PSR-4500/4600 (1989-1990)
It's time to change... but to what?
Release Date: 1989 (PSR-4500)/1990 (PSR-4600)
Introduction: The Beginning of the End of the FM Era
FM sound generation technology was a huge success for Yamaha during many years, mostly during the 80s. It was used in literally dozens of different models, including synthesizers and sound modules, organs, home keyboards, arrangers and even digital pianos. Even today, the iconic DX-7 is still the world’s best-selling synthesizer of all history, with over 300,000 units sold worldwide. FM was not only capable of creating a wide palette of sounds, but also provided new sonic textures that weren’t impossible to obtain with the previous analog synthesizers. For those reasons, FM sounds were used in thousands of commercial recordings and pop hits.
Towards the end of the decade of the 80s, though, things began to change. There was an increasing interest in sampling (PCM) technology, and despite the prices of memory were outrageous, the first hybrid or fully PCM-based models began to show. In 1987 Roland launched its famous D-50 synthesizer, which used a hybrid synthesis technology, labelled L/A (Linear Arithmetic), combining basic waveforms (sawtooth and pulse) oscillators with PCM samples. In 1988, Korg released another iconic model, the M1, which was fully based on PCM technology (called AI Synthesis). Yamaha themselves, in the next generation of synthesizers just after the DX line, the SY line, opted for hybrid or fully PCM technologies. The SY-77 (1989) combined the new AWM (Advanced Wave Memory, a PCM-based synthesis) with an enhanced version of FM, called AFM (Advanced FM). The mid-level SY-55 used just AWM synthesis, whereas the lower end models, like the SY-22 and the SY-35, used another type of hybrid synthesis, combining samples and 2-Operator FM synthesis.
Yamaha decided at that time to use the same technology as of the SY-22, with certain limitations, in their new line of PSR keyboards of 1989, being the PSR-4500 (labelled as X-4500 in the United States) the flagship model of the new line. They labelled the new synthesis technology in the PSR line (not in the SY) as “DASSystem”, a fancy acronym which stands for “Dual Architectural Synthesis System”, stressing the fact that combines two synthesis methods to produce the final sound. This was, thus, the beginning of the end of the FM era, although the next generation flagship model, the PSR-6700 (1991) was still based on the same technology.
The PSR-4500 was followed the next year by the PSR-4600 (1990), but both are actually the same model, just with a different colour scheme and, according to some webpages, with slight differences in some voices and accompaniments. I cannot confirm that point because I don’t own (or ever owned) a PSR-4500, but, in any case, everything commented here is equally applicable to both models.
The fancy-looking "DASSystem" logo
Looks and Design: What the Heck?!
It’s not possible to look at the PSR-4500/4600 for the first time and remain impassible. The design is… erm… interesting, to say the least. You might like it or not, but it’s quite shocking, and, certainly, not a usual one. Many elements, such as the evidently inclined shape that recalls an airplane or even a spaceship, the big lateral “ears” for the speakers, or the strange control bars at the bottom of the keyboard, put this model apart from the rest. It seems that Yamaha wanted to state clear that this was a new stage and a new technology, and the appearance had to reflect this, with a boldly different design from the previous keyboards. Regardless of other positive or negative aspects of this design, a direct consequence is that the PSR-4500/4600 is a very long keyboard, probably the longest 61-key one ever made by Yamaha; in fact, it’s even longer that many 76-key keyboards, and this makes it difficult to find a proper bag for it and affects negatively to its portability.
Beyond the external looks and the first impressions, IMHO the design and construction of the PSR-4500/4600 reveal some points which are sadly a step backwards compared to previous models, and which are even not adequate for a flagship model.
First, it’s built entirely in plastic, and although this is not uncommon even in high-end models, it doesn’t feel as sturdy as, say, the DSR-2000. In fact, it feels rather flimsy, and it has probably the cheapest feeling and sense of poor quality of all the models of the series in this webpage. Older models had some details that showed that we were in front of a “top quality” product, like the foldable panel of the PS-6100/PSR-6300 or the tilt of the DSR-2000. In contrast, the control panel of the PSR-4500/4600 is fixed, in order to reduce costs.
Second, this model, again, uses an external power supply. The PSR-4500/4600 and the DSR-2000 are the only flagship models without an internal PSU. The internal PSU is usually a distinctive feature of the high-end equipment.
