Hard-coded detune of VCO1 and VCO2/ribbon peculiarities/etc.

Forum for fans of the wonderful Yamaha AN1x Analog Physical Modeling Synthesizer, AN200 desktop module and PLG150AN.

Moderators: parametric, Derek, Clyde, Saul, Fozzer

Re: AN1x Ribbon Glitch

Unread postby db7 » Fri Apr 19, 2013 12:22 pm

Box wrote:Then something is off with your AN1x, VCO1 and VCO2 behave in the same manner next to my SY77 and other instruments I have. Odds are it has to do with you having the older O.S. and 1.04 should fix what problems you're having in that regard.
Another reason that it would help a lot if other users with various OSs would do a couple of simple tests and report the results in this thread. I’m surprised no one has bothered yet; it only takes a minute or two.
User avatar
db7
Member
Member
 
Posts: 809
Joined: Tue Jan 01, 2013 12:37 pm
Location: Scotland

Re: AN1x Ribbon Glitch

Unread postby Box » Fri Apr 19, 2013 10:51 pm

I have 1.03 and you 1.02, just need to find someone with the older O.S. and you'll have the new 1.04.

Oh, I started a new thread in the Chill Out Room to go over some other ideas I have in regards to other things.
Ensoniq VFX
Korg Monotron
Korg Triton
Peavey Fury "Raven" 1992
Peavey KB-300
Roland JX-8P
Yamaha AN1x
Yamaha KM802
Yamaha SY77
User avatar
Box
Member
Member
 
Posts: 273
Joined: Fri Mar 09, 2012 9:01 pm
Location: Springville, AL

Re: AN1x Ribbon Glitch

Unread postby SysExJohn » Sat Apr 20, 2013 8:57 am

What do I need to do to contribute to this survey, please?

I think the guy who sold me the AN said that it had the latest version (1.06?) in it.
I'll have to look up the manual to see how I check that, but I do remember having seen something at some point that confirmed that on the display.
As I'm still really a novice with this beast perhaps you could give me some more detailed instructions, Box?

Regards,
John.
MIDI Tutor forum at http://midi-tutor.proboards.com/index.cgi.
UW500; 2xMU1000 + PLGs AN, DX, VL, VH & PF; SC8850; EWI4000s + VL70m; PCR-M30.
AN1x, Novation X-Station 25. E-MU 02 cardbus + 1616m x 2. Sonar 7 PE. XGworks. & SQ01.
Garritan GPO4, COMB2, JABB3, IO, Organs, Steinway, World, Harps.
User avatar
SysExJohn
Member
Member
 
Posts: 100
Joined: Tue Dec 02, 2008 1:00 am
Location: Betelgeuse, shoulder of Orion.

Re: AN1x Ribbon Glitch

Unread postby db7 » Sat Apr 20, 2013 11:33 am

Given that I’ve hijacked poor Box’s thread with this stuff about the detuned VCO2, perhaps I should provide the help. ;) I have bolded some things below, which isn’t to patronise anyone’s reading comprehension but simply to emphasise the key points between the other info.

To check which version of the ROM you have, enter the test menu by turning on the AN1x while holding –, 0, and +. Continue to hold them and watch out for the number being reported on the screen just before the menu itself opens. The latest (and presumably final) version is 1.04. Then you can use the menu if you want. I think you have to turn the synth off and on again to exit the menu, but maybe there’s a button that quits it that I missed.

After that, I would appreciate it if you could let me know which version of the ROM you have and whether your VCO1 and VCO2 slowly drift out of phase with one another while tuned to exactly the same Pitch* (and with no Fine adjustment). The easiest way to test that is probably to perform a reset to a type 1 initial voice (press 130, choose type 1, confirm; no currently stored patch will be overwritten unless you press Store) and turn both VCO1 and VCO2 up to full under Mix/VCF. Another thing that would be worth doing if you can is to check each single VCO, in turn, against another instrument using notes of the same pitch (with only one of the two VCOs active at a time, I mean).

If your AN1x is like mine, you will find that tuning the two VCOs to unison still produces a slowly animated phasing in and out, and that VCO2 is responsible due to being slightly detuned when compared to VCO1 or other instruments playing the ‘same’ pitch. The difference in phase will be slowly moving, not constant, and the detuning will be in VCO2 and not something that can be corrected by the Fine adjustment as it is lower in magnitude than the steps of the latter.

