Polyphony question...

[SciNote]

In the PSR-E433, and other PSR-E series keyboards, for that matter, do certain voices take 2 "notes" of polyphony instead of 1?  And I'm not talking about voices that actually turn on the dual voice and add a dual voice automatically.  I'm just talking about regular sounds that only use one voice.  I'm asking because I'm recording a new song, and I can hear at a section where the bass cuts out, and it is due to polyphony because the bass is there if I just play that track alone.  And yes, I do have some 4-note chords going on some of the other tracks, but even considering that one of those chords is using the dual voice, and even factoring in a few "notes" of polyphony for the background rhythm drums, I still don't see that I'm playing anywhere near 32 notes when the bass is cutting out.  And I'm only using the style for the drums, not the instrumental backing tracks.

This isn't a big deal, as I will just record the bass line to a separate Audacity track when bringing it to my computer, and then put the other five sequencer tracks on their own Audacity track.  Or I may even also put the drums on its own Audacity track so I can independently adjust the level of them.  But I am curious as to why I am having a note drop out.

[SeaGtGruff]

I'm not certain. but I think that XGlite voices use only 1 note of the polyphony-- that is, only 1 element per voice.

I know that certain XG voices do use 2 elements per voice, and for that reason I used to think that those same voices use 2 elements in XGlite. But then I saw an old forum thread which stated that XGlite voices use only 1 element.

One way you should be able to test this yourself is by turning the Release Time up as high as it will go. Then play 32 notes one after the other quickly, such that the 1st note should (if the polyphony is high enough) still be sounding when the 32nd note is played. If you can hear the first 16 notes cutting off as you play the 17th through the 32nd notes, then the voice must be using 2 elements.

I tried this on my PSR-E433 with one of the voices that uses 2 elements in XG, and I couldn't hear any of the notes cut off as I played 32 notes.

[SciNote]

That's not a bad idea.  I'll have to give it a try.  The only problem I can see is that, even with 16 notes playing, that would make enough of a sound where it could be hard to determine if the 16 other notes are playing or not.  But, I guess if I put the first 16 in the low octaves, and then the second 16 in the high octaves (or visa versa), it will be more obvious.  I'm mainly talking about the main panel voices, not the XGLite.  I'm wondering if at least some of the panel voices use 2 elements.

I thought I had read something about this, but I could be mistaken.  I have lost the data list for my E433, and I couldn't find it on Yamaha's site, either.  But they did have the data list for the E443, and while it does mention how certain voices turn on the dual voice, I didn't see anything mentioned about single voices using 2 elements.

[SeaGtGruff]

The voice lists for many of the old PSR models-- the ones that weren't subdivided into the "S" and "E" and "A" and other series-- had an additional column showing whether each voice used up 1 or 2 notes of polyphony, but the voice lists for the XGlite models don't have that.

Then again, the voice lists for the PSR-S models don't have that, either, and they definitely have voices that use 2 or more elements.

I didn't describe the idea for testing the polyphony very clearly, because the way I heard it was to strike the lowest key as forcefully as you can (with the sustain pedal held down, or in the case of the PSR-E433 just crank up the Release Time), and then start at the highest key and work your way down from them. That way it should be easier to tell when the first note gets cut off.

[vbdx66]

Hi,

Just in case this might be useful for someone, I uploaded the data list for the PSR E433.

Regards,

Vinciane

[attachment deleted by admin]

[AnupamEnosh]

The voices which are multi layered could be using two and three notes at a time. I have not tried and tested, but some Panel voices (E453) like Harmonium 2 & 3, Symbiont, SweetHvn, HandsUp!, DynChime, PncyChdz, RSNoizPd, TeknoMan, DreamHvn, BellPad etc. seem to be using multiple timbres of the same/different voice, that could use more polyphony at the same time.

[SciNote]

Well, I tried Michael's suggestion, where I selected a sound and cranked up the envelope release to the max, and then tried to play 32 notes to see if I could hear them cut out.  It's still heard to tell, because even at maximum release, the sound fades out naturally as I play more notes.  But what I did was to start out with high notes, playing chords or arpeggios, and then quickly did the same thing at the lower end of the keyboard, figuring that it would be easier to hear the high notes cut out.  I did it with the string sound, and it seems like the original notes were playing even as I played more than 16 notes, so it seems like there is truly 32 notes of polyphony.  However, I need to try it with other sounds.

If there really is 32 actual notes of polyphony for all sounds that don't turn on the dual voice, then maybe the reason I was having notes cut out had to do with a small amount of release causing chords to momentarily overlap, causing substantially more notes of required polyphony at that moment that the bass note cut out.

[SeaGtGruff]

I think I used the piano-roll editor in a DAW to play the notes rapidly and send them to the keyboard. This has two possible advantages:

(1) You can tell the DAW to play the first note as loudly as possible but play the other notes more quietly.

(2) You can alter the Bank Select and Program Change values that you want the DAW to send before playing the notes.

[pjd]

Hi Bob --

Just a random thought over coffee and you may have already thought of it...

Maybe play 31 notes through MIDI channel 1 at a low channel value (10). Then play the 32nd note through MIDI channel 2 at a high volume.

I guess the method is to chew up polyphony at low volume and try to expose the drop outs at a high, easy to hear volume. Needs to be in a MIDI file -- possibly driven by a DAW -- in order to be repeatable.

I'm just trying to find a way to make the drop outs audible. Lucky for my nephews kids, I gave them my E443. Unfortunate for me as I can't try this experiment on the E443. (The E443 is now 1,000 miles away.)

Take care, love your posts -- pj

[SeaGtGruff]

I like pj's idea of making a MIDI file for testing this, because the keyboard has a function (button) for replacing the Song Melody Voice with the currently-selected Main Voice, so it should be possible to play back the song file while cycling through each of the keyboard's voices.

[vbdx66]

Hi PJ,

It is about time to invest in a PSR E463 or EW410 

Regards,

Vinciane

[SeaGtGruff]

By the way, Yamaha still has the Data List for the PSR-E433 and earlier models, but they no longer have active product pages for those models-- so you have to use the search icon at the top of the site pages to search for the model you're interested in, which should give you results that include a link to the following page:

https://usa.yamaha.com/products/contents/keyboards/downloads/manuals/index.html?k=&c=keyboards&l=en&p=17