What is analogue and digital audio?
Just to recap ...
Analogue audio
Analogue audio is so called because the "shape" or pattern of an electrical or magnetic pressure audio signal is analogous to (looks like) the original pattern of changing air pressure. Audio in nature is analogue until it is converted by our inner ears into an electro-chemical signal that we perceive.
Digital audio
Digital audio is a mathematical description of the pattern of pressure. It is transmitted as a pulse wave, stored as a series of on/off switches (transistors), magnetic pulses, or optical pits and lands etc, and looks nothing like the original pattern of changing air pressure.
Ergonomic advantages and disadvantages
In deciding whether to use analogue or digital audio technologies we are often influenced by practical and ergonomic factors such as ease of session recall, portability, maintenance costs, or the omission or addition of some features on a device. Therefore before we get into which sounds better and how we might design a test to compare and evaluate analogue and digital audio sound quality, here are some ergonomic considerations ..
| Advantages | Disadvantages | |
|---|---|---|
| Analogue | Once recorded to tape, audio is stored/archived Analogue 2" 24 track tape is (was?) a world wide standard Analogue tape has a theoretically better audio bandwidth (this is arguable) Analogue tape is a tried and tested format Editing limitations of analogue tape discourage constant tinkering and procrastination Analogue tape and disc based audio is more difficult for record company executives to mess with! Analogue tape can be edited with simple splices There are many classic well loved analogue hardware processors available (EQ, compression etc) 12" vinyl LPs and 45s have nice big sleeves that are a great canvas for artistic expression Although technically a poor format, folks like the sound of vinyl (not entirely an ergonomic factor!) |
Good analogue is expensive Cheap analogue recording equipment cannot match cheap digital for audio quality(?) Cheaper analogue tape recorders suffer from distortion, cross-talk, wow and flutter, and high levels of tape noise/hiss (expensive ones too but to a lesser extent) Tape is expensive and vulnerable to deterioration Tape is becoming increasingly hard to source Tape recorders need constant maintenance Linear format - tape must be wound/rewound to the location of the recording to be heard Advanced editing of audio, including individual tracks, is difficult if not impossible with tape High noise floor requires noise reduction systems that degrade hi-frequency sounds Harder to synchronise Copying deteriorates sound The electro acoustic problems with vinyl are vast and there is no way records can reproduce a hi-res studio stereo mix masters (limited bandwidth, poor dynamic range, plenty of distortion) |
| Digital | Better sound for cheaper equipment (?) Total recall of session setup and mix status Potential for very low distortion and noise/hiss Potentially for wider dynamic range Variety of recorder options (tape, optical, hard disc, SSD, RAM etc). Non-linear operation (random access) You can access any part of the recording almost immediately Massive range of editing and processing tools Non destructive editing (you can make copies of data and undo decisions) Highly portable Data can be cloned Data can be transmitted over networks Audio can easily be integrated into multimedia (web sites, DVD, games, video etc) Powerful software integration with plug-in effects, processors and instruments Variety of real-time performance and mixing controller options Network retail / delivery |
Computers crash Software compatibility issues OS, (hosts, drivers and plug-ins) Software version shelf life (just when you have a stable system, your computer dies and you must upgrade) Data can become corrupted Data must be archived when hard drives become full Confusing array of formats and systems Sound quality can be too clinical/cold (?) Digital systems can have poor multi-user interfaces Digital audio GUI's tend to encourage visual, rather than auditory, scrutiny Editing parameters (such as when programming a synth or EQ plug-in) are tedious with a mouse |
A word about some digital audio myths
Here are a few myths ..
- Vinyl has a more natural sound, wider bandwidth and dynamic range than CD. This is simply untrue. The electro mechanical problems associated with vinyl are huge and it is inferior to digital in almost every way. This doesn't mean you can't like it though. It is true that because vinyl lacks hi-end detail, deep low end and adds harmonic distortion it may seem to add "warm" to the sound which prefer, but it cannot faithfully reproduce the quality of studio masters. 12" 45rpm vinyl produces the best quality. Unless a recording has been created specifically to be heard on vinyl, vinyl will not reproduce what the creators intended.
- CD quality digital is poor because it cuts out frequencies above the human audio spectrum which have an important effect on the frequencies we can hear. There is little empirical proof that frequencies we cannot hear affect those we can. However, it is true that some analogue audio gear benefits from having a wider bandwidth than the audio spectrum to ensure that unwanted resonance artefacts are moved outside the audio spectrum, and an extended bandwidth certainly doesn't do any harm.