My somewhat lengthy descriptions might make this sound like a lot of work, but it really isn’t! ;) I would appreciate if you and anyone else with a few minutes to spare can report your findings. I presume I’ll know for sure if the detuning of VCO2 is a ROM-specific issue once I receive my chip of 1.04, but that’ll be a while yet. I don’t see how it can be something peculiar to my AN1x, so with the help of other users, we’ll begin to get a picture of whether/when this was changed.

Thanks for the offer, anyway. :) The results will be useful regardless of which version of the OS you have, and the same goes for other readers; once we have each version covered, and possibly duplicates of each for confirmation, we’ll have a better picture of what’s going on.


* Edit: I previously wrote “VCO1 and VCO2 drift out of tune with each other”, but that’s the wrong term. Drift implies an evolving divergence in pitch of one or both oscillators, whereas what occurs here is two oscillators of constant pitches but which differ, creating slowly moving differences in phase, not tuning.
Last edited by db7 on Sun Apr 21, 2013 2:21 pm, edited 1 time in total.
User avatar
db7
Member
Member
 
Posts: 809
Joined: Tue Jan 01, 2013 12:37 pm
Location: Scotland

Re: AN1x Ribbon Glitch

Unread postby SysExJohn » Sun Apr 21, 2013 2:17 pm

Confirmed, ROM version 1.04, now where did I get the idea that it was 1.06 from? Oh well!

Doing as you describe reveals a slight drift between the two oscillators as expected.
I'll try to update this with a comparison to another instrument in due course.

Regards,
John.
MIDI Tutor forum at http://midi-tutor.proboards.com/index.cgi.
UW500; 2xMU1000 + PLGs AN, DX, VL, VH & PF; SC8850; EWI4000s + VL70m; PCR-M30.
AN1x, Novation X-Station 25. E-MU 02 cardbus + 1616m x 2. Sonar 7 PE. XGworks. & SQ01.
Garritan GPO4, COMB2, JABB3, IO, Organs, Steinway, World, Harps.
User avatar
SysExJohn
Member
Member
 
Posts: 100
Joined: Tue Dec 02, 2008 1:00 am
Location: Betelgeuse, shoulder of Orion.

Re: AN1x Ribbon Glitch

Unread postby db7 » Sun Apr 21, 2013 2:25 pm

Thanks for testing!

I now think Box’s AN1x (using 1.03) was doing the same thing all along. He said he could achieve locked phase between VCO1 and VCO2, but he also specifically described a slowly moving interference between the two waves, indicating changing phase. Such movement is not possible without a difference in pitch between them. A simple difference in phase at identical pitches would not cause it. If post-1.02 versions of the ROM had really removed this difference, he would not hear any movement between VCO1 and VCO2, even when they were in unison. The most that would be heard would be cancellation due to the free-running starting phases of the two oscillators, which would vary with their history of pitches and which all of us can already hear readily by playing two notes using the (metaphorically) same oscillator and retriggering them in between appropriate alterations to phase.
User avatar
db7
Member
Member
 
Posts: 809
Joined: Tue Jan 01, 2013 12:37 pm
Location: Scotland

Re: AN1x Ribbon Glitch

Unread postby Box » Sun Apr 21, 2013 6:45 pm

There not being a difference in pitch is what's causing the sound I was describing. I put osc 1 against osc 2 the same thing happens if I were to put osc 1 or osc 2 against DCO 1 or DCO 2 of the JX-8P. Then if I put DCO 1 and DCO 2 of the JX-8P on at the same time the same result occurs on the AN1x with osc 1 and osc 2 being on.

Image
Two waves of the same frequency cause the above on the left. That's what I'm hearing, I'm not so sure about everyone else. If you want, take a chromatic tuner and you'll see that osc 1 and osc 2 are the same.

For a "real world" demonstration take two acoustic guitars and get them in tune with each other on the open strings. Have them face each other and strike the same note, the same thing happens as shown above. To eliminate all other variables, record one note being played and then overdub itself and the same phenomenon happens. Hell, just record yourself playing one note on the AN1x then overdub that.
Ensoniq VFX
Korg Monotron
Korg Triton
Peavey Fury "Raven" 1992
Peavey KB-300
Roland JX-8P
Yamaha AN1x
Yamaha KM802
Yamaha SY77
User avatar
Box
Member
Member
 
Posts: 273
Joined: Fri Mar 09, 2012 9:01 pm
Location: Springville, AL

Re: AN1x Ribbon Glitch

Unread postby db7 » Mon Apr 22, 2013 2:36 pm

Box wrote:There not being a difference in pitch is what's causing the sound I was describing.
That makes no sense. If they were at exactly the same frequency, regardless of phase, there physically could not be animated movement of the interference; it would be constant in nature. I can achieve two oscillators with no difference in pitch (but possibly constant differences due to interference of differing phases) by layering the two VCO1s or VCO2s of Scenes 1 and 2.

