How do *I* think the correction for the PC3 operates ?

[Theo Verelst]

Recently I've made this (Video demo of some (ROM) Kurzweil PC3 "corrected" sounds.) this It might really happen: the PC3 correction thread about a part of my latest PC3 adventure in the specifically intended sense of correcting a lot of strange audio components in it's sound that have bothered me from the start and which I didn't really remember from other Kurzweils and obviously consider unwanted.

 

So, you get the PC3 or supposedly a newer variation with similar chips and functionality out of the box, probably update it to OS2.21 or a later one if that's there, put it on your favorite monitoring or live amplification, and well: try it out. Nice, deep sounds broad sound palette, but, for me pretty soon prohibitive distortion and sound pummel elements that wouldn't make me happy live and didn't make me happy to just connect it up to the DAW and record it.

 

Them there's the question of *how*does*it*work, which has at times been discussed with all kinds of people giving opinions, some strange confirmations of my findings as well, but never a definitive Open Sourcing of the whole PC OS and sound generators, or a definitive outlook on what's right and what's wrong, or how the customer can change the overall sound according to their preferences, except maybe "get and Artis" or possibly some sound advice, etc.

 

I think the instrument at hand (and mind you: I've titled this thread even with the "what *I* think" disclaimer ...!) is an interesting and complicated instrument that easily escapes the standard ROMpler syndrome of samples+some processing, and deserves the corrections I've worked on thus far from some type of global design point of view, Now, like I've said, the part of the design of the instrument I now talk about is at a high level design criteria set, not #of bits, sample rate, type of digital resample filter, standard effects quality, etc. kind of level, no matter how interesting that can be, as well.

 

No, I'm explaining here that my opinion is that there is design at various levels of the instrument that correlates with advanced mixing tools from the highest grade of music makers of the last half century or even longer, and can be tuned to, well, sounding like a a yodeling, peeping bagpipe, a bland and irrelevant sounding sample player with too old sample set, and, as I've tried to demo a bit like an A grade music level instrument which besides acoustic and analog instruments is one of the few who can create sounds that appeal, have good character, mix bands (like blare control and low-mid warmth that doesn't suck after two notes already), realism (to some extend thus far) and sound different than most everything else when put on a decent sound system.

 

Now, I'll try to get there where these sounds, strangely enough by the use of effects, will be changeable from the sounds I've demoed, to other variations that work good, preferably with sliders, to extend the sound palette. I have found a lot more interesting (also non-ROM) sounds than in the demo, which makes me think that is possible. And that includes VA sounds, of which some I have are good enough to take serious in general, as in for decent performances or recordings, which I didn't get to yet in the already lengthy demos.

 

So how do I think it works. First of all the "correction" was done exactly as I've mentioned: an effect path added to the sound (mixed in , added to the main sound, before the master section unless I'm mistaking) which contains properly frequency dependent flanged distortion as one of the main ideas, and the whole sound being (mostly mildly but accurately) equalized by the Master EQ followed by multi-band compression from the built in Master 3 band multi-compressor, adjusted to frequency bands corresponding to the mid-frequency processing I've demoed averaging effects for in the past.

 

Two main scenarios aren't (entirely) true: I haven't exactly subtracted a signal that was otherwise added to the internal sound such as to exactly annul it, and I also don't think I've tickled an internal processing setup that's been explicitly prepared to respond to these effects and parameters. The exact truth I do not altogether know, and certainly cannot prove, but I know for sure there's a lot of stuff going on in the signal path of the machine besides what happens in a straight software or rompler signal path, and there are digital effect preparations in the general operations of this particular machine type that allow for digital processing that comes out to certain harmonic tunings and that can somehow identify sample components and generate pure signal components from that which can come out pretty beautifully well, which includes certain types of compensation for the built in DAC.

 

So I won't waste people time with more potentially alluring dialectics, will record some version of my (Linux) software compensation for the same purposes (and similar sound results) and prepare zooming in to more interesting and well formed digital signals.

