User Review: Yamaha MD8
By Mark Mattson
June 1999

I. History

First and foremost, I am a live sound guy. Over the years, I've done some studio work, both tracking and mastering, but studio recording has never given me the same enjoyment as live reinforcement. I don't have a "project studio", and never intend to. As with many live engineers, I frequently record shows I mix, either as direct board tapes or through a stereo pair of mics at the mix position. Starting (many years ago) with a consumer-grade stereo cassette deck, I progressed first to an inexpensive 4-track cassette, which gave me better quality and more editing capability, and then to a quality 4-track cassette (Yamaha MT4X), which afforded me arguably the best there is to offer in a 4-track cassette format.

Occasionally I would be asked to provide finished tapes of shows to the performers, or even do low-budget demo tapes with this equipment. As much massaging as I might do, the quality and capabilities I desired simply could not be achieved. Four tracks forced compromises in what could be done in re-mix/mastering, and the quality of cassette, even at high speed, didn't rise to what people had become accustomed to hearing.

Ultimately I came to the conclusion that if I was going to continue doing recording of this sort, and if I wanted the results I could hear in my head, I would have to upgrade my equipment. Since sound work for me is, in the end, more hobby than profession, I had a limited budget to work with, and could not justify dropping many thousands of dollars on a full-blown remote mega-track studio. By necessity, this meant that 8, or at most 16, tracks would be within my budget. And so the search began.

II. The Buying Decision

What were my options in an 8 or 16 track recorder? ADAT, Hard-disk based (either PC or stand-alone), Mini-Disk, cassette, and reel-to-reel were all possibilities, with (in most cases) a number of options from different manufacturers in each of the formats. I ruled out cassette immediately as I already knew its limitations. Given that this was primarily for live club work, portability and transportability was an issue, so PC-based hard disk and reel-to-reel (also due to being tape-based) were discarded as options. Stand-alone hard-disk (such as the Roland VS series) looked very inviting, but for my purposes, the lack of individual track-outs was a huge omission. I did not want to be restricted by the internal processing capabilities of the unit, as comprehensive as they may be. The additional issue of limited recording time without dumping to external storage was problematic for my application, where I may only have seconds, not minutes, between songs to work with.

After discarding all those options as un-workable in my application, I was left with ADAT or Mini-Disk (MD). The ADAT is the "industry standard", and for good reason. It affords top-quality, portability, quick storage media change and is easily expanded for additional tracks. Being as wide-spread as it is, taking raw track tapes to a studio for mix-down is common. On the down-side, it is still a tape-based media, which means limited re-use of the media, limited storage time of the media, the very real potential for data loss due to tape errors, and on-going maintenance costs (estimated at $150/year) to keep the complex transport system in good working order. Additionally, ADAT recorders are just that, recorders only. You still must have a mixer to do anything useful with them.

On the MD side of the equation, I was essentially limited to one option, the Yamaha MD8, as no other manufacturer has produced an 8-track MD (although there are a number of 4-track units on the market). I've owned a consumer MD for a few years, and was familiar with the format, including using it for intro tracks for live bands, and recording live shows direct to stereo on it. The MD8 was very portable, allowed quick storage media change, includes a built-in mixer (and therefor unlike the ADAT is a stand-alone unit), and provided digital editing capabilities similar (although not as extensive) to that of a hard-disk based unit. The MD media is, for all practical purposes, reusable indefinitely, and has a storage time well-beyond my remaining life expectancy. On the down-side, the MD based units cannot be expanded as such - 8 tracks is it, the per-minute media cost is relatively high, currently about $1.80/minute (although the unlimited re-use needs to be factored into that as well), frequent disk changes are required, as each disk holds a bit over 18 minutes at 8 tracks, and what is purported by some to be the reason for not using MD, the ATRAC compression that is used.

No article about an MD product would be complete without some mention of ATRAC, but I will not attempt to delve deeply into it here. What I will say is that I will challenge anyone to listen (without prior knowledge) to material that has been recorded on any MD using version 3.5 or greater of the ATRAC , and identify it as such. In other words, while purists can (and will) continue to debate what they can and cannot hear (or think they hear), in the real world, on real equipment, a person listening to an MD recorded using v3.5 or greater ATRAC will have no idea they aren't listening to a top-quality CD. Following my research on the topic, the existence of ATRAC compression became, and is, a non-issue. The unfortunate thing is that current MD products are being judged by the quality of the original compression (v1.0) that was used when MD products first came on the market. To do so is no different than judging the quality of a current sampling keyboard to be poor, because the Ensoniq Mirage was.

