Rediscovering The Yamaha DMP7
Wilf Smarties takes another look at Yamaha's DMP7 digital mixer, and its DMP11 derivative. Has the rest of the world finally caught up with these pioneering products, with their extensive and cost-effective automation facilities?
Sometimes a product comes along that offers something completely new. If that product is also desirable (for whatever reason), it is sure to be a runaway commercial success. Or is it? When I heard about the advent of FM synthesis, my rudimentary background in physics, coupled with a fairly wide experience in sound engineering, enabled me to 'hear' Yamaha's DX7 long before I clapped eyes on it. In fact, I took delivery of mine from the first batch to hit the UK, having placed my order unseen some six months earlier. To say that the instrument did not disappoint would be an understatement in extremis.
The DX7 was one of the two most innovative products ever to come out of the Yamaha stable. The other was the DMP7.
...is a rather ludicrous movie starring big Arnie. It is also the name adopted by Solid State Logic to describe their top-of-the-market range of automated mixing consoles. However, the SSL monster, while offering dynamic automation of faders and mutes, when it comes to all the other knobs, switches, pots and pans does no more than offer you a system by which you can manually reset them to a previous status. A console with Total Recall can take a 'picture' that records every setting, and you have to adjust the console to match.
The DMP7, like many other consoles now on the market over a wide price range, offers mute and fader automation — but much more besides. All parameters can be dynamically controlled, and the DMP7 has an entirely digital signal path. Audio signals are converted into 16-bit digital before all mixing and processing — the DMP7 has on-board digital effects — and are converted back into analogue only at the stereo outputs.
I will divide the main thrust of this retro-view into two halfs. The first will discuss the implications of having a mixing system which can be reconfigured at will; the second will see how good the DMP7 is at doing the job. (I will also be looking at the RTC1 controller, and the DMP11, DMP7's little brother).
This does for mixing what non-linear recording does for recording: from a limited number of channels (or tracks) you can create the impression that many more are actually available. Consider analogue multitrack recording: music scrolls by on parallel tracks. The second chorus always appears after the second verse, and before the 'middle bit'. Different instruments are assigned to different tracks, and are then mixed through separate channels on a mixing board. Not all tracks/channels are used at once, often many less than you might think, but since each instrument needs its own EQ and FX setting, and as most tape tracks carry only one instrument (which might be a sax solo that only shows up in the middle eight), a 24-track recording requires at least a 24-channel mixer, and you also have to allow for FX returns.
There is, however, another way to record audio — random access digital recording. The first system to offer this was the NED Synclavier, initially with four tracks of hard disk recording, then AMS countered with their Audiofile. For several years these two companies had the high value broadcast and A/V markets largely to themselves. Now 8-channel hard disk recording is dropping below the £10k threshold with the advent of price/performance shattering products like Roland's new DM80, first in what is sure to be a line of cheap non-linear multitracks.
Instead of fixed parallel tracks, these systems work on soundfiles which can be called up at will. An 8-channel system will play eight files simultaneously. Where a few seconds before we had a sax solo coming out of output 3, we now hear a backing vocal, next up comes a tom-tom, and so on. With outputs as busy as this, the mixing requirement becomes much more demanding. Either you have to have each output connected to several mixing channels which are unmuted as and when the appropriate sound file is called up, or you have to have a mixing console that is able to respond dynamically to abrupt changes in program material. There are enormous advantages in adopting the second option. The only serious drawback has been the cost. The Trident DI-AN (as featured in a recent SOS article on Soul II Soul) is one of the very few consoles that offer this, but at a price somewhere above where the ozone layer used to be.
There is another environment where outputs have to handle a multiplicity of sounds, and this will be familiar territory to many of you: I suppose the next generation of samplers will have integral full-blown mixers, but for now they present the user with the same mixing problems as do hard disk recorders.
