.53 Odd frequencies when transposing samples

Harbinger · July 18, 2009, 00:25:27

Full Version:
OpenMPT v1.17.02.53

Has the bug occured in previous versions? If yes, please specify version(s): (This option is optional, and doesn't need research. But in case you know...) Not sure.

Description of the bug: When changing back and forth between transposition notes in the Sample page, the value in "Freq (Hz)" does not go back to the original, correct frequency.

How often does it happen?: Always.

How to reproduce (step by step description): Load a sample in the Samples page. Affix the note to a known frequency, say, 22050 Hz, in the Freq(Hz) field. Now click on the Transpose field and change the note to one lower or one higher. Now change it back -- the frequency does not go back to 22050. After a few of these note transpositions, the frequency becomes way off.

However, now that i think about it, i neglected to determine if the bug is in the actual playback or only in the frequency display...

Saga Musix · July 18, 2009, 00:34:52

Sorry but...uhm.... no? This is been working perfectly for me for years.

Transposing works as it's supposed to work here, nothing gets off tune. Maybe you can give me an actualy example file where this happens? I can't really imagine how this would happen. Really, those are functions that have been built into MPT for ages, and I can't spot anythign wrong here in the sample player, the Frequency selector or anything. I need an example.

Harbinger · July 18, 2009, 01:04:15

Well, i know i'm not crazy, Jojo. I was testing it as i was writing up the manual.

I'll make some jpegs and show you what i mean (hopefully tomorrow). :wink:

LPChip · July 18, 2009, 09:11:40

This can happen if the note hasn't been transposed before and is using a strange hz value.

For instance, if I'm transposing samples and one is off, I manually correct it to get them in tune. Would I then use the transpose feature again, it would not get to the finetuned value.

I think this is what happened here. The value was not set to its base value to begin with.

Saga Musix · July 18, 2009, 10:55:52

Yes, but this only happens once and only changes the note by a (very) few Hz. And Harbinger's example, 22050 Hz, is in no way a "strange" transpose value.

Oh, and JPEGs are no use for me, because
1) I prefer PNGs

2) I can't debug them. I need modules, samples, whatever.

Harbinger · July 18, 2009, 20:57:45

I just used 22050 because it is common. When you use the Transpose function to a specific note, then back to the original note, it turns into 22036, and keeps losing the value the more i switch notes using the Transpose popup menu.

But like i said, i didn't test if it was a display error or an actual pitch error.

Saga Musix · July 18, 2009, 22:17:27

And as I said, I can NEITHER find a display error, NOR a pitch error. I load a sample with 22.05kHz sample freq, transpose it up and down and NOTHING happens.

LPChip · July 18, 2009, 22:30:27

Harbinger, how do you transpose up? Do you use the small arrow keys on the frequency? Or are you selecting a different note?

Harbinger · July 18, 2009, 22:49:57

I'm talking about using the "Transpose" field under the Freq (Hz) field. Set a rate in the "Freq (Hz)" field, then choose different notes from the Transpose field, coming back to the original note and checking to make sure the original frequency is in the "Freq(Hz)" field.

BTW, if it helps track down the bug, if i remember correctly, this was for MPTM format, no special tuning.

LPChip · July 18, 2009, 22:52:26

Ok, what I thought.

Can you atleast confirm that you still can do this?

Perhaps create a song with just one sample (can be very short) that has this problem, and send it to us? That would really help.

Saga Musix · July 18, 2009, 22:55:09

Yes that would really help, because as stated very early in this thread already, I can't reproduce this with any setup I try.

EDIT: I can see that this may happen with very low transpoe frequencies, e.g. if you jump down to B-8. However, this is due to the limited accuracy of the translation table and it should never happen with normal transpose values (i.e. anything below and including octave 6). I don't see any problem in that, though, because sampling frequencies below 4000 Hz are pretty much NOT common at all, and it is so NOT common to transpose from G-2 to G-7 and back. If you really mean that this insane change in transposition is so important to be fixed, I'll keep the bug report open, but I won't fix it because I think it's absolutely unneccessary, it's just nitpicking.

Harbinger · July 18, 2009, 23:18:13

No problem, Jojo. I'll do some more tests this week. It may be happening only under certain unusual conditions, so i wanna make sure i give you correct and reproducible information. I'll visit this thread again next Friday. :wink:

Harbinger · July 24, 2009, 21:07:40

OK, i could not replicate this bug. I tried entering the frequency manually then going back and forth between Transpose notes, even in extreme octaves. Tried different tunings, both scales and temperaments. Tried different formats.

Go ahead and close this bug report. Apparently i found it under an unusual case, and if i happen upon that again and the bug shows itself, i'll mention it here. For now, though, the bug is untestable.

Thanks Jojo.