Your technical questions and queries answered
Is it safe to defragment hard drives containing audio?
What is the situation with defragmenting audio drives? I don't generally use long streams of audio for multitrack recording, preferring to drop samples into my sequencer (if you see what I mean). I'm sure that defragging would speed up audio access, but I want to make sure before I attempt anything.
Martin Walker replies: All hard drives get fragmented, simply because we all have occasion to delete files. Once there is a gap in the drive where a file used to be stored, the next time you save any file it may use this gap to store part of its data, and then store the remainder in another empty part of the drive. In time, the contents of the drive can therefore become extremely 'fragmented', and individual files may end up split into small chunks scattered about on the disk. It can take noticeably longer to read and write files on a badly fragmented drive, simply because the read/write heads have to jump about between the individual fragments, rather than smoothly moving through a single contiguous file.
If you run a defragmentation utility program, it simply finds every fragmented file on the drive, and then reorders and rewrites all the data more neatly. This ensures that existing files can be read in the shortest time possible, and that there are no gaps in the existing data to slow down future saves. In the case of multitrack audio recording a fully defragmented drive will not only be faster, but may let you squeeze another track or two out of your audio sequencer. The only difference between long continuous audio takes and shorter sample loops as far as fragmentation is concerned is that the longer your files are, the more potential they have to become fragmented.
You need have no worries about defragmenting any hard drive (audio or not), since many safeguards are built in to ensure that even in the event of a computer crash or power cut during the defragmentation process, no data will be lost. Since hard-drive performance is degraded if there are any fragmented files, some musicians take defragmentation to ridiculous extremes, running utilities after each and every audio take (just to make sure). This is taking things a bit too far, and once a week or even once a month is normally quite sufficient (depending on how many files you save of course, and how full your drive is). Using a separate drive (or a separate partition on the same drive) makes it quicker and easier to defragment your audio data on a regular basis without involving the operating system and application files too (although these will also benefit from occasional 'defragging' as well).
Various defragmentation utilities are available: if you have a PC you will already have the Microsoft one (in the System Tools folder of Accessories). The Windows 98 version not only defragments the files, but keeps a log of how often files are accessed, so that it can also re-order the files as well. The ones that are most frequently accessed are placed towards the outside of the drive, so that they can be read even more quickly. Files that are frequently modified need to be placed on the inside: this is because the drive fills up from the outside in, and therefore modifying a file stored near the outside will always cause a gap and future fragmentation.
The only disadvantage of reordering every file on a drive is that it takes much longer than simply defragmenting the files themselves. If your drive has never been defragmented it may take an hour or more to move the data, depending on how many gigabytes of data you have already stored on it. This is why many people use the excellent Speed Disk utility supplied as part of Norton Utilities, which works much faster. I 'defrag' my drives about once a week, and it never takes longer than about 10 to 15 minutes.
Why won't my computer recognise my MIDI Timepiece AV?
I've just bought a MOTU MIDI Timepiece AV interface, and I am unable to get my Apple Performa 630 to recognise it at the modem or the serial port. FreeMIDI scans for the device, but every time reports that it has found a standard interface, not the MTP AV. If I go along with this and select the Timepiece manually from the menu of interfaces, the software is still unable to locate the unit. I've checked leads, unconnected all the MIDI devices so that all I have running is the MTP AV attached to the modem port, and still no success. The switch to select Mac instead of PC is set correctly on the front of the unit. If it is attached to the printer port instead of the modem port, FreeMIDI reports that it can't access this port because Appletalk is on, even though it isn't. What am I doing wrong?
Simon Stock of Musictrack replies: The version of FreeMIDI you are using may not be compatible with the Performa 630: we have had reports that versions later than 1.35 have this problem on your particular model of Mac. If you need to get hold of a different version of FreeMIDI, please contact Musictrack directly.
If FreeMIDI reports that it has found a Standard Interface, it means that the System Exclusive 'handshaking' between FreeMIDI and the MTP AV was not completed succesfully. This can also be caused by a stream of MIDI data coming into the MTP AV interrupting the SysEx, for example if a master keyboard or drum machine that constantly outputs MIDI Clock is connected: check if any of the red input LEDs are constantly lit to eliminate this as a possible cause.
If you manually set the interface type in FreeMIDI then you must also set the correct speed under 'Interface Settings', as it defaults to 'Not in Use'. However, if the interface is not found automatically when FreeMIDI scans, it is indicative of a problem of some kind.
Finally, as an aside, unless you are using MOTU sequencing or scoring software, creating a full FreeMIDI configuration may be something of a red herring, in that other brands of sequencer (such as Logic and Cubase) do not support FreeMIDI, and therefore do not 'read' the configuration file. You do need to create a FreeMIDI configuration file in order to run the Console editing software, but all this needs to contain is the interface itself.
Can I do what I need to with samplers, and what equipment is right for me?
I am hoping to raise funds to purchase equipment for a home recording studio. With this I then want to write songs professionally for other artists as well as writing tracks for myself and a band I hope to form. I'm writing to you for advice on what equipment I should purchase. I am not looking at purchasing just what is cheap, but rather good-quality equipment that will serve my exact purpose, so I'm looking particularly at equipment which has been heavily reduced from its original RRP.
I am thinking of buying a Korg D8 digital 8-track, which can now be purchased for around £450, and Roland's VM3100 digital mixer, which looks a bargain at its current reduced price of £379. I realise that the D8 cannot record on all eight tracks simultaneously, but since I will rarely require this, does it really matter? Can I not also get around this problem by using a digital mixer with the D8 when I record a live performance?