But there most definitely is a difference in frequency between VCO1 and VCO2 within a given Scene, as can be verified by playing the two in unison, the two an octave apart, or each of the VCOs in isolation against the SY99 playing the same nominal pitch. All produce slow phasing that is conclusively indicative of movement between the waves and thus a difference in frequency from the quoted pitch in at least one of them, specifically VCO2.

For a "real world" demonstration […]
Acoustic guitars and other non-digitally tuned instruments or recording interfaces are irrelevant. I have a fine “real world” demonstration in my living room. I have invited other users to test this and obtained at least one indication that I am not the only person whose AN1x has a slightly detuned VCO2. Your replies have gotten too confusing for me to determine conclusively whether your synth exhibits the same difference, but I highly suspect it does.

I await further tests from SysExJohn and other users. Other things that would be useful to catalogue in addition to the behaviour of VCO1 and VCO2 in unison are behaviours when VCO1 and VCO2 are pitched an octave apart, the interaction of each in isolation vs. another synth (preferably one with a ‘perfect’ waveform and pitch such as the sawtooth, square, etc. of the 77/99), and so on. Again, for me, VCO1 and VCO2 are detuned against each other both at unison and at intervals, and further testing attributed this movement to the slightly shifted pitch of VCO2.
User avatar
db7
Member
Member
 
Posts: 809
Joined: Tue Jan 01, 2013 12:37 pm
Location: Scotland

Re: AN1x Ribbon Glitch

Unread postby SysExJohn » Mon Apr 22, 2013 3:48 pm

If I may add my 2d here, not to be taken as conclusive but merely another opinion.

When, as a lad, I was learning the oboe, my teacher and I had instruments from different manufacturers.
To make mine not "howl" against hers I used to have to pull the staple (the piece of metal onto which the double reed is bound) slightly out (about 2 mm) from the home position within the neck of the instrument. I imagine that my natural A was slightly sharp against hers. (This could all so easily be misinterpreted! :lol: )

We "tuned up" (or rather I tuned down, she was slightly flat :wink: ) until we achieved a kind of unison not dissimilar to the sound I'm hearing with the two oscillators on the AN. In other words we were probably within a cycle per second of each other at A=440, or thereabouts. Much finer adjustment proved nigh on impossible with the sharp/flat roles being reversed. The phenomenon was most noticeable when playing standing next to each other reading from the same score.

Interestingly when playing on opposite sides of the room, the beating wasn't as noticeable, if at all. Room effects I suppose. But then what do I know about acoustics?

Consequently I have been led to believe that most of these beating effects are due to slight differences in tuning between instruments with some kind of proximity effect.

Couldn't we record the output of osc 1 against osc 2 at say A below middle C and measure the exact frequency of each? I suspect we would find a fractional pitch difference. Maybe I'll try that, but I'm not sure whether Audacity can measure fractional pitch differences, although maybe just comparing one waveform against the other will show it.

Regards,
John.
MIDI Tutor forum at http://midi-tutor.proboards.com/index.cgi.
UW500; 2xMU1000 + PLGs AN, DX, VL, VH & PF; SC8850; EWI4000s + VL70m; PCR-M30.
AN1x, Novation X-Station 25. E-MU 02 cardbus + 1616m x 2. Sonar 7 PE. XGworks. & SQ01.
Garritan GPO4, COMB2, JABB3, IO, Organs, Steinway, World, Harps.
User avatar
SysExJohn
Member
Member
 
Posts: 100
Joined: Tue Dec 02, 2008 1:00 am
Location: Betelgeuse, shoulder of Orion.

Re: AN1x Ribbon Glitch

Unread postby Miks » Mon Apr 22, 2013 4:01 pm

Wouldn't it be possible to have both osc's waves on a scope whilst performing those test procedures? Imho this should visualize if there's any detuning between the oscillators?
My gear (in alphabetical order):
Kawai: K1 II
Korg: M1, M1R, Poly-800
Roland: D-50, D-110, D-550, PG-1000
Yamaha: AW1600, DX7IID, EX5R, RM50, RX5, SY99 (with 'Cool Blue' mod), TG500, TX-802, YMM2
Sector101: SYEMB05 (5x), SYEMB06 (3x), EXFLM2 (1kit), MCD Sweet16 (1x)
User avatar
Miks
Member
Member
 
Posts: 803
Joined: Fri Oct 31, 2008 1:00 am
Location: Germany

Re: AN1x Ribbon Glitch

Unread postby Box » Tue Apr 23, 2013 5:43 am

Frequencies are frequencies, it doesn't matter if they're produced naturally or synthetically. They all interact in the same manner regardless of origin. That being said I double checked with the Triton and it being all digital such gave results as shown in the picture I posted earlier. The sound just got louder with no change in sound. I went back and checked the JX-8P and AN1x and they do the phasing thing, not amplifying the signal as the Triton did. So it'd seem all my years of using analogue synthesizers have gotten me used to the oscillators being out of tune with each other. But it's so minute, a fraction of a cent difference for the phasing to be so slow.