- 30 ips 2" analogue tape sounds better than digital. Depends what you think is better. Tape has a reduced dynamic range (heard as a "natural" sounding tape compression), increased distortion and noise, cross-talk, and pitch/speed modulations. These distortions can be desirable and pleasing, but tape isn't technically "better" than today's digital systems.
- There is a wide difference in quality between analogue to digital converters. I, and many other recording engineers, are not afraid to say that we cannot tell the difference between modern converters (last 15 years) when they are running at at least 44.1kHz 24-bit. Any audible differences are vanishingly small and highly unlikely to make any significant difference to the quality of your productions. However, if you can hear differences and they are important to you, that is OK too.
- A 96kHz sample rate produces significantly better audio that 44.1 or 48, especially low and hi end. At 44.1 there are approximately 2,205 sample for every cycle of the waveform at 20Hz. Bass therefore is encoded extremely well and doubling this number does not create an audible improvement. At 20kHz there are only 2.2 samples for every cycle which is barely adequate. So as you increase audio frequency the encoding becomes poorer. The problem is that even doubling the sample rate only increases this to 4.4 samples at 20kHz, hardly a huge improvement when you consider you also double the load on processors and storage devices, meaning you will be able to use less tracks and processing (plug-ins) on a given system. Also, this improvement, if audible at all, is very small. It is worth trying it yourself by recording a single performance at multiple resolutions and comparing the result.
Comparing and evaluating analogue and digital audio quality
The primary debate, however, is not concerned with ergonomics or myths, but whether analogue or digital audio is better at recording and reproducing performances and recordings as they were intended to be heard by the engineers, producers and artists that created them.
So how can we compare them?
It is clear that simply comparing, for example, a vinyl pressing of a recording against a CD will not tell us anything unless we can compare both with the 2-track master heard as the producers/artists intended. We may prefer the sound of one to the other but that doesn't tell us which is objectively better.
Let's agree then, that our comparison should be to evaluate whether analogue or digital more accurately reproduces sound.
First we must establish if we are going to make an objective or subjective evaluation ..
- "Which is better, analogue or digital audio?"
- "Which do you prefer, analogue or digital?"
Clearly this article should concern itself with the first question and ask how we might evaluate a given audio systems ability to record and reproduce audio without altering it.
Finalising the question
Perhaps our evaluation should therefore ask ..
- how well does a given system record audio without altering it?
- how well does a given system reproduce/replay audio without altering it?
We may choose to try and answer both questions using audio test measurement equipment. But as useful as this may be, ultimately we are human and will probably prefer to use our ears! Therefore listening tests will be the preferred methods.
A word about processing technologies
Although challenging, we may be able to design a test using processing devices (such as compressors and reverb technologies) rather than record/replay systems. However, in general, our choice of processing device will be more about the audio character a device imparts and less about how much distortion or noise it introduces, so let's focus on record/replay systems which we primarily want to be transparent (eg flat frequency response, low noise and distortion).
Choosing source material
Our comparison should involve a source signal chain that produces a line level signal that can be fed into our recording system and later replayed through a monitoring system. Both the record signal chain and monitoring system should be identical for our comparison. So if our source signal chain involves a violinist, mic and mic pre-amplifier, the player, instrument mic and mic pre-amp should remain the same for our analogue and digital record/replay systems.
We will need some source recordings for our comparison. There are 4 primary options ..
- Use a 2-track master of a finished recording
- Make a new recording with microphones
- Make a new recording with electronic instruments with analogue outputs (electric guitars, drum machines, synths, keyboards etc)
- Make a new recording with digital plug-ins
Using a microphone captured performance
One option involves making a microphone recording of a musical performance (a musician and instrument). If the performer is available during replay evaluation comparisons, we might even be able to compare the live performance against the analogue and digital recordings, although this might be problematic.
An alternative is to record a player piano (pianola) or orchestrion. The recording can then be compared with an exact repeat performance.
Using a 2-track mixdown master
This solves the problem of comparing recordings to original performances. We simply replay our 2-track master and record it to our analogue and digital test systems. Preferably the master will need to be the highest possible quality, but even an older master tape (perhaps from the 50s or 60s) should have some value for our test.