The prime benefits of having total dynamic control over the mixing process are:
1. A mix can be 'guaranteed' to be 100% reproducible.
2. Different instruments appearing sequentially at a particular input can have their own particular EQ, FX, pan and level settings. Note that any non-MIDI controlled outboard equipment used in the mix will still require you to take note of its settings.
3. EQ movements can be recorded in real time, eg. for audio sweetening or sound crunching. (I can see frequency sweeps becoming as over-used as digital reverb used to be).
In addition, the usual benefits of mute and fader automation of course apply. I hope I have now got you excited about the mixing potential of a system offering real time control over all parameters. Now let's take a closer look at the DMP7, which has been offering you just that for years. Maybe now you will be ready for it.
The DMP7 is an 8-channel stereo mixer. Most interested readers will have come across it at some time or other, but in case you haven't it sports 10 motor-driven faders (for channels 1-8, an assignable FX return, and the stereo mix), a motor driven data entry slider, and arrays of single or multi-function buttons. Over every fader there is an On (un-mute) button, and above this, in the case of the eight input channels, a button called Select. Only one channel may be selected at a time for editing (unless you use the RTC1 or some other controller), except for mute and fader movements, all of which are simultaneously accessible. The selected channel can then have its EQ trimmed by using the EQ band select buttons and parameter cursors situated on the right of the control surface. Though tedious to program from the DMP7 itself, the 3-band full parametric equalisers are powerful and sound good. (There are nine: one is provided on the return of aux 3, which could be a godsend for those looking for filter-swept delays — and who isn't?) There are shelf/bell switches for the upper and lower bands, +/-15dB of cut and boost are provided, and unusually wide Q and frequency bands are offered. (Though not as unusually wide as on the DMP11 — see below).
All digital signal processing is carried out at 24-bit resolution, save in the EQ section which, as a consequence of its demanding nature, requires 32-bit processing. Data conversion and transmission is 16-bit.
Beyond equalisation, the DMP7 has provision for three auxiliary effects sends. The first two of these feed the two identical Yamaha digital effects processors that are built in to the DMP7. Most basic time domain FX are catered for with perfectly usable fidelity (imagine an SPX90 with 20k bandwidth), though there are no harmonising algorithms as found on just about all later Yamaha DSP units. Aux 3 feeds an even more basic FX menu, but can more usefully be routed to the outside world for external processing. Effects programming is as on similar rack-mounting units such as the SPX range, and needs no further comment from me. Setting up an Auxiliary buss mix does, however. This is a peach thanks to the ingenious 'Fader Flip' facility. With this active, the faders of channels 1-8 control the aux sends to whichever of three aux busses is selected, and the aux return ditto. Neat, and very visual. If only EQ select could have been made to call up EQ parameters onto the faders in an analogous manner! Oh well, that's where the RTC1 came in, to sort out the EQ ergonomics.
Other lesser spotted features of the DMP7 include phase reverse on every channel, a stereo compressor on the output, pre/post aux selection, solos all round, and automatic fading.
Connections are all made at the rear. Inputs are via eight quarter-inch mono jacks, each of which has a trimpot (+4 to -20dB) immediately and inconveniently over it. The effects loop uses three similar jacks: one for the send, and a stereo pair for the returns. Stereo mix is offered via unbalanced quarter-inch jacks or balanced XLRs. That wraps it up for the audio, but there are also the three MIDI sockets, ports for 'Digital Cascade' whereby two more DMP7s can be linked digitally, and a stereo headphones jack. At first sight the footswitch socket might seem a little redundant, but it is for a controller and not a switch, and its function (normally a memory-independent overall level control) can be exchanged with the current data entry slider parameter. By this device the cunning programmer can turn a foot pedal into a 'wah-wah' (among other things).
There are four decent tri-colour 8-step LED meters. The stereo mix occupies two, another shows the selected aux send level, and the remaining meter monitors the selected channel's input level. Care should be taken with the latter, since the onset of digital clipping is not a pretty sound. A little red should be OK, however.
