Vinyl to Digits

In my most recent post I covered the importance of cleaning a vinyl (or shellac) record before making a digital dub. At this point I’m going to assume that you have cleaned your candidate for digital immortality. Now it’s time to do the deed proper.

In a nutshell: to digitize a record, you play it while capturing the audio from the turntable with software, just like in the old days when we used to copy LPs to cassettes or open-reel tapes. But of course it’s never quite that simple. Certain issues might need addressing.

Turntables, Cartridges, and Phonostages

Garbage in, garbage out. Your digitized version is no better than the original source. Therefore, first-class digital copies depend on first-class analog playback. Digitize an LP by playing it on a cheapjack record player and your digitized version will sound amazingly like a record played on a cheapjack record player.

Three fundamental components make up a record playing system: the turntable, which not only spins the record but also includes the tonearm, the cartridge, which turns the wiggles of the stylus into electrical current, and the phonostage, which amplifies the cartridge’s output into a sufficiently strong signal for further processing.

Many turntables put all that stuff together into one handy package, giving you a ready-made solution that you can plug right into the AUX input of a home amplifier or the audio IN of a home computer. But the higher up the audio food chain you go, the more separate it all becomes. HIgher-end turntables do not come with cartridges installed; you buy one separately and either install it yourself or hire an expert to do it. The phonostage is a separate component in most higher-end systems, although many amplifiers include one as well (look for a “Phono” selector on your amplifier.) If you really sail up to the highest reaches, the motor, the tonearm, the headshell, and even those itty-bitty wires that connect the cartridge to the outside world are separates.

However you choose (and there are lots of choices!) make sure you’re happy with the sound of the playback system. If you don’t like it, you won’t like your digitized copies.

My solution: For LP’s, VPI Scout II turntable with a JMV-9 tonearm and Grado Sonata 1 Reference cartridge, plugged into an NAD P-1 phonostage. For 78s, Rega Planar 78 turntable with a Grado 78E cartridge, also plugged into the NAD P-1 phonostage.

Analog-to-Digital Converter (ADC)

For the analog sound from your turntable to become a stream of digits, it must run through an analog-to-digital converter, or ADC. Just about all computers contain an ADC—they can’t have an audio input jack otherwise. In most cases, computers take a standard stereo 1/8″ stereo plug, usually used for microphones but perfectly capable of accepting any high-gain input, such as the output of a cassette player, music player, or phonostage.

However, a computer’s built-in ADC isn’t anything to write home about. It will do the job, but not well. If you’re serious about creating fine digital copies, you need a dedicated ADC. Most connect to the computer via USB or Firewire, and most require software drivers. Once the ADC has been installed, it shows up as an Input source in your computer’s “sound” control/preference panel. Just select it and you’re ready to go.

My solution: Apogee Duet external Firewire interface.

USB Turntables

Present-day “USB” turntables combine all of the above into one package—turntable, cartridge, phonostage, and ADC. Therefore you can plug the turntable into your computer via USB, select it as your Input source, and you’re ready to roll. Keep in mind that you’re swapping convenience for better sound quality, or at least step-by-step control over the ultimate sound of your recordings. But that convenience is very compelling.

Software

Now for the software to use: there’s a lot of stuff out there. Open-source folks swear by Audacity, a free jack-of-all-trades program that has a lot to recommend it. DAWs like Logic Pro, ProTools, or Cubase make first-class platforms for vinyl restoration as well.

My own choice is Adobe Audition, part of Adobe Creative Suite 5.5 for the Macintosh. Here’s why:

  • Audition saves the recordings automatically while they’re being made
  • Audition offers non-destructive editing tools that are particularly useful for joining the sides of 78 rpm albums
  • Audition contains a first-class suite of processing aids, and works well with VST and AU plugins
  • Unlike LogicPro or Cubase, Audition doesn’t bother with MIDI recording or processing; it’s designed exclusively for digital audio.

Recording

The recording process itself is straightforward: start recording a track in your software and play the record. Provided you have chosen your ADC as your Input device, the output from the ADC will be captured by the recording software. When the side ends, stop the recording and then you have a complete track representing one side of a disc. Your recording will be uncompressed, full-quality audio.

An important consideration is the bit rate and word length to use when making the recording. Audio CDs are at a bit rate of 44.1 kHz, considered by many audiophiles as too low for optimal reproduction. Personally I think that a bit rate of 96 kHz is overkill. So I have settled on 48 kHz bit rate and a 24-bit word length. If I want, later on I can downsample to 44.1 kHz and 16-bit word length, for example if I’m going to make an audio CD of my files. But usually I just leave them at 48 kHz.

You adjust the bit rate at the system level, not at the program level. On the Macintosh, you use the “Audio MIDI Setup” program for that purpose. Windows offers something comparable, no doubt in the Sound control panel.

Post-Production

You’ve got the record digitized, but you’re not finished yet. Post-production comes next.

