Tweaking polyphony of PR-E:0387 RD-1000 without compromising sound

[xp30]

Context: This week I have watched a Youtube video, complaining about the polyphony of the Fantom, and the tone PR-E:0387 RD-1000 was used as one example. I was curious, looked at the tone, and thought this is a good example for providing some hints for tweaking polyphony. Beyond that, there is no overlap between my post and the video.

It is helpful to know that the tone RD-1000 existed in previous models, for example in my XP-30 from 1999. The synth engine gained more and more features over time. However, these old tones were simply migrated and only use the subset of features that existed when they were initially created.

If you try to follow these steps, consider (temporarily) deactivating all effects (MFX and reverb), because it makes it easier to hear the differences.

The tone uses 3 partials: The first partial uses wave C.24 MKS-20 P3 A and is panned hard left. The second partial uses wave C.25 MKS-20 P3 B and is panned hard right. Beyond the wave and panning, there are only two other options that differ between the first two partials. These two partials make up the main sound. The third partial is some kind of key-on/hammer noise (C.593 Piano Thump) that is only audible above C4.

Let's start with the first two partials. Given that they are almost identical, we can replace them with a single partial. Let's deactivate partial 2, change the panning of partial 1 to center, and then add C.25 MKS-20 P3 B as right wave. By doing that, we actually lose some volume. For this tone, the volume of partial 1 is already at 127. For this reason, I would increase the common volume a bit, and then reduce the volume of partial 3 to compensate for that.

I mentioned two more differences: First, AMP Bias Level was -10 for partial 1 and +10 for partial 2. We can use AMP Pan Keyfollow to mimic the effect, by setting it to +10, and setting AMP Bias Level to 0. Second, AMP Time Keyfollow was +20 for partial 1 and +30 for partial 2. I would just ignore this difference and leave it at +20.

So far, we have replaced two partials with a single (stereo) partial, which halves the number of consumed voices for this part.

Let's look at partial 3. As mentioned before, the key-on noise is only audible above C4. That's because of AMP Bias Level and AMP Bias Position. On my XP-30, the partial still consumes a voice if played below C4, even if it is not audible. I guess it is still the same on the Fantom. We can fix this by limiting the key range for this partial in the keyboard settings, that means setting Key Range Low to C4.

The other interesting thing is that partial 3 has a very long AMP Envelope Release Time. This could be an issue. However, I do not think it is an issue in this case. Partial 3 uses a short non-looping sample, and in such cases, the voice automatically stops with the end of the sample, and frees up the voice. I can confirm that my XP-30 works this way.

If there are doubts, then we could tweak the envelope. I guess the AMP envelope looks so bit strange (long decay and release time), because the key-on noise is supposed to be independent of how long the key is pressed down. On the Fantom, we have better support for such sounds, which makes it much easier to shape the envelope. We would first go into the Control section and change Envelope Mode to NON-SUS for partial 3, which gives us a lot of degrees of freedom for shaping the envelope.

The important point about partial 3 is: The sample is short, and we only consume voices for a short duration. That means partial 3 has a much smaller effect on polyphony than partial 1.

In summary, we have effectively halved the number of consumed voices, or doubled the polyphony for this tone. We did not have to compromise on the sound.

This is not specific to tone PR-E:0387 RD-1000. Most of the tones were not designed for the Z-Core synth engine, and for many/most of them, polyphony can be easily tweaked by using newer features. In particular for tones in the bank PR-E (XV5080) and bank PR-C and PR-D (SNS).

[neomad]

Thanks XP30 ! you rock.

Watched that video, your solution solves the problem of most of old programs not designed for Z-core but migrated to it...

Just wondering if Roland will ever replace the old (and used again and again since decades) samples in the Fantom. For instance, acoustic instruments and brass are almost un-usable to me in studio, and barely usable on stage. Was planning to get a Fantom 0 for stage but not anymore...

[xp30]

Watched that video, your solution solves the problem of most of old programs not designed for Z-core but migrated to it...

I might also post something about tweaking polyphony for pad sounds. In the video, a pad from bank PR-B was used. AIUI, this bank was created for the AX-Edge. I guess for a keytar, wasting polyphony was not a concern. I took a quick look. It seems the polyphony can also be easily doubled for this sound, with minimal changes to the sound, in particular when layered with a piano sound.

[xp30]

I might also post something about tweaking polyphony for pad sounds.

I am doing it here, because it is actually quite similar to the previous example.

The tone PR-B:0162 Soft Pad 1 consists of two partials, one panned to the left and the other one panned to the right. Similar to above, we can merge the two partials into one partial. We disable partial 2, and increase the AMP level of partial 1 to 112. We select A.293.Warm Pad A as right waveform for partial 1. We change pitch fine tune to 0, and instead set stereo detune to +1. We change the pan to R10, and stereo width to 70. We pick a cutoff frequency between the old values 503-526, for example 510, and similarly for the AMP attack time, I pick 120 (previous values were 104 and 192).

This sounds close to the original (while only consuming half the polyphony). However, there are differences, because we can not mimic the different cutoff frequencies and the different AMP attack times. One more thing: The tone uses analog feel (12), which also behaves differently when using 1 stereo vs 2 mono partials.

Having said that, I think the difference is negligible when used for layering with a piano.

For playing live, the original is maybe not a good choice. The left channel is brighter than the right channel. I think stereo width is nice, in particular for those who can hear both channels (e.g. musicians with in-ear monitoring). However, in this case, I think it is a bit too much.