Pan positions are indicated by a row of eight 3-LED displays, and are accessed by hitting the master Pan button, selecting the required channel, and entering a position via the Data Entry slider. Four large Status LEDs let you know whether you are in normal channel, or one of three auxiliary send, modes, the latter accessed via Fader Flip as previously described.
The manual is not over large, and is writ in three languages, so you will not have too much swotting to do before you get down to work. It is clear and unambiguous, and includes much detail regarding MIDI implementation.
Here the DMP7 is in a price/performance league of three (its only competition being the DMP11, and Fostex's new DCM100, a line mixer with eight stereo inputs which also allows total dynamic automation of all parameters, though its EQ is very basic and it has an entirely analogue signal path with no digital effects. See last month's SOS for a full review).) There are two main ways of controlling mix automation: static (snapshot), and dynamic (ie. continuous real-time parameter changes).
There are 30 user memories for whole desk set-ups on board the DMP7, and a further 67 available in the (free) plug-in RAM (à la DX7). You can recall memories over MIDI by program change commands. I am not personally a big fan of program changes per se, unless I can write the contents of the programs onto my song disks: I like to 'burn in' current parameter settings. (Emu's Proteus is the master of this art.) Among other MIDI bulk dump features (most usefully controller/parameter maps: see later) is one which stores the contents of all 30 programs into a sequencer or data filer. I would have liked to have been able to store current desk status as a SysEx snapshot, which has the added advantages that no program change command is required, and it only takes 1/30th of the memory! Though the DMP7 cannot directly do this, an equivalent routine is available in reverse using Cubase in MIDI manager mode, more of which later. [Note from Yamaha: It is possible to emulate the 'Request' command from the RTC1 which is the equivalent of recalling memory 127. A DMP7 receiving this command will output the current desk status.]
Dynamic automation is, of course, the DMP series' forte. There are 206 variable parameters on board, and all are recordable over MIDI, (though you can record 'only' 128 simultaneously: see later). You could record an EQ sweep onto a sequencer track, then edit that information to make the sweep ultra smooth. The possibilities for creative and fine-tuned mixing are legion. Most MIDI digital FX give you the option of switching between programs and editing parameters in real time over MIDI, and the DMP7 does too. Not enough of you explore these wilder regions of the MIDI landscape I fear, to your eternal disadvantage. (I'll bet it has never even occurred to the vast majority of MIDI synth owners to record edits in real time).
On the negative side, the cost (in hard cash and signal degradation) of installing large numbers of A-to-D and D-to-A convertors has resulted in a console with no insert points. For example, if you want to use an analogue compressor on a vocal track, you must apply it before the signal enters the desk, ie. pre-EQ, not always the most desirable order of play.
The specified dynamic range of the earlier DMP7s has been upgraded from its original 88 to 92dB, which is of course 4dB worse than CD. The stereo fader takes the mix noise down to virtual silence, though. Subjectively the noise floor seems to lie somewhere between DAT and a good analogue tape machine at 30ips. Unless you gate or fade/mute the stereo output, you will not be able to go from "I can't tell if it's switched on" to "Ouch" during a typical mix. However, you will be pushed to hear any noise when music is playing, provided that you have optimised input gains.
Recording fader and pan movements into a sequencer is straightforward. Replay duplicates the movements made, and a mix can be built up with overlays, controller information edited under the microscope etc. Experienced sequencer/synthesizer/sampler users will be in familiar territory: many modern instruments are pannable and fadeable over MIDI anyway, though at the time of the release of the DMP7 this was not the case. However, real-time equalisation is something else again...
Adjusting EQ frequency or Q is slightly 'steppy'. Filter sweeps are glissando rather than portamento. In the midst of a busy dance beat I found this limitation to be virtually un-noticeable. The slower the sweep, the smoother it sounded, not surprisingly.
