How to Rip Cassette

Guide by err0r - Posted here for safe keeping

There's no one true method to ensure perfect cassette rips, but there's certainly many things that you can do to ensure your ripping adventures start out on the best footing. This guide lists what I consider to be a good foundation, plus some tips on how I process my ripped audio.

Section 1 - What you will need

• A cassette that you want to rip.
• A cassette deck or player with line out capability - please do not use the headphone output
• A soundcard with line in capability (just about all of them - can you see a socket labelled with a microphone or arrow in? You're good!)
• An appropriate audio cable to connect the output of the cassette player to the soundcard line-in - this will probably be 3.5 mm stereo jack (soundcard) to RCA plugs (cassette player), but some cassette players will use other connectors such as 3.5 mm stereo jacks or DIN plugs. Check and purchase the appropriate cable

Software to record your rip. This could be as basic as Windows Sound Recorder (though I recommend against it!) to as complicated as a DAW (Digital Audio Workstation). Audacity is a good compromise! I use Ableton Live Lite 7 that came with my soundcard.

Section 2 - Ensuring a good rip

There's four main things for a good cassette rip. They are in rough order of decreasing importance:

1. A good cassette. And by good I mean preferably new. Giving a tape a few spins in a hi-fi deck isn't going to destroy it, but the less plays it has had and the less exposure to heat, dirt, and sources of magnetism (like a tape head!) the better the signal on the tape is going to be. Personally, the first time I play my tapes is when I rip them.*

2a. A good tape deck. As the vinyl aficionados will tell you, just because you can rip an analogue source with your equipment doesn't mean that you should. Luckily cassettes don't require preamps of matched impedance or fussy selection of deck, tonearm and cartridge - just a single deck of adequate quality.

What counts as a good deck? Opinions will obviously vary, but in my opinion a 2- or 3-head deck made between 1980-1990 is a solid bet. If you want to record tapes, stick to 3-head systems, otherwise 2-head systems are fine for playback. A good deck should list features such as bias/tape selector and Dolby noise reduction, whilst features such as tape auto-reverse are unnecessary and can even be problematic, so try to avoid them. Rather than fussing over particular brands and models, concentrate on getting a deck in good condition, e.g. stored inside and rarely used. Old or well-used cassette decks can suffer from a number of problems. In order of decreasing importance they include: deteriorated rubber belts & rollers; internal dirt & oil build up; head wear; and magnetised heads. Some things like dirt on the heads or head magnetism are easily fixed by you, others (particularly worn heads) may make a deck uneconomical to repair. Buyer beware! My own deck is a Sansui D-550M that I picked up off eBay for about $60 and it came in perfect condition. Sometimes you just get lucky.

2b. A corrected azimuth. Azimuth refers to the angular relationship between the cassette tape and the recording/playback heads. When a cassette is recorded, this angle (ideally 90 degrees) is determined by the positioning of the head used to magnetise it, and can differ between machines. If played back on a machine where the azimuth is aligned differently, the cassette sounds muddy and the high frequencies are attenuated.

On some decks it is possible to manually adjust the heads using a screwdriver. During playback, gently turning the screw supporting the playback head will rotate the azimuth and noticeably modify the quality of the sound. Locating the perfect azimuth can produce a vastly superior cassette rip.

Consult the following guides for information on how to do this:

• What is Azimuth ? An Explanation
• Endino's Cassette Tranfer How-To
• Tutorial - How to optimize a cassette tape output

3. A good recording environment. What am I talking about? I'm talking about a recording environment that's as free as possible from electrical and RF noise as possible. This is important because as ripping cassettes is an analogue process your recording will pick up whatever electrical noise is around. I was going to rate this condition as being less important than a good soundcard, but trust me, high noise or pops and clicks are much more objectionable than mild distortion or non-linear frequency response!

OK, all well and good, but how do you test how much electrical noise is present and how do you minimise such intrusions? To test for environmental electrical noise hook everything up (cassette deck -> soundcard) and turn them on, then start recording silence (i.e., leave everything on but don't play a tape) for 30 minutes. Have a listen to the resulting recording and check out the spectrogram - there will be a low level constant hiss, but it shouldn't vary much and there should be no pops or clicks. Any variation in the noise recorded means that you're picking up electrical interference from somewhere. To minimise electrical intemperance, avoid setting up your equipment near anything with a motor in it (particularly big motors) such as air-conditioners, fridges, room fans etc., avoid recording whilst there are electrical storms about, and avoid turning on or off electrical equipment near to - or on the same electrical circuit as - your computer and tape deck. If your electricity supply is subject to frequent brown-outs, i.e. your lights often dim for a fraction of a second, then plugging your computer and tape deck into a UPS (Uninterruptable Power Supply) is a good idea.