 

No need to talk about this, or read this if you don't want to, I know I'd have wanted someone to give me those effect settings years ago, would have saved me a lot of work and recording troubles, but anyhow, not everyone has the same interests. Isn't it.

 

Theo V.

[area51recording]

Maybe instead of endlessly trying to "correct" the PC3 it's time to just buy a new keyboard. Unless "correcting" the PC3 is actually your thing, in which case, carry on...

[Theo Verelst]

You didn't really try to read well, isn't it ?

[MurMan]

I did read it.

 

What *I* think is that you are not "compensating" for the DAC or any other imagined flaw in the PC3.

 

I think you are adjusting the sound to your personal tastes, something that most of us do.

[Theo Verelst]

Are you a bit scientifically formed in the field at hand ?

[MurMan]

A bit. I'm an EE and have implemented VoIP codecs and DSP functions that are in use commercially. I'm not a PhD theorist.

[Theo Verelst]

Given a background with some normally proper theoretically foundation, you should be able to remind yourself of the (mathematical) sample reconstruction theorem.

[Theo Verelst]

(Darned spell correct) theoretical.

[MurMan]

Theo, if you have come up with a novel method of improving the sample playback process, you need to do the following:

 

1. Get a patent attorney

2. Write a formal paper

3. Get the paper peer-reviewed

4. Stop wasting your time on a musician's forum. If you've really developed something, get it out there.

 

In the meantime, color me "skeptical".

[MathOfInsects]

You didn't really try to read well, isn't it ?

 

Am too!

[DulceLabs.com]

You didn't really try to read well, isn't it ?

 

No it is that for this not.

[area51recording]

You didn't really try to read well, isn't it ?

 

Awwww.......busted

[Ulf]

OK, English is not my native tongue, I never got my EE degree, and there is no tl;dr (shouldn't there be one in a post like this), but SERIOUSLY?

[WillNeverPost]

That's a mistake, IMO. Taking Theo seriously, I mean. It only encourages him.

[smanzella]

I have always found that the best way to improve any sound that is sampled, synthesized, modeled, or otherwise is to simply practice and learn to play it better. My .02. YMMV.

[Kenny Ingram]

The PC3 correction is here: the Forte. By the Forte's standards, it's sounds are much more out of date. Thread after thread the pattern is that there needs to be a million correctionshow about you upgrade to the Forte?

[AnalogGuy1]

Theo - you may have some interesting ideas here, or not; from what you've written I can't tell. (I am a DSP theorist). But you're posting them the wrong forum. Try the DSP developer's section over at KVRaudio. It's a good group, and every one of them will be intimately familiar with your interest in the sample reconstruction theorem. Suggest you also post there a block diagram explaining your thoughts so it is not just words. But I think you'll just get flack posting in a musician's board if you want to generate real discussion. Do that at on the DSP subboard of KVR, and here share with us your stories of crazy things said to you at gigs! Then you'll have the forumites laughing with you, instead of getting annoyed by intimations that they can't understand your posts.

 

Here's an example of mine: "Do you need help carrying that?" said by a friendly patron watching me schlep my 50 lb RD-700 to my car as we were tearing down. Nice thought, except as she rose to help she was so drunk she teetered, and then fell back in her seat.

[Hoochie]

A bit. I'm an EE and have implemented VoIP codecs and DSP functions that are in use commercially. I'm not a PhD theorist.

 

OT kinda but.

Because I've been doing this forever.

 

I cut my teeth back in the 1A2 days of Western Electric.

Diode ring arrays. Clocking relays. Real Buzzers. Real light bulbs.

Call directors with thousands of connections hard wired.

Static that could be anywhere in all of that.

"Bugs " that were really crawling around inside pooping on everything.

All the things that make a phone tick.

I miss it all sometimes.