Third, in a clear step back compared to the previous DSR-2000, the PSR-4500/4600 lacks an LCD screen, which could have been very useful considering that it also incorporates synthesizer functions. Instead, it has two 7-segment LED screens instead of just one, but anyway the information, other than the number of the voice and the accompaniment, is much more difficult to interpret.
Fourth, flagship arranger models, as a rule, mount high-quality keyboard mechanisms, usually better than those used in the cheaper home keyboards. Nonetheless, Yamaha decided to mount a cheaper keyboard in the PSR-4500/4600, of the same type and quality than the inferior models. The keys in the PSR-4500/4600 are noticeably noisier than those of the PS-6100/PSR-6300/DSR-2000 and don’t feel as good.
Fifth, another weird decision was to put two bars at the bottom of the keyboard, which take almost the same width as the keys themselves, instead of the usual pitch bend and modulation wheels. Besides the fact of the unusual design (I haven’t seen anything similar in any other keyboard from Yamaha or from any other brand), this leads to some practical problems. For instance, it’s necessary to apply significantly more force to operate the bars than in a typical wheel, and their movement is pretty much slower. And moreover, it’s virtually impossible to control both at the same time with just one hand (usually the left one).
Finally, there are also some cuts in the connections department: the keyboard lacks a midi thru terminal, and the headphones jack is in the rear side, which is more uncomfortable and avoids the need for a specific circuit board on the front side.
The Hardware: What’s DASSystem About?
On the hardware side, the most relevant novelty of the PSR-4500/4600 is that, for the first time, there is a major change in the sound synthesis since the FM, which was present (with minor changes) in the previous three flagship models.
As already told, Yamaha used, as a marketing catch, the label “DASSystem” to state that this was a new and a different technology than before. The “DA” part stands for “Dual Architectural”, so this is telling us that the new technology combines two types (architectures) of synthesis to create the final sound. Concretely, it combines PCM-based (digital sampling) synthesis, which is called “AWM” (Advanced Wave Memory) by Yamaha, with the good old FM (specifically, 2-Operator Frequency Modulation synthesis).
All the magic is made by the new (then) TMC3493PH chip, also known as GEW-5. The voices can have two elements (1 AWM waveform and 1 FM element), or four elements (2 AWM waves and 2 FM elements). The GEW-5 chip is 8-part multitimbral and provides 8 voices of polyphony, but a very interesting feature is that many GEW-5 chips can be combined in a chain in order to increase the polyphony, while sharing the same waveform ROM data, so if a keyboard implements two GEW-5 chips (as in the case of the PSR-4500/4600 or the SY-22, for instance), there is a total of 16-voice poly for the same set of sound patches.
The maximum amount of shared waveform ROM data by all the GEW-5 in a chain is 2 Megabytes (16 Mbits) of 8-bit or 12-bit samples. That’s not very much, but it was a decent figure in that time. In fact, the PSR-4500/4600 have just 1 Megabyte (8 Mbit) of waveform data (cheaper models of the PSR line had just 512 Kbytes and a single GEW-5 chip). Most relevantly, The AWM synthesis part of the GEW-5 is very limited, as it lacks a very important feature: the filter. What the AWM part can do is basically to reproduce waveforms in different pitches and amplitudes (volume), but it cannot control the harmonics through a LPF or whatever other kind of filter, so the resulting sound is quite static and not very expressive. There is an amplitude envelope and a LFO (which can affect the pitch -for vibrato- or the amplitude -for tremolo-), but that’s all. The FM part can provide more expressiveness through dynamic changes in the timbre, responding to key velocity or to the envelope of the modulator operator, for example.
Just because the GEW-5 can handle PCM samples, there’s no need in the PSR-4500/4600 for specific hardware for the drum and percussion sounds, like the RYP4 in older models. The keyboard has a total of 42 drum/percussion sounds plus 14 sound effects like laughs, applause and the like. Regarding drums and sound effects, as a new feature this model includes 8 pads for playing sounds manually (up to three different sets can be programmed, plus 1 more of pre-programmed sound effects).
