware. Even in the case of converting vinyl or cassette to digital, this system is probably a bit of overkill.
    The photo to the right shows my rinky dink little set up for converting vinyl albums, or casettes, into digital. For other sources, such as (cringe) 8-track, I would do something similar. My turntable is at the top of the photo, next to an external 120gb drive. The turntable is connected through the front panel jacks, while the casette deck below is connected via a rear panel jack. My Sun SPARC machine (Unix) shares a shelf with Mogul. Actually, there is an even more rinky dink way of doing this which involves a splitter/converter for the use of standard RCA jacks in a 1/8" input in the back of a sound card.
    For those using a turntable, you may wish to be aware of the signal levels on most turntables, and input lines. For this to make any sense, to anyone younger than about forty or so, a little history is in order. Once upon a time, there were two types of phonograph cartridges in widespread use, the ceramic, and the magnetic. The ceramic cartridge was the original high fidelity type, invented in the thirties by RCA, and the forties by Columbia. These initially used crystals, but soon utilized a piezo system which gave a longer playing time, and tracked at 6 grams, as opposed to the 10 grams of the crystal versions. These cartridges were introduced along with a new style of record, made from vinyl, instead of the old shellac discs. The old shellac discs, used since the turn of the century, broke easily, had a very poor signal to noise ration, with lots of hiss, and had a frequency range of 50 to 8khz (some of the very earliest, before the advent of electronic recording methods in the mid 1920's, had a range of 50 to 4khz). The new vinyl records did not have the hiss, and could reproduce sounds up to 12khz. They also offered, at 45 or 33.3 rpm, longer playing time than the old 78 rpm shellac records.
    The new vinyl albums were great, but by the sixties, electronic recording gear had improved to the extent, that it was possible to extend the range of recorded material. This in turn made greater demands upon the cartridges, and the records. In answer to this, the new magnetic cartridge was developed. This cartridge was capable of lighter tracking, generally two to four grams, and faster response than the old ceramic styles.
    There were two issues with the new cartridges, however. The first was that they had a considerably lower output than the old ceramic types. Outputs varied somewhat from cartridge to cartridge, but the typical ceramic cartridge produced 0.5 to 0.7 volts, though some of the earlier versions made as much as 3 volts. A typical magnetic cartridge would put out 5mv (0.05v), though some could produce as much as 50mv. The Magnetic cartridge also required around 50k ohms impedance, as compared to the 1m to 2m required by the ceramic cartridge. The second problem with the magnetic cartridge was that they did not play flat. They required special equalization in order to faithfully reproduce the recording. Both of these issues were dealt with by the use of a phono preamp, or by a special phono connection on the amplifier.
    Original line inputs were developed for the voltage produced by the ceramic cartridge. When home tape recording and audio gear began to appear, it was designed to operate at this same level, in order to conform to existing equipment. Where a tape source can be connected directly to the line in on your sound card, the phonograph will have to be connected through a preamp, or through a system amplifier. There are some exceptions however. The Sony turntable which I use, has a switch for use with a standard line in. This activates a built in preamp/equalizer within the turntable itself. This is a nice touch, and something that would have been very welcome back in the sixties and seventies when the turntable reached it's peak of popularity. The other exception would be if you could find an old style turntable with a ceramic cartridge, though I strongly recommend against this approach. The old ceramic cartridges were heavy, wore out records, and did not sound as good as the magnetic cartridges.
    CD is a little different. Audio CD can be recorded pretty easily, through my computers built in CD drive. Unlike most computers, this one has the CD connected to the sound card via digital cable, rather than the standard audio cable. What this means is that I can do direct digital to digital dubs and copies. In most computers, CD audio is sent down the cable in analog form, and is converted back into digital when it is dubbed.