Miks wrote:Wouldn't it be possible to have both osc's waves on a scope whilst performing those test procedures? Imho this should visualize if there's any detuning between the oscillators?

Or just use a tuner that gives out the frequency in hertz for a definitive reading. I have one but people are sleeping and since it uses a microphone it'll have to wait.
Ensoniq VFX
Korg Monotron
Korg Triton
Peavey Fury "Raven" 1992
Peavey KB-300
Roland JX-8P
Yamaha AN1x
Yamaha KM802
Yamaha SY77
User avatar
Box
Member
Member
 
Posts: 273
Joined: Fri Mar 09, 2012 9:01 pm
Location: Springville, AL

Re: AN1x Ribbon Glitch

Unread postby db7 » Tue Apr 23, 2013 5:04 pm

Box wrote:Frequencies are frequencies, it doesn't matter if they're produced naturally or synthetically.
My point was that it does matter when we are talking about minute differences in pitch and/or phase vs. the possibility/expectation of perfectly locked waves. The latter are the province of digitally controlled synths only, for constant synchronisation between two oscillators/notes isn’t something that can be guaranteed in analogue ones or non-synthetic instruments such as acoustic guitars, oboes, etc. The minor difference in pitch in the AN1x is due to its VCOs not being exactly the same frequency, whereas one might expect identical frequencies fairly logically, and they’re certainly possible from its digital nature – unlike in non-digitally controlled instruments, hence my assertion that they are hardly relevant here. Whew! That was a mouthful. Do you see what I meant now?

I went back and checked the JX-8P and AN1x and they do the phasing thing, not amplifying the signal as the Triton did. So it'd seem all my years of using analogue synthesizers have gotten me used to the oscillators being out of tune with each other. But it's so minute, a fraction of a cent difference for the phasing to be so slow.
Great, so I was right after all? :roll: :) [edit] Thanks for confirming, of course! :D [/edit]

Box wrote:
Miks wrote:Wouldn't it be possible to have both osc's waves on a scope whilst performing those test procedures? Imho this should visualize if there's any detuning between the oscillators?
Or just use a tuner that gives out the frequency in hertz for a definitive reading. I have one but people are sleeping and since it uses a microphone it'll have to wait.
It would be nice to see concrete measurements like this, so thanks in advance to anyone who performs them. This has further encouraged me to get a proper analyser in my signal-chain, at least on the PC… which could help greatly with understanding FM synthesis, too, since FM Theory and Applications is nice for a theoretical description but leaves most of the practical results to the imagination. :P

[edit]
SysExJohn wrote:Couldn't we record the output of osc 1 against osc 2 at say A below middle C and measure the exact frequency of each? I suspect we would find a fractional pitch difference. Maybe I'll try that, but I'm not sure whether Audacity can measure fractional pitch differences, although maybe just comparing one waveform against the other will show it.
It would certainly show the difference with sufficient probing, but I don’t know whether the effect could be summarised well (e.g. visible while maintaining discernability of the individual cycles). Anyway, something that could make this easy to check would be setting up two Scenes, one with VCO1 only and one with VCO2, setting Separate to full (Yamaha, y no spread of oscillators within a single Scene?), and recording a stereo wave into a PC for checking of the two channels relative to one another. Again, maybe it’s not worth it, though, if it wouldn’t be very informative visually. I can try this again later but would need to transport the screenshot or whatever from my PC to an online one… I did the same thing before with a mono signal and confirmed the existence of the slight phasing between VCO1 and VCO2.
[/edit]

In anticipation of someone asking why this small difference in VCO2 matters, I’d better state that I’m not claiming it’s an instrument-breaker, abnormal compared to other instruments (especially genuine analogue synthesisers), or anything. The point of all this has been to determine whether it was a general and unavoidable feature of the AN1x, which we now seem to have confirmed, and later to obtain some more concrete data on its nature: we know it’s VCO2, but it’d be nice to get some shots from an oscilloscope, an idea of the difference in tuning, and so on.