4. A good sound card. Most soundcards (including onboard sound) released in the last 3 years are OK, but stepping up to something like one of the Creative Audigy or X-Fi cards, or something from the ASUS Xonar range will improve things. These will all record at qualities up to 24-bit 96kHz and greater than 100dB SNR. There are bigger and better cards out there, but any of these are more than enough for great cassette rips. I personally use an ASUS Xonar D2X that my wife bought me for my birthday. Who's a lucky boy? Me. ^_^

It's worth getting to know what your soundcard and recording software can support. All soundcards & recording software will support recording at 16-bit 44kHz, but some will allow you to use ASIO (Audio Stream Input/Output), JACK or CoreAudio device drivers, or to record at higher bit-depths. Read your manuals. Note that there's little point in using sample frequencies higher than 44kHz - most cassettes won't exceed this bandwidth.

Remember, in the end just plunking a tape into any player and running the line-out into any soundcard's line-in will give a recordable signal and for rare tapes people may be willing to overlook quite a bit just to hear the music. But that's no reason not to try and do the best job that you possible can! And considering that less than $100 can turn a craptastic set-up to quite a good set-up that will stand you in good steed for more than just cassette ripping I don't see why you would settle for second best by choice.

Despite new tapes being preferable, it may be the case that you have just discovered a rare classic hiding away in a dusty attic or the like that's twenty years old or more. Look out! Sometimes very old tapes can suffer from mold or oxide shedding. Mold will show itself as discolouration of the tape, and as well as sounding bad it can cause the tape to stick to itself inside the cassette shell - if you try and play such a tape it could very well snap. It's safe to say you should seek another copy if this ever occurs. If you desparately want to rip this, for personal pleasure or for great bounty you should really be talking to a tape restoration professional. Oxide shedding is far less common but has the ability to really gunk up your machine. If the tape seems to be shedding "glitter" or flakes of brown gunk don't play it - again, take it to a restoration professional. These are the worst things that can happen though - your old tapes may play and sound just fine.

Section 3 - Audio processing

Audio processing can be broken into three parts - recording, sound processing, and track splitting. Let's tackle the recording stuff first.

Recording.

As alluded to earlier, 16-bit 44kHz is perfectly adequate for ripping cassettes. If you can use ASIO mode (Windows), JACK (Linux), or CoreAudio (Mac OSX) then by all means do so. If your soundcard supports recording 44kHz audio at 24-bits then it will help to do so, just don't worry if you can't. In whatever software you are using to record, set your recording gain sliders to maximum. These may be labelled line-in or microphone. The next bit is important, but once set won't need to be changed again. Using a loud section of tape (the loudest you can find ideally!), adjust your cassette deck's output level so that the recording level meter is as loud as possible without hitting the maximum level (0dB). In a perfect world this level would correspond to about 3/4 of your deck's maximum output level (i.e., dial or slider at 75%), but it's quite likely that like my decks, the maximum output level comes nowhere near 0dB, in which case it's best just to crank it to maximum and be done with it. For those pros keeping track at home, for a half-way decent deck it is most likely that the inherent noise in the system is more objectionable than any distortion caused by cranking the deck output to 100%, so maximum signal to noise is the name of the game. Of course, you should try a range of levels (and noisereduction settings) and see what works best for your system. My Sansui deck sounds best at 100%, my Nakamichi is best at 75%.

Some tapes may display the Dolby symbol, a logo akin to two Ds, one reversed. This means you will need a machine with Dolby noise reduction, and to engage it upon playback. If your machine has multiple options for Dolby noise reduction (e.g. Dolby B, C, or S), then you should set it to Dolby B.

Now to actually start recording! For this, turn everything on, start your program recording, and then after two to four seconds start your tape. Keep recording until the tape comes to a complete stop, then hit stop recording in your recording software. Why do we start recording before we start the tape? So we can have a quick check of how much noise is due to electrical interference and so we have a clear "event" that signifies the start of the tape proper. Your recording software should save the resulting rip as a WAV file somewhere - we now need to process this audio to bring it to an uploadable condition.

