Offline

I downloaded some single cycle waveforms from adventure kid's website.

He says he has pitched the samples at "D2+2"...(some of them are "D1+2")
Does this mean "D octave 2 plus two cents", like a slightly sharp D2?

I looped them in LGPT (with root note set at D2) and tested the pitch with a guitar tuner.
From D2 (going up in semitones) until about D4ish all notes registered as correct...as I got higher than that the notes were flatter (seems normal for a short looped sample spread out across many keys in a tracker).

On the LGPT instrument screen 'loop end' shows as '258' making me think he might mean 256samples + 2samples....

"+2" has confused me and my guitar tuner may not be good enough to register +2 as a sharp note/out of tune... so I came here to ask if anyone knows what it means...any help will be very much appreciated

Last edited by Mister Sombrero (Jul 3, 2013 3:50 am)

Offline
Sweden

To find out exactly you would need to know the sample frequency and the phase increment per sample for each note. Not sure what's up with notes going flat past D4, though. With fixed point phase counters (which I assume LGPT uses) the precision should actually increase the further up you go. Can you hear the flattening yourself, or is it possible that aliasing throws your tuner off? Maybe the tuner is just beter at registering a difference in pitch at higher frequencies. That is, the difference in pitch in terms of cents stays about the same, but as you go up the scale the difference in terms of Hz grows.

Last edited by boomlinde (Jul 3, 2013 7:29 am)

Offline
NC in the US of America

Your tuner doesn't happen to be "optimized for guitar frequencies" does it? That could possibly be limiting it's abilities to detect the pitch of that waveform

Offline
Montreal, Canada

Something something about number of playable samples in a single millisecond and resulting fractions being rounded off.

You notice higher frequencies going out of tune better than lower ones. That's why nobody notices when the bass player fucks up his part during a gig smile

Offline

I think your right about it being a problem with the tuner, that its possible aliasing throws the tuner off and that the tuner may be optimized for guitar frequencies. I used it to try and check the pitch of some of the AKWF fm waveforms and it could not even register that a note was playing.
If I use the perfect waveforms the note always shows.

thanks, your right...in theory the pitch should stay accurate as Im using SCWFs and not other types of short loops or one shot samples spread across the keys..."difference in pitch in terms of cents stays about the same, but as you go up the scale the difference in terms of Hz grows"

Although the guitar tuner was registering flattened notes I couldnt make out whether the notes I was hearing were flat or not by ear.
Thanks for the tips about sample frequency and phase increment.

I sent adventure kid an email asking about whether the D2 is raised by 2 cents or what he meant by "+2". If I get a reply I'll post the answer here. I should buy a more accurate tuner for future use. Thanks for all your help. Much appreciated smile

Last edited by Mister Sombrero (Jul 3, 2013 3:50 pm)

Offline
Unsubscribe

You can also get a wav sample of a test tone and use it as a tuning reference in piggy:
http://onlinetonegenerator.com/

Just make the tone you want for your root note put it in one channel, and then adjust the instruments root note until it sounds right in another.

Offline

Thanks for the link. That will be very useful

Offline
NC in the US of America
n00bstar wrote:

Something something about number of playable samples in a single millisecond and resulting fractions being rounded off.

You notice higher frequencies going out of tune better than lower ones. That's why nobody notices when the bass player fucks up his part during a gig smile

Yeah, man. Bassist I know tunes by ear because he has perfect pitch, but his perfection is a bit flawed. He has trouble finding the keys of songs because of this, lol. Whenever I have to play his bass I whip out my tuner and have at it.

But despite that it really doesn't sound so bad as long as he's not playing through his backup guitar amp, since that doesn't bass as well as his bass amp does.

Offline
Mister Sombrero wrote:

think your right about it being a problem with the tuner, that its possible aliasing throws the tuner off and that the tuner may be optimized for guitar frequencies. I used it to try and check the pitch of some of the AKWF fm waveforms and it could not even register that a note was playing. If I use the perfect waveforms the note always shows.

being a scientist, i can't resist to bore all of you with some theoretical details regarding this wink
especially for fm sounds is makes perfect sense for the tuner not to register the basis frequency at all, if you consider how fm spectra differ from those created by a plucked string instrument. for plucked strings, you always get the strong "base frequency" (first mode of oscillation) and only the overtones above that (being integer multiples of the base freq). an example can be seen at the bottom of this page: fixed string vibration
fm spectra look kinda different, as can e.g. be seen here: fm spectra in the second picture. first, the spectrum is always symmetrical to the base frequency, and second, depending on the modulation index, the amplitude of the base frequency might be smaller than that of the sidebands (look at modulation index 2 in that picture, for example). in addition, you might get aliasing effects if the sidebands go above/below the possible range of sampled frequencies.
hence it's not surprising that the tuner which probably just looks for the lowest/maybe strongest frequency component has a hard time recognizing the base note.