The MD8 still posed a serious issue, lack of media portability. Given that few (if any) studios would have an MD8 in their inventory, I would be unable to take the media to them for mastering. After considering this issue, and its importance in my application, for some time, I ultimately arrived at the solution of eventually purchasing a DAT deck to master to. DAT is probably more common than even ADAT, and since I don't own (or intend to purchase) a CD burner, a very transportable media for getting the music on CD, should that be the desired goal. So having tossed about all the various issues, I decided that the best set-up for my application, was an MD8, a DAT deck (used), and a good supply (20 initially) of Data Disks. I was able to make the entire purchase for roughly $1,800.

III. Putting It To Use

At about the same time I was making my purchasing decision, one of the bands I engineer regularly (Shaman Spell) was making connections with promoters and record labels in an effort to move their cause forward. The people they were making contact with wanted to hear new material, as their debut CD (which I wasn't involved with) was now about 14 months old. The band had two new songs that were pretty well wrapped up, and two more that were about half way through development. A plan emerged to produce a 4-song demo CD for distribution to interested clients using my equipment. We had a time-frame of 4-6 weeks to get the project finished. The band members were leery about using the Mini-Disk format. They'd "heard" that the quality was poor. I tried to assure them that it wasn't an issue.

The first show I did with Shaman Spell after getting the MD8, I set it up to record direct, using all 8 tracks. One of the advantages of the MD8 over ADAT is that you can bounce down on top of existing tracks - you don't have to leave tracks open, so I could record a full 8 tracks, and bounce to 2 later. This particular band is 4 piece (hard rock) consisting of lead, acoustic, bass, drums, and vocals. I set the system up (utilizing both direct outs and subs from the mixer) with channels of lead VOX, backing VOX, lead guitar (DI), acoustic guitar (DI), bass (DI), kick, snare, and toms/cymbals. I essentially just left the on-board EQs (three band, swept mid) flat, got the levels hot without clip, and let it fly. Watching the elapsed time and changing the disks frequently was a bit if a pain with everything else going on, but not terribly difficult.

In reviewing the recordings, I was very pleased with the performance of the unit. Everything sounded great, although quite sterile to my "live guy" ears, as there was practically no room ambience and no FX. I had learned about the unit, gotten comfortable with it, and so my mission was accomplished.

Two days after the show, one of the band members called me to ask how the recording had turned out. It seems he had made contact with a new club owner that was considering hiring the band, but the owner wanted a demo tape. Could I put something together? By the next day? Being the glutton I am, I agreed. I set up the MD8, patched in the small FX rack that I use for "briefcase" jobs, and set up one Aux send on the MD8 for vocal FX, the other Aux send for some over-all verb to put back in some ambience. After some modest tweaking of the channel EQs, and getting a starting point on the channel levels, I patched the output to my MT4X (set for "consumer stereo" mode) and away we went. In about 4 hours I had transferred 2 hours of music to cassette. It was a "quick 'n dirty", admittedly, but it sounded phenomenal in comparison to anything I'd been able to do in the past with the MT4X by itself. I was quite pleased. The band member would pick up the tapes the next day and select a few songs to dub to the actual demo tape for the client, and do the dubbing himself.

Also during this show the band played a new song they had been working on. Although I liked the song, and felt it was worth working on more, I felt it needed to be tightened up and shortened. With the MD8 I was able to program a "cue list", setting markers at various points in the song (which are fully editable and can be moved back and forth until you're happy with the placement). I then used another function to create a "play list". On playback, the MD8 jumps back and forth to wherever you've programmed it seamlessly. In this way, I was able to edit the song (reduced the guitar solo, added a second chorus at the end, and so on) and dump it off to cassette so the band could hear what I thought the arrangement should sound like. Had I wanted, I could have saved this information to the MD as well.

Three days after that (Thursday) I received a call from the same band member. Plans had changed in a big way. A radio station that produces a weekly "local" show had called. They'd heard that Shaman Spell was doing some exciting things and new music, and wanted to schedule them for the show...10 days later. Suddenly we were on a very tight deadline. I had not yet purchased a DAT, there was no way the band could schedule studio time on that short notice. It was crunch time. The first decision that was made was to focus on the just the two songs that were pretty well wrapped, and set the other two aside. We'd go ahead and put just two songs on the CD. I set about purchasing a used DAT.