 

Now I'm chasing echoes, one way audio, routing issues, packet loss, ghost calls

Codec mismatches , jitter , buffer over under/runs, m-law , u-law, video conferencing , IM/chat integration ,ACD , hosted, premise, network analysis, fiber, POE, desktop, CRM integration, presence mobility, blah, blah , blah.

My buffer is is over run.

Now the static is " I c-- bare-- --re a w--- y---e saying." ( I can barely here a word your saying) or "you sound like your underwater" or "I call you and it doesn't even ring. You can hear me talking? Weird!"

 

And I do a lot of editing of what are loosely called instructions

so that ordinary people can understand this technology and use it

in ways that make their jobs easier and more efficient.

 

So ya. I can follow some loosely organized data points and make some

WAG's about intended results.

 

Color me mystified as to what this is about.

Audio is a very personal thing.

We all "hear" differently and it's not something that

a baseline or control can easily be attached to.

My ears are whack from too many years of rock and roll.

 

Differentiation though can be easily shown or "heard ".

By giving us examples. Now the dog "can" hunt.

 

All this is to say;

After years of reading between the lines,

looking at odd things from within and without,

in a very technical field of study to get at real issues,

I'm not groking this. ( for you fellow Martians.

If you have not read Heinlein, I highly recommend doing so. )

 

 

 

John

 

 

 

 

 

[Tom Williams]

I ain't got no EE, though I have read about and mostly understood issues of musical acoustical science for something like 50 of my 57 total years. And Theo (whom I consider part of this forum family, and do not want to alienate) is as inscrutable to this native Appalachian English-speaker as he is to the rest of you. But the following:

an effect path added to the sound which contains properly frequency dependent flanged distortion [...] equalized by the Master EQ followed by compression

...looks like an aural exciter, which would indeed tend to pep up the sound. I gotta say I find it interesting, and I hope to try it out one of these days.

[jimmymio]

The PC3 correction is here: the Forte. By the Forte's standards, it's sounds are much more out of date. Thread after thread the pattern is that there needs to be a million correctionshow about you upgrade to the Forte?

I am a dissatisfied PC3 owner and I was wondering about this myself. Are you saying, Kenny, that the overall quality of the Forte's sounds are better than those of the PC3?

[Hoochie]

I've not owned a PC3 but as a Forte owner all I can say

is the audio quality itself is as good

if not better than anything

I've ever played.

 

This coming from an ex owner of

CP4

Kronos

M3

Motif

 

 

Current owner

Nord Stage 2

Forte 88

MOTU ultra lite / Mainstage

 

Maybe the PC3 does have something that needs correction.

But based on Kurzweils Name Branding of

"It's the Sound" I find it hard to believe.

 

Unless your monitoring with a gain stage and monitors built

in a bunker under Funk & Wagnells front porch

of Unobtainium to most anybody

except the landed gentry.

Or a pair of electrostatic headphones that,

once again, are a ridiculous luxury that

pencil scratches in an Amish monastery

can be heard through.

"You there! You sharpened that pencil

with a pencil sharpener didnt you. I can tell by

the lack of telltale dihedral/ quadrilateral scratching

whitling knifes cuts will cause as the lead is drawn across the paper!!

Your banished for using technology!!

 

John

 

 

[stoken6]

I have always found that the best way to improve any sound that is sampled, synthesized, modeled, or otherwise is to simply practice and learn to play it better. My .02. YMMV.

You see, I think that's what separates Theo from most others on this board. I read Theo's posts (as best I can) and I see someone who is genuinely fascinated by, and driven to improve on, the smallest imperfections in the sound quality he hears. Everyone else is interested in the musicianship, the playing. Who cares if a reconstruction error causes a +0.6db blip at 17kHz lasting 182 samples?

 

I would defend Theo's right to post here, no question - but I can't help feeling (as others have mentioned) that we would all benefit if instead of posting

remind yourself of the (mathematical) sample reconstruction theorem.

that you explained the theorem (perhaps with a link to more sophisticated material) and why it's relevant here.