Another remarkable feature of the hardware is that the PSR-4500/4600 includes, for the first time in a Yamaha’s arranger, a DSP chip, specifically the YM3413, for the reverb effect. The keyboard is very limited regarding digital effects, though. There’s only one effects processor for the overall sound (it applies equally to all tracks), with the three following modes: Stage (reverb), Hall (reverb), and Echo (delay). There are no parameters at all to edit, so it’s not possible, for instance, to select which parts are to be affected or not, the amount of reverb (send level), the reverb time, the reverb delay, or whatever. The quality of the effects processing in the 80s wasn’t either very high, and sadly the PSR-4500/4600 is no exception, as it feels quite grainy and unpolished. Moreover, in the 80s and early-mid 90s there was a tendency to abuse of the reverb effect, resulting in a sound which is quite drowned under the effect. Nevertheless, it’s better having this than nothing 😉
What Does It Sound Like? A Mix of Flavours
This is for sure the most interesting part. One would suppose that a relevant change in the technology of sound generation implies a relevant difference in the sound character, and, in general terms, this is so in the PSR-4500/4600. All the previous models relied on Frequency Modulation synthesis (FM), which is capable of many different sound textures and expressive sounds, but it’s not very adequate to emulate the timbres of real instruments. The main advantage of the PCM samples is that, as they are digital recordings of actual sounds, can reproduce the original sound source accurately, which is very good if you’re trying to sound more or less as a real instrument, but, on the other hand, it’s more difficult to manipulate and tends to be quite static. Manufacturers try to overcome this by sampling different “layers” of the original source in different octaves and amplitudes, and through the use of digital filters, but this requires more processing capacity and lots of available memory, and in the 80s and 90s memory was a scarce and expensive resource, and processing power was much more limited than today.
The PSR-4500/4600’s technology doesn’t implement a digital filter, and the available ROM memory for waveforms (the “raw data” of sounds) is limited to just 1 Mbyte. This means that there’s not space for many nor for very long samples, and this has an impact on the overall quality. Moreover, the quality of the digital recording itself depends on two factors: the sampling resolution (8 bit, 12 bit, 16 bit, 24 bit…), and the sampling frequency (11 Khz, 22 Khz, 44.1 Khz, 96 Khz…), and the GEW-5 seems to be able to manage 12 bit @22 Khz samples at most, which is quite a low quality (as a reference, standard CD quality is 16 bit @44 Khz). 8-bit and 12-bit samples tend to “break” quite quickly when they are reproduced at a lower frequency than the original sampling (for instance, when the original sample is a C3 and you play it in a C2, and when there’s no sample of a lower frequency (i.e., you have a sample of a C3 note but not of a C2, for example), the quality loss is evident.
That said, it’s clear that many “acoustic” voices sound more realistic (or less unrealistic, if you prefer) in the PSR-4500/4600 than in previous models. This does not mean that the result is always better, as sometimes a less realistic sound can be more expressive or “musical” than a poor sample. An obvious example is the acoustic piano. The piano sound of this model is honestly very poor, as it clearly lacks more samples, both in different frequencies and intensities (velocities). The piano voices are flat and dull, and this is a pity, because the acoustic piano is a very important sound, very expressive, versatile and suitable for almost any style of music, and it’s usually the first that you hear when playing a keyboard for the first time, and tends to give the player an approximate idea of the overall sound quality of a keyboard. For this reason, manufacturers pay special attention to it, and it’s not surprising that even low-end models such as the Yamaha PSR-E373 have very good acoustic piano voices.
As told before, for the first time in a Yamaha’s arranger we find DSP effects: we can choose between two types of reverb (stage and hall) or an “echo” (delay without reverb of any kind). The quality, though, is rather poor and the effects are not editable at all. Besides, there is also a rather obscure “ensemble” button. The first impression is that this is a chorus/detune effect, similar to what there was in previous models, but honestly, I’m not sure, because in some cases it appears to alter also the timbre of the voice (that’s evident for instance in the “Electric Piano 1” voice), and not just to apply chorus, so it could affect to the different elements (probably the FM ones) of the voice.