Live performances

This can be a real challenge. It is one of the few instances where the break out box, and other features of this card are really needed. There are essentially two types of recordings that this set up will permit. The first is if the wave form, which essentially turns the computer into a digital tape recorder, and the second is through Midi.
    Midi is more like a control system than a music recorder. It does not record waveforms or sound at all. Instead, it records commands. Music is not really played in Midi; it is constructed. This offers some real advantages to the composer, or to the electronic music buff. Midi can score music played into a computer, it can also modify and change the sound, or apply other sounds, including those stored in a sound font.
    A sound font is a collection of sound either recorded, or created by the user. This is similar to the use of a wave table, for playing back Midi over the web. A wave table works by using a collection of stored waveforms, one for each instrument. As an example, the sound card may have a waveform stored for a piano. The midi file will have instructions saying retrieve the piano waveform, and to play certain notes at certain intervals using this waveform. This permits Midi files to be very small, compared to recorded music files, since the Midi file is little more than a glorified text file.
    Sound font sounds can be created by the user, either by recording sounds and saving them off as sound fonts, or by editing sounds in a wave editor, or modifying and processing sounds to change them. Thus you can record a glass breaking,  make it a sound font file, and then use it as an instrument to play back Midi files. You can also use the sound font files as a wave table for live performances.
    Interestingly, you do not even need to use or be able to play an instrument to make music in Midi. You can select a sound or an instrument, designate which notes are to be played, and the measure of the notes, and the music will be played by the computer. Midi is a great tool for setting up background music to set against a solo, or for hearing compositions.
    Recording live performances as waveforms, and then converting them into MP3 files, or some other type of audio file, is made much easier through use of the break out box. The front mounted break out box of the Audigy allows me to set input and mike levels, and to connect through a variety of sources.
    One of the shortcomings of this card is that it really only has one mike line. This means that I can plug a mike or instrument into the mike line, another instrument into the Midi input, and perhaps another into the firewire or direct digital input. This makes doing real studio type recording session rather difficult, unless I decide to go with a simple stereo pair of microphones. Still, a considerable amount of really good work has been done with a simple stereo pair, and this is more than good enough for someone at my skill and involvement level. For those with the knowledge (and the money) to make use of them, more elaborate break out boxes can be added through the front mounted firewire, optical, or digital inputs.
Dubbing and mixing
    Strictly speaking, you probably don't need a really special card to do dubbing and mixing, once the audio stream has been captured; but the better your card, the easier the time you will have, and the more likely it will be that you will be able to find a good software package that will be compatible with your hardware. A good card will also save you some time and effort. Though the computer's CPU will do the actual signal processing, it will be the sound card which converts the signal for playback. Without a good sound card, real time playback is nearly impossible.
    You don't really even need a special sound card to capture the origonal audio either, though a better card will give you a much better sounding stream. Any sound card can be used for doing digital sampling, with the proper cabling. The standard sound card or on board audio in your computer right now can be connected to a cassette, or with a proper preamp, to a turntable. You can also connect from the liine out on your amplifier, reciever, or av center, to the line in on the sound card. This can all be done through the use of a mini plug to RCA jack converter, which is shown in the photo to the right.
    These converters are available at most electronics stores, computer places, or stereo shops, as well as some discount stores. They range in price from about ten dollars, to nearly forty dollors, depending upon the length, the make, and the quallity of the wire. I picked up several sets for about ten dollars each. Some after market sound cards will include a set for free.
    These cables are not limted to be used for connection to standard audio cards. Though the front panel jacks, and breakout boxes are convenient, many high end cards are also connected through the use of adapter cables. I use this type of cable to connect to my casette deck, and to play my computer through my stereo.