Monitoring
We should use a single monitoring system to compare our recordings. We will be unable to make any kind of evaluation if our monitoring system is incapable of reproducing accurately the audio spectrum with as flat a frequency response as possible. This means we must consider factors such as ..
- amplifier slew rate - must be sufficient not to smear transients
- D to A and A to D converters must be reference standard or at least not budget devices
- stylus - our record player stylus must be in reasonable condition
- headphones - must be of sufficient quality with a flat frequency response for the audio spectrum
- speaker drivers and cabinets must have a specification that minimises distortion
Good monitoring systems are not cheap and most consumers will not have high quality replay systems. Nevertheless, our is an objective evaluation and we need to hear the results of our recordings as well as possible, therefore we should use the highest quality studio monitors available to us. This isn't to say that differences won't be discernible on lesser systems, you just may find some important artefact are less audible.
We can of course choose not to use our ears at all, and employ test equipment to measure noise and distortion, but for the most part we will want to evaluate by ear.
Record/replay systems
Now we need to consider our record/replay systems. Here are the primary ones still in use today (although some only just!) ..
| Consumer record | Consumer replay | Professional (studio) record | Professional (studio) replay |
|---|---|---|---|
| Analogue | |||
| Compact cassette | Compact cassette | 24 track 2" | 24 track 2" |
| 2 track 1/4" | 2 track 1/4" | 16 track 2" | 16 track 2" |
| Vinyl 33 rpm 12" | 16 track 1/2" | 16 track 1/2" | |
| Vinyl 45 rpm 12" | 2 track 1/2" | 2 track 1/2" | |
| Vinyl 45 rpm 12" | |||
| Digital | |||
| DV video camcorder (uncompressed PCM audio) | DV video camcorder (uncompressed PCM audio) | DAW / audio interface / soundcard / stand alone converter (uncompressed PCM 24-bit) | DAW / audio interface / soundcard / stand alone converter (uncompressed 24-bit) |
| AVCHD video camcorder (uncompressed PCM audio) | AVCHD video camcorder (uncompressed PCM audio) | Hardware recorders (such as RADAR) | Hardware recorders (such as RADAR) |
| Mobile phone / iPod (uncompressed PCM audio and MP3) | Mobile phone / iPod (uncompressed PCM audio and MP3) | ||
| Tablet (uncompressed PCM audio and MP3) | Tablet (uncompressed PCM audio and MP3) | ||
| PC soundcard / audio interface (all PCM and compressed formats) | PC soundcard / audio interface (all PCM and compressed formats) | ||
| CD-R stand alone hi-fi (uncompressed PCM audio) | CD-R stand alone hi-fi (uncompressed PCM audio) | ||
| DVD video | |||
| Blu-Ray video | |||
Whatever we choose we will need to ensure that our choices are equitable. We can agree that comparing compact cassette with an HD Pro Tools rig will not be equitable and we might also decide that our replay devices (eg turntable, CD payer) should be of similar cost, or perhaps the best available at any price.
Suggesting a test system for an equitable comparison
We can design a range of systems for our evaluation, but each should have identical source material/signal chains and monitoring systems.
Perhaps that best comparison we can make is to specify the best money-no-object systems for comparison so we can evaluate the potential of analogue and digital
Here are 3 suggestions ..
| Comparison 1 | Comparison 2 | Comparison 1 | |||
|---|---|---|---|---|---|
| Analogue | Digital | Analogue | Digital | Analogue system | Digital |
| 1. Source material: Microphone / performance | 1. Source material: Microphone / performance | 1. Source material: Microphone / performance | 1. Source material: Microphone / performance | 1. Source material: Microphone / performance | 1. Source material: Microphone / performance |
| 2. Record to professional analogue tape multi-track | 2. Record to professional analogue tape multi-track | 2. Record to professional DAW (24-bit / 96kHz) | 2. Record to professional DAW (24-bit at 96kHz) | 2. Record to professional DAW (24-bit at 96kHz) | 2. Record to professional DAW (24-bit at 96kHz) |
| 3. Mix to 2-track analogue tape master | 3. Mix to 2-track analogue tape master | 3. Mix to 2-track analogue tape master | 3. Mix to 2-track analogue tape master | 3. Bounce to 2-track PCM 24-bit / 96kHz master | 3. Bounce to 2-track PCM 24-bit / 96kHz master |
| 4. Cut / press to vinyl | 4. Press or burn to CD | 4. Cut / press to vinyl | 4. Press or burn to CD | 4. Cut / press to vinyl | 4. Press or burn to CD |
| 5. Replay with record player, pre/power/amp and monitors | 5. Replay with CD player, pre/power/amp and monitors | 5. Replay with record player, pre/power/amp and monitors | 5. Replay with CD player, pre/power/amp and monitors | 5. Replay with record player, pre/power/amp and monitors | 5. Replay with CD player, pre/power/amp and monitors |
We may also decide not to evaluate multi-track systems as part of the process in which case comparison systems might be ..