Then there is the synthesizer. Indeed, calling this a “synthesizer” is very benevolent, as it’s just a few editable parameters. The structure is very similar to the previous DSR-2000, although there are even less parameters, and the editing process is less straightforward (no cursor buttons and no LCD screen). Moreover, there are no memory registers for the edited voices. These are the parameters:
1) Spectrum: As in the DSR-2000, this alters the timbre of the sound. It affects only the FM element of the voice, as the GEW-5 has no digital filter for AWM voices. Usually, the higher the value, the brighter, harsher and more “metallic” the sound.
2) Attack 1: Affects the attack part of the envelope of the FM modulator, so it can be used to alter the timbre of the voice during time.
3) Attack 2: Affects the attack part of the amplitude (volume) of the voice.
4) Decay 1: Affects the decay part of the envelope of the FM modulator (the timbre).
5) Decay 2: Affects the decay part of the amplitude of the voice.
6) Release: Affects the time it takes the volume to vanish after the key is released.
7) Vibrato Depth: Quite self-explanatory, although it actually affects either the amplitude (tremolo) or the pitch (vibrato), just like in the DSR-2000.
8) Vibrato Speed: Sets the speed of tremolo/vibrato.
9) Volume: The overall volume level of the voice.
10) Velocity Sensitivity: Sets the dynamic range (amplitude variation) of the voice according to key velocity value. A setting of 0 implies no sensitivity (basically for organ sounds).
Another novelty is that, for the first time in a flagship arranger model, there is a demonstration tune. In older models there wasn't any, perhaps because in those years built-in demos were seen as something of "cheap" models and improper for "serious" equipment. Anyway, in the PSR-4600 there are three different demo songs. Neither of them, though (in a common practice in all the following models) is an example of the arranger functions of the keyboard, but instead are multi-track sequences using the internal sounds.
Retaking the structure of the older PS-6100 and PSR-6300, the PSR-4500/4600 has "Solo" (monophonic) and "Orchestra", and "Orchestra to Lower" sections (with 4 fixed keyboard split points). This time, though, there are no separate sound sets for each section, and any of the 100 preset voices can be freely chosen. Moreover, there is a Dual Voice function for the "Orchestra" section.
Here there are some examples of the preset voices of the PSR-4600:
A partial view of the control panel, with the preset voice list
The Arranger Functions: Bye Bye "Rhythms", Hello "Styles"
We are very used to the concept of "styles" when thinking about accompaniments in arranger keyboards. But this has not always been the case. In the 80s, when there were severe limitations in multitimbral capabilities and in polyphony, automatic accompaniments basically consisted in a drum pattern, a bass line and rhythmic chords, or perhaps a simple arpeggio. The final result could more or less resemble a certain music genre, but the limitations were obvious.
The PSR-4500/4600 introduces significant changes in how accompaniments are conceived. From the beginning it's evident that the arrangements are more complex, and there are two "chord" tracks instead of just one. That is, an accompaniment now has drums, bass and two different chord/melodic lines. Also, the number of accompaniments grow significantly, to a total of 100. Moreover, there is a new function, called "Solo Styleplay", which in theory sets the most suitable voice for the style and adds some effects such as echoes, harmonies and/or arpeggios to "enhance" the result.
Is all this too nice to be true? Sadly yes, at least to some extent. On the one hand, sometimes (more often than not) the arrangements have quite elaborated or complex melodic lines, arpeggios, harmonies, etc., which might sound more or less good in themselves, but often make it difficult to fit well with the lead part (the melody you are playing), especially if you have to switch between different chords quickly. This is an example (only minor Root and IV chords are played):
Second, in this models Yamaha clearly focused more on quantity than in quality. Although some of the styles are quite good, many of them are mediocre and some others (specially within the Latin and World sections) are even ridiculous or laughable.
Third, things doesn't go better (indeed, it's often the opposite) when the "Solo Styleplay" is activated. Too many times you wonder why someone would even think that the chosen voice would be suitable for the style, and too often the applied effects (harmonies, arpeggios, echoes, etc.) only make things worse. And not to mention that when the Solo option is active, the melodic performance turns monophonic.
In a very similar vein than the old PSR-6300, there are three fill-in patterns for each style, a variation, and an intro/ending. The user can program up to 5 original accompaniments, each one with rhythm, bass and two chord tracks. The available voices for the chord tracks are not all the presets, but a selection (about 40-50 voices to choose from).