I am also intending to buy a sampler, and would like to know more about what it is possible to do with samplers. If you're writing music using a sampler and say, for example, you sample brass, guitar or strings, how do you get the sampled sounds to play in the key of the song you're writing? Will a sampler convert the sampled sounds into the keys you want to use? Can you purchase long sample solos of guitar, bass, or brass, for example, and get these solo sounds to play over parts of your track in the right key? And are sampled sounds now so good and plentiful that one could put together tracks using the preset patterns, without having to rely on musicians? Finally, can a sampler take a track recorded by an artist and separate all the sounds on the track? For instance, say I wanted to sample Donald Fagen's voice without also recording the music in the background, or vice versa, could I do this with a sampler?
I'm not a very good keyboard player, but there seem to be many DJs mixing and remixing music successfully who, I assume, have little or no knowledge of music theory, nor can they play a musical instrument -- so how do they do it? What tools do they use? Is it a case of getting a sampler and CD samples and using the CD samples as you would preset patterns to build up a track?
Assistant Editor Sam Inglis replies: To answer the last question first, I think you're probably right that most of the DJs who make remixes have little or no training in 'traditional' music skills such as keyboard playing. It's also true that samplers are probably the most important tool in remixing and much dance music generally.
Samplers can manipulate samples in a number of ways on playback, including changing their pitch -- so you can, for instance, sample a note or riff and play it back a couple of semitones higher or lower to fit in with the key of your song. However, the further you change the pitch away from the original, the more noticeable the processing becomes, and shifts of more than a few semitones either way often sound very unnatural. Also, playing back samples at different pitches varies their length as well as their pitch -- so if you raise the pitch of a bass line, for instance, you will also make it faster. Some modern samplers will perform true pitch-shifting (ie. changing the pitch of a sample without altering its length) and time-stretching (vice versa), but this process usually has some audible side-effects.
You can use samplers in two basic ways. On the one hand, you can use them like a keyboard or sound module, where every note plays back a single note at the appropriate pitch (so, for instance, you might load up a set of electric piano samples into your sampler, whereupon playing a Middle C on your keyboard would make the sampler play an electric piano note at Middle C, playing a D would produce a D, and so on). In this sort of arrangement, your sampler would be loaded up with what is called a 'multisample': because shifting the pitch of a single sample by more than a few notes tends to sound unnatural, the sampler will actually contain lots of samples of an instrument at different pitches -- so while the C sharp above Middle C in the electric piano sample would just use the same sample as the C, played back at a slightly higher pitch, the C two octaves above or below would use a different sample. In general, you'd use this sort of sampler program if you wanted to play or record an electric piano part (or clarinet, guitar or whatever) but don't own a real electric piano.
The other basic way of using samples is what is sometimes called 'phrase sampling'. In this case, your sampler is loaded up with a whole set of completely different samples such as drum loops, short bass lines or synth parts, guitar riffs, vocal phrases and so on. You would then construct a track or parts of a track either by triggering these manually (each would be assigned to a different key on your keyboard) or, more usually, from a sequencer.
Dance tracks, including remixes, are usually constructed in the latter way. You can buy sample CDs for both purposes -- ie. CDs which contain multisamples of instruments playing single notes, and CDs containing various phrases and drum loops. There's a huge range of both available, and it's certainly possible to put together dance tracks just using a few samples from commercially available sample CDs. Most 'real' dance tracks tend to be a bit more sophisticated than this, however: either the creator will use his or her own samples or synth parts, or record some new vocals especially for the song, or use parts of an existing recording. They will also often use sample-manipulation tools such as filtering, phasing, flanging and other effects. However, as you say, it's certainly not necessary to use any 'real' musicians at all, if you have a suitable set of samples, and many hit dance tracks are created just using samples.
When it comes to remixing, the remixer will usually be working from the original multitrack recording of whatever it is that is being remixed, so he or she will have access to the component sounds (drums, bass, vocals, and so on) individually. This makes it easy for them to take the bits they want and replace the bits they don't. If you're sampling from a record, CD or other form which is already 'mixed down' into stereo, your options are much more restricted. In your example, for instance, if you wanted to sample Donald Fagen's voice, you would inevitably get the background music as well; there's no way to separate them. In general, if you want to do remixes, you really need to have access to the original multitrack recording, or at least to some of its component parts.
With regard to your equipment choices, the Korg D8 is a good digital 8-track if, as you say, you're sure you'll never need to record more than two tracks at once. The Roland VM3100 is, likewise, a good mixer. However, I'm not sure that they would make a particularly good combination: since the D8 has its own mixer built-in, it's not really designed to be used with an external mixer like the Roland -- in order to make much use of an external mixer, you'd need to have individual outputs for each of the D8's tracks, so you could send each of these tracks through a separate channel on the Roland mixer, but in fact the D8 has only stereo outputs. Having an additional mixer wouldn't allow you to record any more simultaneous tracks on the D8, though it would provide extra inputs which could be mixed together to be recorded via the D8's stereo input.
If you were to buy both, then, you would effectively be buying two mixers, which you wouldn't be able to take full advantage of, and which you might well not need. You could consider two other possible approaches. The simplest option would be to spend the same amount on just buying one unit which combines a mixer and hard-disk recorder (such as Roland's VS880, Akai's DPS12 or, if you can stretch to it, Korg's new D16). Alternatively, you could buy a mixer such as the Roland and a digital 8-track recorder which has no mixer section, such as Fostex's FD8, or a second-hand Alesis ADAT. If you do this, you need to make sure that your mixer and recorder are properly integrated: for instance, to properly interface Roland's VM3100 with a digital 8-track such as the Fostex or an ADAT, you'd need an additional converter box to convert digital signals from Roland's proprietary R-Bus format to ADAT-format.