Trim off any leading or trailing space from the track

It only takes a moment, but it’s worth doing: strip off any leading space that is silent or records the thwump of the tone arm settling into the groove. Do the same with the end of the track. If your software offers non-destructive editing (you’re never applying your edits to the original file, in other words), I recommend that you use it. In Adobe Audition, applying the razor blade tool to the track as it appears in the Multitrack window is non-destructive. Actually opening the track (by double-clicking) and then selecting-and-deleting the section is destructive. Your choice.

Join individual tracks into longer sections as necessary

Typically you need to do this only with 78s, but once in a while you also come across an LP that splits a long movement across both sides. The process isn’t particularly difficult if you’re using software like Adobe Audition, Logic Pro, Cubase, or ProTools — all programs that record in blocks that you can move around at will. You slide those blocks of sound together, overlapping gently where necessary to achieve a smooth join, and trimming as necessary. It takes some practice and a lot of listening—and in some cases, you’re well-advised to have a copy of the score handy so you know precisely where that join should be made.

Optional but recommended: bounce the trimmed and joined track to a new track

If you have been moving segments around and trimming them, it’s best to “bounce” those completed segments to an altogether new track, i.e., copy the joined segments into a single track. Whether or not you’re working with joins, if you have been making non-destructive edits, you need to “bounce” to a new track as well.

If you are bouncing old tracks to new, then all of your subsequent post-production is applied to those bounced tracks.

Apply your choice of processing to the track

This one’s the 500-pound gorilla. Just how much noise suppression do you want? Some people insist that their dubbed LPs have the same noiseless background as a CD, and so they’ll use noise suppression vigorously to achieve that goal. Be advised, however, that noise suppression removes some of the original signal, no matter how sophisticated the algorithms. My processing software of choice is a suite of plugins from iZotope. I use the “Declick” module regularly; it removes clicks and pops without otherwise impacting the recording. But I don’t use iZotope’s “Decrackle” module as a rule, because it changes the audio; ditto iZotope’s “Noise Removal” module, unless something truly egregious needs fixing. In extreme cases I might open up the “Repair” module and perform some very specific edits, but that’s a rare occurrence for me.

I “normalize” each track in order to adjust the volume. When a track has been normalized it means that the volume has been raised so that the loudest part is at the level you specify, and the rest is proportionate to that. I normalize to -3 dB, to me a healthy high end that allows for plenty of dynamic headroom. Note that I would never in a million years apply dynamic compression to a recording — i.e., increasing the volume on the softer passages. Normalization is not dynamic compression.

Finally, I always apply a 60-cycle hum removal filter, since a low-frequency hum is endemic to all records, regardless of their speed.

Apply any frequency adjustments you think necessary

Older recordings might require some frequency-curve tweaking—but use a light hand. 78s in particular might come out a bit funny on modern equipment because they were recorded with the assumption that they would be played back with certain automatic frequency adjustments. The Macintosh program Amadeus Pro offers presets that are designed for recordings from particular labels during particular eras. They might help.

However, I always err on the side of caution and eschew audio tampering unless something is really bugging me. I’m particularly bothered by blares in the low treble region, so in the case of a particularly offensive recording I might drop the range from 4.8 – 5.2 kHz by a few decibels. But as a rule I leave the sound alone. It’s always tempting to tweak, but most of the time you wind up doing more harm than good. And with LPs made from 1951 onwards, you rarely need to do anything about frequency responses. If it was a cruddy recording to begin with, equalization isn’t going to help much. Lipstick on a pig and all that.

Split the long track into individual tracks (movements of pieces or individual songs, for example) if you want

It’s a bummer for your finished dub to be a big long file containing the entirety of an LP side. You might want split it into individual files, one for each track on the original disc. Although some programs purport to automate that process for you, by inserting splits in the silent parts of a file, as a rule they are more trouble than they’re worth. I use a program called Amadeus Pro that allows me to split a file into smaller sections based on the location of named markers that I have inserted myself. I got through the file (largely by ear but also watching the onscreen waveforms), put in my markers, and then tell Amadeus Pro to spawn off individual files for those marked segments.

Convert the track(s) into whatever digital format you might prefer—mp3, Apple Lossless, FLAC, whatnot

Lots of programs can convert audio from one format to another. For straight-up conversions I use XLD, which can convert just about anything to anything. But in many cases I have already taken care of the format conversion during the previous step when I split out the file into individual tracks, having Amadeus Pro store the split files in my format of choice (Apple Lossless.)

Last Thoughts: Using One-Stop Software

There are programs out there that automate most of the process I’ve outlined before: you start the recording going and they take care of the rest. Probably the best of the bunch is PureVinyl from Channel D Software. I tried PureVinyl and rejected it—I didn’t like the frequency curves it applied to the recordings, for one thing. But you might find something like that ideal for your purposes. Fortunately, most of those programs offer an initial trial period, so you don’t have to be out a pocketful of change just to find out what you don’t want.

This entry was posted in Uncategorized. Bookmark the permalink.