These sorts of things are good to have available online for reference. I’d like eventually to put together a table of how parameters such as VCF cutoff map to certain values of Hz and other ‘real life’ figures, and info on the native detuning of VCO2 will be a nice addition to that sort of reference. I know there’s a list of frequencies for the LFO on the list on Yahoo; however, I’ve not come across any other translations of the AN1x’s native ranges to their equivalent measures. That is, of course, barring those for the effects provided in the official Data List, which seem somewhat strange in their inclusion when the more relevant main parameters are not described!

Working on such a reference could be a nice prelude to starting something similar for the 99/77… but I won’t even consider that for a while. ;)
User avatar
db7
Member
Member
 
Posts: 809
Joined: Tue Jan 01, 2013 12:37 pm
Location: Scotland

Re: AN1x Ribbon Glitch

Unread postby Box » Tue Apr 23, 2013 6:14 pm

So there's nothing wrong with the AN1x then in that regard. :P Every VA and analogue I've had are the same way, so over time I just came to expect it as the norm. Though the real question is how far off are they. It'd have to be a fraction of a cent for the phasing to be as slow as it is when put against a stable digitally produced waveform. It's hardly noticeable though and is close enough to "be in tune" as any acoustic instrument could be close enough to be in tune. Also it's not just VCO2 that does it, VCO1 does it as well. If VCO1 were perfect it wouldn't beat against the perfect sine wave of the SY77.
Ensoniq VFX
Korg Monotron
Korg Triton
Peavey Fury "Raven" 1992
Peavey KB-300
Roland JX-8P
Yamaha AN1x
Yamaha KM802
Yamaha SY77
User avatar
Box
Member
Member
 
Posts: 273
Joined: Fri Mar 09, 2012 9:01 pm
Location: Springville, AL

Re: AN1x Ribbon Glitch

Unread postby db7 » Tue Apr 23, 2013 6:21 pm

Box wrote:So there's nothing wrong with the AN1x then in that regard. :P Every VA and analogue I've had are the same way, so over time I just came to expect it as the norm.
Newbs like me would not necessarily come to the same conclusion. :P Or people who would expect it to be an option if anything, rather than an enforced feature that’s always present and can’t be deactivated… which also describes me! Again, it doesn’t ruin anything, but it’s definitely noteworthy IMO.

Though the real question is how far off are they. It'd have to be a fraction of a cent for the phasing to be as slow as it is when put against a stable digitally produced waveform.
Agreed that an attempt to quantify this would be nice to have for the sake of reference. I also want to collect screenshots of the various single waveforms for that purpose. Then there are the more involved things I’d like to quantify, but that’s more work!

It's hardly noticeable though and is close enough to "be in tune" as any acoustic instrument could be close enough to be in tune.
Indeed, but again, we’re dealing with a digital instrument, and the difference is significant for those users who might have expected to be able to obtain perfectly synchronised intervals, which can now be seen to be impossible within a single Scene and only achievable by layering identically numbered VCOs of two Scenes. Once again, that doesn’t really ruin anything – especially when one considers that reinitialisation of phase on key-down, which also alters the relationship albeit in a static way, is missing from the AN1x – but I think these are things that are worth documenting nonetheless.

Also it's not just VCO2 that does it, VCO1 does it as well. If VCO1 were perfect it wouldn't beat against the perfect sine wave of the SY77.
I didn’t notice this during my admittedly quick test of triangle(-ish) waves from the AN1x vs. the SY99; perhaps you’re correct, but I’ll check again later.
User avatar
db7
Member
Member
 
Posts: 809
Joined: Tue Jan 01, 2013 12:37 pm
Location: Scotland

Re: AN1x Ribbon Glitch

Unread postby Miks » Tue Apr 23, 2013 7:13 pm

Another idea: checking those frequences by using a high-resolution frequency-meter may clarify if the osc's are 'in tune' or if one of them is slightly off...
My gear (in alphabetical order):
Kawai: K1 II
Korg: M1, M1R, Poly-800
Roland: D-50, D-110, D-550, PG-1000
Yamaha: AW1600, DX7IID, EX5R, RM50, RX5, SY99 (with 'Cool Blue' mod), TG500, TX-802, YMM2
Sector101: SYEMB05 (5x), SYEMB06 (3x), EXFLM2 (1kit), MCD Sweet16 (1x)
User avatar
Miks
Member
Member
 
Posts: 803
Joined: Fri Oct 31, 2008 1:00 am
Location: Germany

PreviousNext

Return to Yamaha AN1X, AN200 & PLG150AN

Who is online

Users browsing this forum: No registered users and 1 guest