Audio file types
    Though there are hundreds, or perhaps thousands, of audio file formats, there are only a few dozen which are well known, and only a few of these are in common use. I do not include streaming file types here, because they can not ordinarily be saved or worked with (though there are ways). DRM stands for Digital Rights Management, and refers to an encryption method used to prevent dubbing or free distribution of files. DRM capable audio file types can be set up to be uncopyable except by an original owner, or by a person possessing an encryption key or password.
File Type
Native Program
MB per minute
10 mb


2:1 to 120:1
1mb (10:1)

Real Audio

Microsoft’s Windows Media


Standard Music CD format
10 mb
16 bit

Microsoft Audio Video Interleave

Midi file format

Creative format

Quick Time


    The only formats with which I am really familiar are MP3, WMA, RAM (or RM), and wave. Actually, we are all familiar with wave. Wave is the original Windows sound format, and is the format used for all of the Windows sound effects, and themes, It is lossless, but also creates very large files. Wave files can have either 8 or 16 bit resolution. We are also all familiar with the CDA format, since it is the way in which audio CD's are recorded. Like wave, CDA is lossless, with no compression, and produces very large files. CDA is rendered in 16 bit form.
    Real Media files are capable of utilizing very high compression, but are unsuited to any really high quality work. They are very popular on the web, because of their ability to be rendered down into very small file sizes, and even to stream live across a low speed connection. They also offer a form of DRM, though it is a rather easily broken form. The only real suitable files types for high quality audio, with reasonably files sizes are MPEG, MP3, and WMA. Two of the three offer DRM, MP3 does not, which goes a long way to explaining it's popularity among those who trade music, and who keep large personal music archives. From a compression and quality standpoint, the three formats are said to be equal in quality, with a bit of an edge in compression ratio going to WMA.
    From a personal subjective point of view, I prefer MP3 to WMA. Both have their own methods of compression, but the WMA version seems to be more intrusive. MP3 seems to have more hiss than WMA, but WMA produces a definite tinny artifact. Though these improve with both formats, when lower compression settings are used, they never completely disappear. The most common compression ratio used with MP3 is 12:1, which is said to offer CD quality. All compression algorithms throw away or modify some data. You don't get something for nothing.

    For distribution, or archiving, there are a couple of choices here. The most popular is probably just to burn the files onto CD. This gives the option of either using the standard CDA format, which can be played on any CD or DVD player, or to save the cuts as MP3 files, or in some other form which is playable on a computer. The other choice is to transfer the files to an MP3 player, or to save them on a memory card. The MP3 format is particularly attractive because of it's universal popularity, and the considerably smaller file sizes, when compared to standard CD cuts.
    CD is probably the most popular choice because of it's widespread use, and the low cost of the media and of the recorders. This is also old, and tested technology, with few technical issues, and is a very stable platform for storage. CD's are said to have a longevity of better than thirty years. Today's CD media is capable of holding 700mb of data.
    Shown in the photo at left, is my little CD stack. The bottom unit is an external USB Cd burner. The unit above it, which looks like a second burner, is a USB label printer. This printer uses a thermal print process to print a label directly onto a standard CD
print before recording. It will print text or graphics, imported from standard graphic files.
    Other options for labeling disks are the use of standard sized stick on labels, or of special blank disks which allow for direct printing on the disk surface. None of these options are particularly cheap, and some of the labels can actually cost more than the disks to which they are affixed.
    An intersting new development is that of the Lightscribe media, and the Lightscribe capable recorders. Lightscribe uses the same loaser which writes the files, to draw on the opposite surface of the disc. At present, this is stricly black and white, or more accurately, tan and gray; but who knows what the future might bring. As of this writing, Lightscribe capable recorders are compatative in price with standard recorders, though the media is anopther story. Lighscribe blank dvd's are presently (mid - 2006) selling for about a dollar each. Lightscribe CD's cost around fifty to seventy five cents. This makes them little diferent in cost from standard media, once the cost of the expensive labels, or direct labeling ribbons is taken into consideration.
    There is also the option of using recordable DVD media, though this is not yet as desirable a method as using CD's. The recordable DVD format is still undergoing development, and there are still issues being resolved regarding digital rights management. There is also the cost of the media and of the drives. Still, DVD is probably the wave of the future. The new audio DVD format uses 24 bit resolution, as compared to the 16 bit resolution of audio CD. The audio DVD format also has a copy prevention component, something which was never incorporated into the audio CD format. As with the recordable CD, the recordable DVD offers the option of storing audio cuts as files.