At rest the EQ is certainly powerful, each band having a range and scope beyond that of most conventional equalisers, at least with regard to Q and frequency. Using the RTC1 to control a channel's EQ over MIDI resulted in a credible simulation of the operation and feel of an analogue parametric. However, when I first tried recording this MIDI data into a sequencer and then playing it back, the slight glissando noted earlier became discontinuous nonsense. Upon further investigation I found that the RTC/DMP7 combination has too many variable parameters (206) for the 128 controllers in the MIDI spec. (The DMP7 is a single channel device). The default set-up here has note on/note off data controlling all EQ parameters, and note off data is reinterpreted by virtually all sequencers these days either as note on velocity 0 or, more usefully and usually, note length. Perhaps this MIDI data conversion flummoxes the DMP7. When I tried to get Cubase MIDI Manager objects to 'learn' note data from the DMP7, the Status Byte was found to be missing from the transmission — another possible explanation for the problem? Whatever the reason, the bottom line is that you cannot use note commands to control parameters requiring dynamic automation via a MIDI sequencer. It is not even possible to use notes for switching functions like muting. (I have Cubase 3 and an MC500. I expect a similar result would be found using other sequencers).
How many DMP7 users have never recorded an EQ sweep successfully because of this? If only they had read the manual. It is a fairly straightforward procedure to remap the entire 206 parameter set to something more sensible, using controllers where continuous sequencer playback of data might be required. Best leave the most boring parameters (Phase, Solo, perhaps Q) either non-automated, or perhaps consigned to switching via memory location program changes (though this would destroy an ongoing dynamic mix).
The RTC 1 is a remote controller capable of driving up to four DMP units. Its most exciting feature is the inclusion of nine rotary data encoders giving immediate access to all EQ parameters of the channel currently selected. For some reason these are not of the 360 degree type, necessitating the use of null point indicators. I suppose this means that they appear even more like conventional EQ knobs. Buttons are of the type found on the DMP11 (see later). The RTC1 enables fader subgrouping over up to four DMPs. Some multiple DMP users find this feature alone justifies its expense.
Other than the fact that it is a remote, giving programming access to equipment that is out of reach, there is not much more that the DMP's own control surfaces can't handle as well (or better). An example of a downside to the RTC: selecting a channel via the DMP7's channel select buttons calls up the input onto a meter. Not so via the RTC.
The RTC1, unlike the DMP7, has no provision for parameter/controller re-mapping. When I found this out, it temporarily dropped off my shopping list. Though I loved the feel of the EQ controls, they are of next to no use to me if they can't be recorded into a sequencer. (This is not the case when it is controlling a DMP11: see later). Hence the need to create a MIDI manager to do what the DMP7 does least well: EQ ergonomics. FX editing is no better or worse than on your average 19" unit. EQ cannot be subjected to the same ergonomic constraints and survive. I have developed a simple but effective MIDI Manager template on Cubase, giving controller access to eight channels of level and pan, Q, frequency and gain for each of the three EQ frequency bands, and channel mutes. This makes editing the DMP7's 'conventional' mixer functions much easier than when going through the DMP7's own cumbersome hierarchical menu (with the exception of the faders and mutes which are all simultaneously accessible). I will construct a further MIDI Manager template specifically for master out, returns and FX programming.
Here I will dwell on the differences between the budget DMP11 and the DMP7. Though the DMP7 can be rack mounted, it is primarily designed as a flatbed mixing console. The DMP11 has no such pretensions. In appearance it looks like a REV 7 with an extra 2U of faders below. Styling is considerably more basic than its big brother, and so are the facilities in some respects.
There are two, not three, aux sends, and two, not three, frequency bands in the parametric EQ. The upside is that both bands are sweepable throughout the entire audio spectrum, from 32Hz-18kHz centre frequency. Show me the analogue parametric that does that! Aux 2 connects to the outside world, as does aux 3 on the DMP7. Effect 1 is similar to 1 & 2 on the DMP7, though the inclusion of three harmonising programs is a welcome addition. Effect 2 is, as you might expect, super basic.
One really annoying omission is that of the stereo EQ-able FX return of the DMP7. Here, if aux. 2 is feeding into an external processor (which it probably will be), the return would have to come up through one (or two) of the eight input channels, or summed with the output of the DMP11 via a second mixer.