Over the course of a barrage of phone calls, it was decided we would use one of the band members' rural home to record. His large, hard surfaced living room would serve for tracking and detached, large dry garage would become the control room. Due to conflicts with day jobs, the band would rehearse on Saturday evening and we would record on Sunday. I got the DAT into my hands Saturday to discover some shipping damage, another minor heart-attack. Fortunately I'm a Pro-Audio tech and was able to make repairs. The unit seemed to be functioning well, although I didn't have time for full-blown testing.

IV. Trial By Fire

Sunday morning I arrived at the house at the appointed hour (8:00 am), to discover the band had decided to substantially re-write one of the songs the night before. This was not a good sign, but they assured me it was much better (I liked it the way it had been). The really bad news, however, was that band had also decided to get rather blotto after rehearsal, and were now quite hung over. I set about getting my gear set up while they tried to nurse themselves back to health. We were ready to start tracking at 11:00 am.

Going in, we knew that we had a limit of 8 tracks to work with at a time. Both the songs we were doing started with guitar, and no drums for about the first 30-45 seconds. This meant that rather than recording a full 8 tracks of drums first, we would be recording 2 tracks of guitar and 6 tracks of drums. Not really a problem, as such. It just meant that everyone had to be on their toes.

The drum tracks were laid down for the first song (the one they hadn't re-written) with nominal difficulty, and once we were satisfied with them, mixed down to 2 tracks. Knowing that once the drums were tracked, the band would be able to dismantle the drum set and open up the room, we proceeded to track the second song. The drummer knew his part, the guitar player was shaky. After a few false starts, they made it through the song with minimal errors. Utilizing the exceptional programmable punch-in capabilities of the MD8 (resolution is 86 frames per second) we went back and fixed the guitar players' parts. The MD8 performed wonderfully (once I figured out how to set it up for punching), even if the guitar player was struggling. We mixed these drum tracks down to 2 as well, and were ready to tear down the drum set and move on.

We proceeded through the various instrumental tracks, some of which required a fair amount of fixing with punches, others that didn't. Although the re-written song now sounded much better, the limited amount of time the band had spent on it showed. They were literally re-writing and learning parts as we did the punches. Around 6:00 pm a light rain began falling.

By 8:00 pm, we were ready to start the vocal tracks. At this point, we had mixed the bass and drum parts together onto 2 tracks, the lead and acoustic guitars were mixed into two tracks, and so we had four tracks open for lead and backing vocals. The first song went well. This one hadn't been re-written, so everything came together smoothly. I had 8 tracks on the MD, ready for mix-down and mastering. Outside, we had heard thunder approaching for some time, and as night fell, very large and intense thunderstorms moved in. We, of course, hadn't listened to a radio or TV all day, and were caught somewhat by surprise. With the deadline we were on, we pressed on. The house was well-insulated enough that the rumbles of thunder weren't creating large problems on the vocal tracks, since I was high-passing them at the mixer anyway.

During the first pass recording the lead vocal of the second song, at about 30 seconds from the end, a stroke of lightening flashed nearby and the power failed for a few seconds. Once everything came back on, I reset the digital gear, listened to the intro for the song on the MD, and everything sounded good. Still feeling the pressure to get this project done that night, we restarted tracking the vocal part from the beginning, rather than trying to punch in the tail end. About 1 minute into this take, the power failed for a few seconds again. Again I reset everything, got the gear ready to go and started from the beginning to get the vocal track. It was the last thing we needed. 10:00 pm was approaching by now. Just 5 more minutes and we'd be done tracking.

As we recorded this time thru, I stepped away from the controls briefly to stretch my legs. I thought I heard something odd in the monitors, but when I walked back over, everything seemed fine. About 2 minutes in, the power failed momentarily again. Clearly, this was one serious thunderstorm (we would later learn it was the worst in a number of years), and I was starting to get quite concerned about my gear, deadline or not.

I prepared to try to get the tracks again, but first wanted to listen to see if I had actually heard something the last go-round, or if it was just the snap of a lightening stroke. As the song played back, at about 1 minute, there was a very brief, very loud, digital "burp". Then again at about 2 minutes, and again near the end. Each time the power had failed, this digital "burp" had been put on the media. Very loud, on all the tracks (including those previously recorded), there was no covering it up. We were all devastated. Many hours of work had been destroyed and the deadline still stood. I told the band there was nothing I could do. We'd have to re-record the entire song, but there was no way it was going to happen that night. It was already 11:00 pm, they were all still nursing hang-overs, and most of us had to work early the next day. It was decided we would reconvene on Friday evening to re-track the song, knowing the band also had to play a show Saturday and be ready for the live radio broadcast on Sunday. No one was in good spirits. I left most of my gear there, but took the MD8 home with me to listen to what we had done.