 

I've learned so much here from people who do that: on technique, voicing, timekeeping, synthesis techniques and other topics.

 

Cheers, Mike.

[stoken6]

I have always found that the best way to improve any sound that is sampled, synthesized, modeled, or otherwise is to simply practice and learn to play it better. My .02. YMMV.

You see, I think that's what separates Theo from most others on this board. I read Theo's posts (as best I can) and I see someone who is genuinely fascinated by, and driven to improve on, the smallest imperfections in the sound quality he hears. Everyone else is interested in the musicianship, the playing. Who cares if a reconstruction error causes a +0.6db blip at 17kHz lasting 182 samples?

 

I would defend Theo's right to post here, no question - but I can't help feeling (as others have mentioned) that we would all benefit if instead of posting

remind yourself of the (mathematical) sample reconstruction theorem.

that you explained the theorem (perhaps with a link to more sophisticated material) and why it's relevant here.

 

I've learned so much here from people who do that: on technique, voicing, timekeeping, synthesis techniques and other topics.

 

Cheers, Mike.

[stoken6]

I have always found that the best way to improve any sound that is sampled, synthesized, modeled, or otherwise is to simply practice and learn to play it better. My .02. YMMV.

You see, I think that's what separates Theo from most others on this board. I read Theo's posts (as best I can) and I see someone who is genuinely fascinated by, and driven to improve on, the smallest imperfections in the sound quality he hears. Everyone else is interested in the musicianship, the playing. Who cares if a reconstruction error causes a +0.6db blip at 17kHz lasting 182 samples?

 

I would defend Theo's right to post here, no question - but I can't help feeling (as others have mentioned) that we would all benefit if instead of posting

remind yourself of the (mathematical) sample reconstruction theorem.

that you explained the theorem (perhaps with a link to more sophisticated material) and why it's relevant here.

 

I've learned so much here from people who do that: on technique, voicing, timekeeping, synthesis techniques and other topics.

 

Cheers, Mike.

 

[Kenny Ingram]

The PC3 correction is here: the Forte. By the Forte's standards, it's sounds are much more out of date. Thread after thread the pattern is that there needs to be a million correctionshow about you upgrade to the Forte?

I am a dissatisfied PC3 owner and I was wondering about this myself. Are you saying, Kenny, that the overall quality of the Forte's sounds are better than those of the PC3?

Yes, that's exactly what I'm saying. The PC3's architecture is from before 2010. The Forte is STILL being updated to this day. It only takes about thirty seconds of one Forte promo video to see that this is exactly what it is: a PC3 upgrade. Enough of this scrambling around and flaunting your EE degree because it ultimately makes you an amazing sound designer: the Forte was specifically designed for this purpose. You have AMAZING software/sound guys like Dave Weiser who have literally put hours and hours of time into making sure the Forte is the best Kurzweil keyboard released so far. If you're dissatisfied with the sounds on your PC3 and want an upgrade, go for the Forte.

[Dave Weiser]

The PC3 correction is here: the Forte. By the Forte's standards, it's sounds are much more out of date. Thread after thread the pattern is that there needs to be a million correctionshow about you upgrade to the Forte?

I am a dissatisfied PC3 owner and I was wondering about this myself. Are you saying, Kenny, that the overall quality of the Forte's sounds are better than those of the PC3?

 

The Forte's overall sound quality is nicer. Better converters, signal path, etc.

 

More importantly, the Forte's new samples (Steinway, Yamaha C7, Rhodes 77/73, Wurly, Clav, Harpsi, Celesta, Glock) are roughly 1000x bigger than the corresponding samples in the PC3. I'm not usually a number spec kind of guy but those figures do a reflect a huge difference in quality.

 

The Forte is a newer board (2014) than the PC3 (2007) and has now (with OS updates) become Kurz's flagship - so I would expect to see this kind of improvement.