They may look inelegant, but having eight Clip LEDs monitoring all inputs simultaneously is a vast improvement on the DMP7, where only the selected input is metered. Also, I wouldn't swear to it, but I think that the onset of clipping in the DMP11 is less drastic than on the DMP7.
Since the faders do not move there is no need for the manual/auto switch which stops the DMP7 faders from flying while you are updating a mix. Here a null point must be crossed before updating occurs, as on the RTC1 EQ controller. Fader positions can be monitored as vertical bars on the LCD, in normal and 'Fader Flip' modes.
The first thing I did with the DMP11 was to try recording an EQ sweep into a sequencer. It worked first time. (Shock!) Controller information only is used here, albeit over two MIDI channels. The second thing I noticed was that the sonic performance was better than on the DMP7. This was not simply because continuous input level monitoring made it easier to optimise signal level on the DMP11. A careful A/B test showed a noticeable discrepancy in fidelity between the two units in favour of the DMP11. In particular, the DMP7 seemed to smear out transients, whereas the '11 sounded harder, more 'time aligned'. Subjectively, the DMP11 was damn close to CD or DAT in its sonic precision.
Further research into the reason for this sonic anomaly revealed that the DMP7 uses earlier DEQ1 chips, whereas the DMP11 has Yamaha's current DEQ2 chips installed. (Yamaha suggested that the DMP7 I was testing may have been out of spec, as both DMP7 and DMP11 should sound identical. My answer is: the difference was initially not obvious; and that a major DMP7 user in Edinburgh also thought the DMP11 more transparent. Also, the claim that the DEQ1 and DEQ2 chips are sonically identical might be correct, in which case the perceived difference may come from the way data is handled post-DEQ processing. As mentioned before, data transmission is 16-bit throughout for the DMP7. In the DMP11, data leaving the DEQ2 chip does so at an improved resolution of 24 bits. I would be interested to hear other DMP users' thoughts on this matter.) Incidentally, when the DMP11 was released it was able to take advantage of better A-to-D and D-to-A convertors than were available to the DMP7. It is possible to retrofit these into older DMP7s — contact Yamaha for details.
Uglier they may be, but the DMP11's programming keys have a better tactile response than those on the DMP7. And while the moving faders of a DMP7 can feel decidedly sticky, those on the DMP11 offer smoother and more precise control.
Here's what I plan to do. My Roland S770 and SP700 (the forthcoming playback-only version of the S770/750) will be parallel feeding into 16 channels of my Tascam MM1 mixer (to gain access to its four aux busses) and a pair of Yamaha DMP11s, via a normalising patchbay, giving potential insert points. In this configuration the MM1's aux sends would be pre-fade relative to the DMP11s, which is not usually desirable. However, by attaching a 16-channel MIDI-VCA automation (10 a penny these days) to the inputs of the MM1, synchronising its MIDI controller-channel fader assignments with those of the DMP11s, and driving the whole shebang from Cubase MIDI Manager, the MM1 will be able to provide four virtual post-fade effect sends to augment the meagre two of the digital mixers. Beat that for a cost effective and (almost) totally automated mixing system! Oh well, please yourselves.
If you are heavily into MIDI and debt, the DMP11 has to be the best buy. The DMP7, which has sold better (600 in the UK), is easier to use as a stand-alone mixer in situations where fader positions have to be true at all times, such as live mixing. The RTC1 remote controller is a luxury item whose value has been reduced by the advent of MIDI management programs. For the DMP range there are no obvious competitors. Investigate if you can. The future has been with us for some time now.
Yamaha DMP7 £2,936 inc VAT.
Yamaha DMP11 £1,495 inc VAT.
Yamaha RTC1 £709 inc VAT.
Yamaha MLA7 £335 inc VAT.
Yamaha-Kemble UK, (Contact Details).
Feature by Wilf Smarties
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!