Some examples of accompaniment styles (together with melodic lines):
A partial view of the panel, with all the styles listed at the top
Other Functions: The Music Programmer, the Roll Bar, the Registration Memory, Amplification and MIDI
As usual in top-of-the-line keyboards, the PSR-4500/4600 has a built-in multitrack recorder, labelled in this model as "Music Programmer". This time, though, its has only four tracks (Solo, Orchestra, Accompaniment/Orchestra and Bass/Orchestra). The last two tracks can be used to record the auto accompaniment. On the other hand, now it's possible to record more than one song (up to 4). "Editing" functions are as limited as in the DSR-2000, as the only thing than can be done is to re-record a track from a certain point to the end. The manual does NOT specify the memory capacity of the multitrack recorder, so you could have a bad surprise, especially when recording more than one song.
With no doubt, one of the weirdest things of those models is the strange control bars at the bottom of the keyboard. While the sort one at the left is just a pitch bend controller (BTW, there seems to be no way of setting the pitch control interval), the long one at the right (called "Roll Bar") can have more functions than just modulation. The control panel has dedicated buttons for selecting the function of the roll bar. The available options are: Modulation (pitch cycling to create a vibrato effect), Orchestra Brilliance (controls the amount of brilliance of the "Orchestra" voice), Solo Expression (controls the volume or amplitude of the "Solo" voice), Orchestra Expression (the same but with the Orchestra voice), Accomp. Expression (the same but with the accompaniment parts), and Master Expression (it affects the global volume of the keyboard, in a similar way as an expression pedal). It's nice to have different options, although only one at a time can be used.
The PSR-4500/4600 also offers the possibility of registering all the panel settings in memory (although this does NOT affect the synthesizer settings) and recall them with the touch of a single button (the Registration Memory), as in the older DSR-2000. But also, as in the previous DSR-2000, the keyboard can only store four different panel settings.
In the amplification department, the PSR-4500/4600 repeats the 2 x 8 watts amplifier combined with 12 centimeters speakers as the previous DSR-2000. In contrast to the DSR, though, the design of the speaker cabinet is much bulkier, contributing to the great lenght of the keyboard. This extra space, though, hasn't sadly much impact on the sound quality. Still, the overall sound is gritty, with undetailed highs and poor lows, so using external amplification/monitoring is advisable whenever possible. At least, Yamaha finally dropped the RCA connectors for the audio output in favour of the more standard (in pro audio) jack terminals. There are also a couple of RCA connectors for mixing the audio signal from an external source.
Regarding to MIDI, there is a relevant change this time. While on the one hand there is no MIDI Thru terminal, on the other hand the default MIDI mode is now mode 3 (omni off, poly), so this means that the PSR-4500/4600 is ready from the start to be used as a multi-channel (8-part) sound source. The MIDI implementation is still quite poor, as the keyboard does not respond to very common MIDI controller messages such as panning, epxression or aftertouch, and, in a clear step back from the previous models, the PSR-4500/4600 has no portamento capabilities (I suspect that this is a limitation of the synthesis architecture itself). At least, it's less of a headache to use the keyboard as a MIDI sound source in a DAW, despite the fact that, for some strange reason, the voice numbers of the panel do not correspond to MIDI program numbers, and it's necessary to have the sound table of the manual at hand to know which preset to select when using the DAW.
These are some examples of MIDI sequencing using the PSR-4500/4600:
Verdict
In my opinion, with these models, which combined a very aggressive looking and a new synthesis technology, Yamaha tried to state clearly that this was something different to what was seen before. The result, though, could have been significantly better, as in many aspects those models don't seem to meet what it's expected in a flagship keyboard. This is my own list of pros and cons:
Pros
Latest available technology for sound generation (back then), which produces a more realistic sound for "normal" instruments than the previous FM technology
DSP effects, albeit very limited
Important increase in the number of accompaniments
Output jacks instead of RCA connectors
Multichannel MIDI mode as default
Cons
Highly questionable aesthetic design
Build quality is rather poor, with an obvious decrease in quality when compared to previous models
Pitch bend and roll bar are very odd, and difficult/hard to operate
Does not really take all possible advantage of the new synthesis technology
Styles focus more in quantity than in quality
Synthesizer functions are very limited, and in many aspects such as multitrack recording or registration memory is just as limited (or even more than previous models