Prophet 3000, courtesy of Matrixsynth
A sampler, introduced by Sequential Circuits in 1987. The Prophet 3000 was intended to be a high-end product among the samplers in the market at the time. It was one of the first samplers capable of stereo sampling, and its user interface was advanced for the time. However, it never got a chance to realize its potential, as it got caught up in Sequential's financial problems and subsequent purchase by Yamaha.
Unlike the Prophet 2000, the 3000 was offered only as a rack mount unit. The main 2U high rack mount box contained only a few controls; the bulk of the control functions were done using a tabletop remote control box, which came with the unit, and connected to it via a 10-foot (3.3m) cord. The remote control box featured a 8-line by 48-character LCD display, which also had graphics capabilities. Immediately below the display was a row of six function buttons; displays on the screen were divided into a number of "pages", and the functions available on a given page were indicated on the bottom row of the display, immediately above the buttons. Other controls on the remote consisted of a large "jog wheel" rotary encoder used to enter parameter values, and a set of up/down/left/right cursor movement buttons.
The unit was 8-voice polyphonic, although playing stereo samples required two voices per note played, reducing to 4-note polyphony in this mode. The audio chain of the voice architecture was very basic; other than the sample playing mechanism, each voice had a digital VCF capability (presumably a four-pole low pass algorithm, although the manual did not specify), and a VCA (an analog VCA, oddly), and there was a capability to pan voices when playing mono samples. The VCA had a dedicated five-segment envelope generator (with the "alternate release" mechanism found on other Sequential models of this era), and another was available as a source in the modulation matrix. Two digital LFOs were provided.
The strength of the unit was in its sample processing and playback capabilities, which were advanced for the time. All sampling was done at 16 bits, and the unit was capable of sampling at 16, 32, 44.1 (CD rate) or 48 KHz. A variety of sample recording starting methods were available: manual, threshold (started recording when the input signal reached a certain amplitude) or footswitch. Upon making a sample, the 3000 ran algorithms to detect both the frequency of the sampled sound, and choose what it considered to be optimum loop points. (How well this worked depended on the sampled sound; the manual pointed out that the frequency detection would revert to manual mode if a noise-like source was sampled.) Each sample could have two loops, a sustain loop (which played in a continuous loop while the VCA envelope was in its sustain phase), and a release loop (which played in a loop during the envelope release phase, until the envelope reached zero). Other then the necessity of the sustain loop preceding the release loop, the two loops could be positioned anywhere in the sample. (Reverse playback of a sample was possible, but looping was not supported for reverse playback.) Graphic editing on the LCD display assisted the user in manually choosing sample start and loops points when desired, A patch (called a "preset" in the manual) could have multiple samples, which could be distributed across the note span and/or set up for velocity switching, as desired. Various forms of amplitude manipulation such as compression and normalization were available.
Multitimbral operation was possible by placing the unit in MIDI mode 4. This allowed individual voices, or groups of voices, to be mapped to different MIDI channels. Each channel acted effectively as a separate synth; each could have its own programs and samples loaded (within the limits of available memory), and each had use of the full MIDI note range. In addition to the stereo outputs, eight individual outputs were available; a given channel could be mapped to an individual output, or a pair of outputs for stereo operation.
As was the case with nearly all samples of the era, the Prophet 3000 was ultimately limited by available sample memory. Most units came with 2 Mbytes of sample memory, which yielded about 23 seconds of stereo sample memory at 44.1 KHz. A few units came from the factory with 4 Mbytes of memory. Sample storage was via 3-1/2" floppy disk; the built-in floppy drive only accepted quad-density floppies, which allowed the full contents of memory to be stored onto a disk set consisting of a maximum of two discs. Because the operating system had to also be loaded from disk, the unit automatically saved the operating system code onto every sample disk set; this way, the #1 disk of any sample set could be used as a boot disk.
The Prophet 3000, as it turned out, was destined to have a short and troubled history. Per Robert Alexander's article in Sound on Sound, Sequential was able to build 67 units before operations ceased in December 1987. These were all shipped to Sequential's European distributor, but they had significant problems and many came back for warranty service. Sequential had parts on had to build about 185 more; Yamaha funded engineering to fix the most significant problems, and assemble more units from the parts stock, all of which were sold by 1989. However, Yamaha was at the time on the verge of introducing its own sampler, the ill-fated TX16W, and so it had no interest in building more Prophet 3000s or making any further improvements. Thus, three planned capabilities never made it to market. One was that the sampler actually contained a SCSI interface for connecting a larger-capacity hard disk drive; Sequential had planned to offer one of its own, possibly based on a Syquest cartridge drive, which would have greatly improved the usability of the synth vs. loading and saving samples on floppies. A second was that Sequential had intended to offer an expansion chassis which would have expanded the unit to 16 voices, and possibly also provided for memory expansion. The third, hinted at in the manual, was that the unit included inputs and outputs for SMPTE signals, which possibly would have been used to add workstation features such as a sequencer that could be synchronized to tape.
Remaining units are considered moderately collectable, although they seldom appear on the market. The 4 Mbyte memory ones are much more desirable. They have not had major reliability issues other than the floppy drive, but a number of operating system bugs were never fixed.