Ideas for using MD for live multi-track recording.
The venue where I record typically supplies a mono desk mix (or a least not a very wide stereo mix). In this case I will record the desk signal onto one channel, and on the other I use a shotgun mic at the back of the room (the ability to adjust the pan of the input level on the MDS-E10 makes balancing the two different input levels reasonably easy).
As a live recording, you may find the desk mix a little dry and the room mic of course way too mushy and distant. When you get this two track recording home, it's time to mix down. If though, you mix these signals together with a little bit of spatial pan, you will find you get a reasonable stereo mix. You will find different parts of the spectrum fall on different channels, so that it will sound like a real stereo mix, rather than a lop-sided mix of two mono signals. You also get to hear the response of the audience (if this is a good thing) and it can also correct the imbalances found in sound re-enforcement (eg. Vocals are louder on the desk because they need to ride above drums and guitars which don't need to much aplification). One thing to note is that in large room, you may a few milliseconds gap between the desk and the room mic. In this case, I used a delay to rectify the problem.
But what happens if you want to record more than two channels with MD? You could go and buy one of those multi-track MD units, but you will find they are more like a Porta-studio than a unit for multi-track mastering. The multi-track MD units use the MD Data format and also may not have a digital line out. My goal in finding a multi-track MD solution is to be compatible with MD Audio and also be able to get coaxial or optical transfers to the PC for mix-down.
As I own an MDS-E10 rack unit and an MZ-R501 portable, the simple solution was to record on both units at the same time. While this is a cheap solution for those who, like me, own a few MD units, there are issues which must be considered.
At the start of the recording you will need a short sharp sound which appears on at least one track of both recordings. Once you get the signals into the PC, you can synchronize the tracks (by trimming the files to the same crest of sharp wave form). Another issue, which for you may not be an issue, is time shift. Its not practical to use two cassette desks as a multi-track recorder because even under ideal circumstances using identical recorders, one deck will run a fraction faster or slower than the other. Even the smallest difference over time will multiply and the signals will get out of sync. As MD is a high precision digital format, these issues are not so noticeable. I ran a simple test where I recorded onto two recorders at the same time, a few mins of blank audio with a sharp noise at the start and end. I then used one of those recorders to play back both into my PC via a digital link. I then measured the gap between both of the sharp sounds down to a sample level. I could not determine any difference. While this is promising it is far from conclusive. Different brands and recorders may act differently. You also may get away with playing back on the same machines which did the recording (if it records a little slow it will play a little slow also, negating the error).
Other things to consider are the type of music you are recording and the amount of spillage between machines. High spillage, which is slightly out of sync, can cause phase cancellation when mixed.
You may not want the Philharmonic Orchestra to get even a single sample out of phase but dirty, sleazy rock and roll, that's a different story.
Another important thing to remember is that a portable recorder won't have a very accurate level meter. I found a good record level by running tests using a small mixer. I refer the peak levels on my mixer to determine if I am clipping the recorder. Not highly accurate, but it shouldn't cause you too many problems if after this, you still give it a bit of headroom.
Oh, and don't get excited when you find that your recorder has some kind of sync connection. They are not very accurate (well, not accurate to the phase of a waveform). Using this, or the same remote control on two machines may help keep things relatively close, but manual syncing will still need to be done.
I plan to record an up coming project this way using two MD units and by recoding the MIDI instruments by logging the control signals, rather than as audio. If you use the MIDI instrument to create that sharp sound, you can re-record the MIDI later from the sequence and still be in synch (if your MIDI recorder is accurate). This gives you finer control of the recording process but depending on your instruments and setup as a whole, may not be practical, or repeatable.
I would not recommend anyone wanting to get into multi-track recording go out and buy a pile of MD recorders. These solutions are just ideas for people who already own the equipment. The disadvantages are that it's fidgety and time consuming to assemble your mixes. The advantages are that it's cost effective if you already have the gear, and that the entire event can be stored on a few cheap MDs. While a dedicated hard disk recorder is technically easier to record with, it may be difficult and time consuming to archive your recordings.
After I record a gig on two track, I have a quick listen. If it's usable I will do a mix down, if it's not so good, I just throw the disc into a box. I have recorded 30 live recording in 12 months for under $100 Australian ($3 per disc). No need to be burning CDs from a HD multi-track.
The Dream Machine: I would love it if someone built a dual deck MD which used the same master clock for both decks to allow for perfectly synchronous 4 track recordings. Another solution is a machine which records 2 (or more) Stereo tracks at the same time.
You could also use this deck to make perfect (ATRAC data) copies of MD tracks. I know what Sony will not like this idea as ATRAC data protection is part of their copy protection `philosophy' (monopoly?), but why not allow it on such a machine if the copy prohibit bit is not set. It's not set on anything I record.