V. Resurrection

By Tuesday evening I had enough energy back to sit down with the MD8 and listen. As I was listening, I was thumbing thru the manual, trying to come up with a way to fix the disaster. Suddenly I came upon what looked like a possible solution. Although it was probably not intended for this purpose, the MD8 allows you to copy a definable section of one track to a different track (or so the manual says). My problem was that I had to work within one track at a time. I sat down and started playing with it. Before long, I had programmed precise cut in and out points surrounding the section where the burp occurred. I went looking on the rest of the track and found a "good" identical section I could use for replacement. I put the process in motion, then went back and listened. It worked! I was ecstatic. The area containing the burp had been replaced with the good area, and I absolutely could not hear the break points. I went on to treat all the tracks in this way and eliminated the burps. While I was at it, I had noticed a section where the guitar player had chopped off a chord just a hair too early. I used the same method to copy 13 frames of his tracks forward to cover this error. Again, it was absolutely seamless. The wonders of digital technology had come to the rescue. In all I had fixed a total of 6 measures and 13 frames across all the tracks. The entire process (there was a bit of a learning curve here) took about 2 hours. Far less than re-tracking would have taken, with far less effort. While I was at it, I used the "song copy" function to duplicate each of the songs not once, but twice on each disk, just in case there were any more catastrophies. If I had taken the time to back up the work regularly from the beginning, a major headache could have been averted.

When I called the band the next day to tell them, needless to say they were very happy campers. We still had to get the last vocal tracks and do the mastering on Friday, but a tremendous burden was lifted. Compared to what we had been facing, this would be a piece of cake.

VI. Tying Up The Loose Ends

The session on Friday went well. Although everyone had worked that day, we were in good spirits. Tracking went smoothly, and we moved on to mastering. The first song went to DAT without any hitches, and everyone was pleased with the results on playback. The second song needed a little more massaging, possibly because it hadn't had as much time to "grow on" the band, but it came together as well. On listening to the playback from the DAT, suddenly there were sections that sounded terrible. The audio would drop in level and get fuzzy. I knew it couldn't have been a problem with the audio path itself, as I was monitoring thru the DAT deck, so I thought possibly just a bad section on the tape. We went back and re-recorded the master, but in the next area of the tape. On play-back, the same problem. Frustration starts to set in. A third try, it sounds even worse.

In desperation, I switched to a new tape, and at the same time, I "work" each of the controls a few times on the DAT before beginning, under the presumption that a control may have some oxidation on the contacts. This time everything works well. On playback there's no nasties, and we wrap it up.

Saturday morning, one of the band members took the DATs to an old friend of mine that has a project studio, and the ability to go from DAT to CD. He loaded the songs into Sound Forge, took a look, and proclaimed no need for any tweeking at all. In 45 minutes, 6 CDs had been burned, and the project was complete. The band declared the CD (with no prompting from me) to be "awesome", and I've heard no more concerns about the "quality" of the Mini-Disk format. As I'm typing this (Sunday evening), I'm listening to the radio show, and although I don't care for what the radio station compression did to the dynamics, it still comes across as some great music. I haven't even actually heard the CD yet, except on the radio. I can't wait.

VII. But Wait, There's More

Since this band has a tendency to write songs on the fly during shows, and then not be able to remember what they'd done after the fact, I had decided to start recording all of their shows. I took the MD8, two condenser mics, a stand, and cables to the Saturday night show. Using the first two channels of the MD8, which have XLRs with switchable phantom, as well as inserts, I set the unit to 2-track mode (there is also a 4 track mode), giving me 74 minutes of recording time per disk and recorded the entire show on 3 disks. Sure enough, they created a song late in the show, which I'll be able to dump to cassette so they can do as they please with it. Once I've done that, and reviewed the mix for my personal benefit, I'll simply erase the disks, and be ready for the next time around.

VIII. Conclusions

I'm sold. For my applications, I couldn't ask for more. There are editing and MIDI capabilities I haven't even touched on here, mostly because I haven't actually used them yet. This unit has already shown itself to be a very capable and powerful tool, and the issue of "quality" is a dead one. For somebody looking for a multi-purpose 8-track device that is willing to accept the (few) limitations imposed by the MD8, I recommend this unit without reservation.


If you have questions or comments about this article, I would prefer they be posted to the discussion group that led you here, for the benefit of all concerned. I will respond to emails, but I will not respond to, or acknowledge, emails requesting information about the MD8 which is available from your local Yamaha dealer.

If you want, you can Email me here.

Pro Audio/Electronics and Stuff