Offline

Adventure kid has confirmed that the "+ 2"  means 2 note octave cents...
D2+ 2 note octave cents (slightly sharp)

Offline

Sorry for the double post here.

I tried my guitar tuner with renoise and every note using renoise's included chip instruments registerd as being perfectly in tune (including FM sine).

LGPT oscillator tuning test 1

I took the A440 test tone sample from http://onlinetonegenerator.com/...
I reduced it to a mono single cycle waveform.
I loaded it into LGPT.
Switched to oscillator mode (loop end was 64 / 100decimal).
Notes played in tune if root note was set at C2.

LGPT oscillator tuning test 2

basic waveform samples taken at A440 from audiocheck.net (440Hz, 1 sec, -3db, 8Khz).
Cut them down to single cycle waveforms using renoise, chopping at zero crossing to get exactly 1 cycle.
Loaded them into LGPT.
Changed to oscillator mode...(loop end was 12 / 18decimal), Root note was A...
every note played out of tune...
Changed to root note C...still out of tune (consistently by one semitone)..
After I changed the root note to Csharp every note played perfectly in tune.

LGPT normal loop mode + oscillator mode tuning test

Extracted the single cycle waveforms from renoises included chip instruments (triangle smooth, triangle complex, chip triangle, chip square, chip sine FM, chip sine, chip sawtooth, chip saw smooth)

They will be in tune in normal loop mode if root note is C3.
They will be in tune in oscillator mode if root note is C6.

---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

if LGPT expects A440, root note C in oscillator mode...the renoise single cycle chip samples should be more out of tune in normal loop mode as the root note was C and not A.

Also these waveforms were behaving really strange in terms of tuning:

basic waveforms at A440 from audiocheck.net (440Hz, 1 sec, -3db, 8Khz) loop end 12

(because they played absolutely perfectly in tune across all octaves and notes, but only if root note was Csharp)...WTF?

Last edited by Mister Sombrero (Jul 5, 2013 4:25 pm)

Offline
Unsubscribe

Sounds like a bug report.

Offline

The C# thing looks fishy indeed.

One thing that is important to understand is that the very goal of the oscillator mode (in theory) is not to have to care about the note that was originally played when sampling the cycle. So if you have a single cycle sampled from an A4 note, to get correct pitches you need:

+ In regular loop mode, you should set the root key to A4 to get correct pitches (like in regular samplers)
+ In oscillator mode, you should *leave* root key to C. The oscillator mode re pitches the sample data automatically so it's tuned.

This doesn't explain the case you have where you need to put root key to c#. If you want to send me the project you are using, I could look at it.

Note also that the oscillator mode has been know for heaving tuning quirks since the beginning. It's not enormous but it's there. If I remember well, I had found a way to fix it but most of my song turned into uninteresting mud so I decided to keep it that way.

Hope this clarifies the situation a bit

Offline

I decided to give up on the A440 waveforms downloaded from tone generator websites and just stick with the SCWF chip samples from renoise.

These samples are at C (C4 I think).
in loop mode they are perfectly in tune over all notes and octaves..
In oscillator mode they are almost in tune over all notes and octaves, just slightly flat.
(The samples are in renoise 2.8 instrument folder - chip - chiptriangle, chipsine, chipsquare, chipsawtooth)

As for the adventurekid waveforms,  "+2" means they are D, sharp by two note octave cents.
My guitar tuner registers his perfect waves as being in tune in normal loop mode at detune values from 79 to 7D.
Using AKWF samples with the test tone sample and tuning by ear will be awesome too.

I dont think I really need to use oscillator mode anymore...as everything can be perfectly in tune anyway...
Next, ill check out creating hand drawn SCWFs for LGPT in milkytracker...should be cool

thanks for reading my questions and helping me out.

Last edited by Mister Sombrero (Jul 5, 2013 11:51 pm)

Offline
Geneva, NY
Mdashdotdashn wrote:

If I remember well, I had found a way to fix it...

I think you called it PFIN. tongue

Offline
Unsubscribe

I guess the next feature is the ability to do a pfin adjustment at the instrument properties level, because I think it isnt a full cent sharpness.. to my ears tuning up makes it still off a tad bit.