Sound processing.

Sound processing is the art of making the recorded audio sound as good as it possibly can. This involves things such as maximising the album volume (normalisation) and reducing the level of tape hiss (noise reduction). Technically any audio editing software could be used for these things, but for the purposes of this guide I'm going to stick with the program I use for my rips: Adobe Audition 3.0. You can find this program on What.CD either as a stand alone program or as a part of some Adobe suites.

• Step 1: Open your raw audio file in Audition (or editor of your choice) and set your display to waveform (that squiggly line)
• Step 2: Now we want to normalise the music. You can do this manually by applying gain or amplification until the biggest peak in your waveform is at least -3dB but doesn't go above -0.5dB, or you can do it automatically by using the normalise command (Audition - Effects - Amplitude and Compression - Normalize (process)...). I personally like to do it manually as it allows me to use the same gain setting for both sides of the tape. That way if one side was meant to sound louder than the other that dynamic is retained. For me, I generally have to apply 10-12dB of gain to my rips, but a few will need more and a few will need as little as 8dB - it all depends on the tape.
• Step 3: Find a section of blank tape at least four seconds long and capture a noise print (Audition - Effects - Restoration - Capture Noise Reduction Profile) of this section. The blank tape I'm talking about can be found at the beginning and end of each tape side, between the audio component of the tape and the tape leader. We want to capture a noise print of blank tape and not of the tape leaders or audio component as this section will allow us to remove the maximum amount of tape hiss whilst disturbing the audio signal as little as possible.
• Step 4: Select the entire tape and apply noise reduction (Audition - Effects - Restoration - Noise Reduction (process)...). What settings to use here is the black box of cassette processing. I have mine set to the following: Noise Reduction Level - 75%, FFT Size - 8192, Reduce by - 24dB, Precision Factor - 10, Smoothing Amount - 1, Transition Width - 10dB, Spectral Decay Rate - 65%. This is fairly strong noise reduction and I rely upon capturing a good noise print to ensure it's only removing noise and not music along with the fact that my system has fairly low levels of noise in the first place. For your first rip it pays to try a lot of different settings for this and save each of the results using a good naming scheme that gives details of the noise reduction, this way you can listen to each of the different results and see which suits your style best. The trade off is between tape hiss which may be loud but easily ignored when listening to music, digital artifacts of the noise reduction process (a rapid low level squelch sound) which whilst fairly quiet are very objectionable, and reducing things so hard that the music is damaged. If you can't decide, try putting some samples up on MediaFire or the like and getting a Mod to have a listen to them.

For (much) more information on noise reduction, please see noisereduction.

Track splitting.

There's three different possible scenarios here - tapes that contain a single track per side, tapes that contain clear gaps of silence between multiple tracks per side, and tapes that mix tracks together or run different tracks hard up against each other. If you're lucky enough to be ripping a tape that contains a single track per side, then all you need to do is to trim off unnecessary silence from before and after the music. I personally trim things so that there's two seconds of silence before and after the visible spectrogram.

If there's multiple tracks per side but they're clearly separated by silence then after trimming off the leading and trailing silence as above one will need to cut and paste each track into a new separate audio file and save them. If the gap of silence between tracks was less than four seconds I just cut exactly in the middle of the gap, but if the gap is longer than four seconds I trim the silence to a maximum of two seconds before and after the music. A note when saving files in Soundbooth - make sure that when saving you check the bit-depth. Soundbooth automatically sets new files to 32-bit float - change it to 16-bit before saving. If you recorded in 24-bit, now is the time to save to 16-bit WAV.

The last scenario can be quite the dozy - tapes that mix tracks together can often be more trouble than it's worth to split into separate tracks, but if your tape lists a tracklist and the mix transitions don't take too long then it's worth a shot at splitting them anyway. Whilst the exact point of splitting will be somewhat arbitrary, with the ultimate choice being wherever you think sounds best, it's important your split occurs at a section where the waveform exactly crosses zero. To do this switch to waveform view and keep zooming in on the section where you want to make the split, then find a suitable crossing to make the cut at. If you don't do this both of the tracks resulting from your cut will terminate with an abrupt thump.