I seem to have found a bug in the NS2... but want to confirm

[hand solo]

Hey

I was making a patch for Lucy In The Sky with Diamonds... In the process I seem to have found a bug, and I was hoping some fellow NS2 guys can load up this program and see if it behaves the same way!

First load up the program (it's for the LSD intro and has a bass guitar on the left and a electric harpischord on the right.
The samples are:
- RMI Electra Harpsi PH from here: http://www.nordkeyboards.com/sound-libr ... ectra-368x
- Hagstrom H8 Pick_Damp from here: http://www.nordkeyboards.com/sound-libr ... ss-guitars

To replicate the issue:

1: When you load the patch, play the right hand part, the sound is completely wrong:

• LSD_stored.mp3

2: Tweak the Flange Amount knob on Slot B so the display reads 0.1 - now it sounds as I intended it to, but if you store it, it reverts to the initial sound!
It seems this param is being incorrectly stored in memory...

• LSD_correct.mp3

3: As a control test, turn off the flange effect completely - it sounds like this:

• LSD_no-flange.mp3

[neolithic]

I had another look at this today and weirdly, the patch seems fine, so it must have been me...

[pablomastodon]

...cue theme from "Twilight Zone..."

the day I came home with my Wave, fresh off the plane from Sweden, I stayed up late with headphones and observed all manner of strangeness -- patches changing wildly and randomly without me touching anything, parameters changing willy nilly, definitely ghosts in the machine. I went to bed worried sick about what was wrong with my new baby. In the morning everything was fine and it hasn't had so much as a slight hiccup since then, now four years ago. Perhaps I was hallucinating...or perhaps the Murphy Law gods were just letting me know that I am powerless to stop them from wreaking mischief and havoc any time they so choose...

Pablo

[FunKey]

I was going to test this, but haven't had a chance yet. If it doesn't always happen, that kind of confirms my suspicion that there is some kind of timing artifact in certain situations. I've run into this kind of problem in the synth section by using the "detune" mode with a detune amount of 0 (not permanently, of course). In that case, it made sense to me because two oscillators are running at the exact same speed, and their phase difference might be set randomly when loading the program. I still haven't looked at your program, but I suspect there is also something that's accidentally "locked in" with the flange effect. If it happens again, maybe you could try changing the "rate" knob of the flanger.

Or maybe Pablo is right...

[neolithic]

Thanks guys. I am slightly reassured by your experiences, as I am usually pretty thorough in testing something before I leap to conclusions - most of the time it is down to human error
However, as a software dev I know that things like rounding errors can sometimes cause weird results and be hard to pin down, so it wouldn't surprise me if it was a real issue.
I'll try to experiment some more with reproducing it, just for the sake of my own sanity

[jaume]

do you have this sound for nord electro 3?? i only need the archive in this termination .nsmp

[neolithic]

I could sample it sometime when I have a chance - I'm a little busy so next week realistically.

Your post reminded me about this issue again. I did some further testing, and the issue is reproducible:
- load the program, edit the flange amount
- save the program, flange amount appears to have jumped to a wrong position.
- load a different program, then go back to the saved LSD program again. Sounds fine.
Verdict: I think it's a bug, but it doesn't bother me too much now I know the workaround. Maybe Pablo if you see this you can advise what the process is for reporting a bug? Via local customer support? Not sure.

[pablomastodon]

Not testing on your program above, but just testing by setting flange amount, storing, changing, etc.

When you hit the STORE button the first time and things start blinking, one can turn the main selector knob to scroll through potential program locations. While doing so, the instrument will actually load the programs in those locations so that you can play them to see whether or not this is something you wish to overwrite. Well, immediately upon hitting the STORE the first time the program of the existing location loads to provide you with that same possibility (before you turn the knob). If the only thing you've changed since the last STORE is Flange Amount, then hitting the STORE button once will give the visual result of having that Flange Amount revert to where it was previously. Hitting the STORE button the second time stores the program with the new value. Loading another program and then returning to the subject program will correctly load the newly saved flange amount value.

At least, I think that's what's happening. This thing has been out there for more than four years and thousands of people have been playing the heck out of them. Such a basic and obvious bug could not escape notice for so long if it were an actual bug. Not likely anyway...

Does this possibility answer your riddle or are you seeing something other than what I described?

Bless,

Pablo

[neolithic]

Thanks Pablo - your input is always greatly appreciated! It's slightly different in that it is definitely the saved program which is erroneous not the 'compare' program. It only really crops up on this program because the flange amount is almost zero: 0.1 which is the lowest non-zero setting possible. In most programs this difference would probably not be audible, but in the MP3s above you can hear it makes a difference.

[pablomastodon]

I'll have a go over the weekend using your program and see what happens for me.

Bless,

Pablo