About digital music
    At the risk of destroying a sacred icon, in truth digital music recordings  are not intrinsically better than analog recordings. In many ways they are not as good. What digital recordings have going for them is the integrity of their material. A digital recording will sound as good after the 100th dub, copy or play, as it did when it was first made. Analog recordings degrade at every play, and degrade greatly, unless very great effort is made, at every copy.
    An analog recording is actually a direct image of a waveform, including amplitude, and frequency. When an analog track is recorded, a needle, or a magnetic coil is directly moved by the sound source, and leaves a track on a tape or on a metal cylinder. When an analog recording is played back, this process is reversed, and the groove moves the needle, or the track of charged particles moves the coil of a playback head. This directly and mechanically recreates the original wave.
    The problem with the analog method is that it records everything, including noise, and extraneous vibration. It also picks up any background noise every time it is copied, processed, or dubbed. In addition to this, any inconsistencies in speed, pitch, or reproduction in any piece of gear is picked up, and will be faithfully recorded in any future dubs down the line. As if all of this were not bad enough, every time you play a tape, or a record, it is worn and erased, just a tiny bit.
    When a digital recording is made, the original sound waveform is measured, and plotted. It is then converted into a stream of numbers. Every number represents a height in amplitude. In a 16 bit recording, which is what an audio CD uses, there are sixteen different levels of amplitude possible. The human ear can determine a change of 3db, so if each of the 16 different amplitude levels were to be rendered as a 3db difference, we would have a dynamic range of 48db. If this were done directly, then this would be the dynamic range of audio CD's. However, the converters used are designed to be able to average out between steps. What this means is that the converter will produce an extra level of range between each level, essentially doubling the dynamic range to 96db. If you look at the specs for an average CD player, you will see that this is just about the rating listed. So much for dynamic range, but if this is the only value represented by number, then what about frequency?
    Frequency is determined by how the various amplitude measurements are stitched together. The diagram below shows this pretty clearly, and also shows the great weakness of digital recording. Much of the data is thrown away to be filled in either by the converters in the player, or by the human brain's wonderful capacity to detect patterns, and to complete them when they are incomplete.
    So in summary, sampling rate determines frequency; bit resolution determines dynamic range.

Quality of components
The most important factor for high quality digital audio recording is having high quality converters which are external to the computer. This is, unfortunately, something that my set up does not have. A 24 bit 96 khz converter mounted on a PCI card inside the computer will pick up pretty much noise, as much (if not more) than 16 bit 44.1 khz converters.
Sample rate conversion
You lose a lot of fidelity and accuracy if you need to convert sample rates. Reducing sampling rates intentionally throws away a considerable amount of information which can never be recovered. If your primary output is 44.1 khz, the default for CD audio, do yourself a favor and record, mix, and output to 44.1 khz.
bit depth is more important than sample rate
If you wish for a high-fidelity alternative to 16 bit 44.1 khz, it is better to increase the bit depth (to 24) than to increase the sampling rate. Converting bit depth does not introduce any distortion (unlike converting sampling rates), though you can lose some very soft detail when stepping down to 16 at the end.
Why does bit depth matter more than sampling rate? It's a matter of the human ear, which is only accustomed to hearing and resolving frequencies from 20-20,000 hz., but can resolve about 130 dB of dynamic range. The dynamic range of CDs and 16 bit digital audio is, at best, 90 dB; the dynamic range of 24 bit digital audio is 109-120 dB, depending on the quality of the converters. You see, the DVD audio standard allows a wider variety of amplitudes, or volumes, than CDs (which is its main improvement).