| Comparison 1 | Comparison 2 | Comparison 1 | |||
|---|---|---|---|---|---|
| Analogue | Digital | Analogue | Digital | Analogue system | Digital |
| 1. Source material: Microphone / performance | 1. Source material: Microphone / performance | 1. Source material: Microphone / performance | 1. Source material: Microphone / performance | 1. Source material: Microphone / performance | 1. Source material: Microphone / performance |
| 2. Record to 2-track analogue tape master | 2. Mix to 2-track analogue tape master | 2. Mix to 2-track analogue tape master | 2. Mix to 2-track analogue tape master | 2. Record 2-track PCM 24-bit / 96kHz master | 2. Record 2-track PCM 24-bit / 96kHz master |
| 3. Cut / press to vinyl | 3. Press or burn to CD | 3. Cut / press to vinyl | 3. Press or burn to CD | 3. Cut / press to vinyl | 3. Press or burn to CD |
| 4. Replay with record player, pre/power/amp and monitors | 4. Replay with CD player, pre/power/amp and monitors | 4. Replay with record player, pre/power/amp and monitors | 4. Replay with CD player, pre/power/amp and monitors | 4. Replay with record player, pre/power/amp and monitors | 4. Replay with CD player, pre/power/amp and monitors |
Establishing some evaluation criteria/benchmarks
It is important that we define some agreed criteria for evaluation and perhaps some agreed terminology (we need to agree the interpretation of the terminology too!). The primary objective criteria may be ..
- distortion - how much has the timbre of the sound been altered and is there unwanted clipping type distortion
- noise - how much has been added
- frequency response - has the frequency response been changed
- amplitude - has the volume envelope of the sound altered, perhaps exhibiting compression
- phase - has the phase relationship been altered, resulting in such effects as a widening or narrowing of the stereo field
Making the comparison
Now we have defined the question, and established some equitable equipment, processes and criteria for undertaking a comparison, we can proceed.
Ultimately we we will need to use our ears and decide which system reproduces the original source audio more accurately.
Which sounds better?
It is not the intention of this author to tell you which is better, you must decide for yourself, but I can give my opinion.
The electro-magnetic challenges of recording and reproducing audio in an analogue system are formidable. Anyone who has studied how record players work will appreciate the challenges that had to be overcome to produce acceptable sound quality. They are legion! So too are the challenges of converting an analogue line level signal to a pattern of magnetic flux on tape and then retrieving it.
Early digital systems suffered from poor anti-aliasing filters, which are employed in the signal path to ensure no harmonics can enter a recording system which are greater than twice the sample rate (see Nyquist). Also the noise shaping circuits in CD players were unrefined. This is no longer true today and even the most cost effective converters are capable of impressively wide and flat frequency responses and wide dynamic range.
My experience is that only the best analogue tape based systems can approach the frequency response, noise and dynamic range performance of CD quality audio and that vinyl is a severely compromised format.
There are those that maintain that restricting the frequency response of CD to the theoretical upper limits of the audio spectrum at 20kHz excludes higher harmonics which may impact the audio spectrum, but I can hear no difference between 44.1kHz and 96kHz audio encoded at 24-bit.
All the hi-resolution analogue and digital 2-track masters I have made are reproduced more faithfully by CD than both vinyl and pro-sumer grade 1/4" stereo tape machines (Revox B77 since you asked!).
Despite all this, I still listen to vinyl records every week. I like the sound of vinyl. But I no longer believe that what I am hearing is what the creators were hearing in the recording studio.
Please feel free to disagree, but not before conducting your own comparative listening tests.
In conclusion
I hope this article has helped you to consider the challenges of comparing analogue to digital and that you will be better equipped to carry out your own tests to determine both which is best and which you prefer.
Further reading ..
What is an analogue to digital converter?
How do analogue to digital converters work?