Audible when rendering xanadu.csd to dac on the current develop branch
on Ubuntu 16.04. The first two instruments produce a warbling sound
that did not used to be audible and that is clearly wrong. I suspect
some regression in the delayw, delayr, or deltap3 opcodes. CPU use
also seems high but not sure what it used to be.

Regards,
Mike

-----------------------------------------------------
Michael Gogins
Irreducible Productions
http://michaelgogins.tumblr.com

Not aware of changes in those opcodes.  Have you tried bisecting it?

About to sleep so may look in the morning.



Sent from TypeApp



On Oct 29, 2017, 01:59, at 01:59, Michael Gogins  wrote:
>Audible when rendering xanadu.csd to dac on the current develop branch
>on Ubuntu 16.04. The first two instruments produce a warbling sound
>that did not used to be audible and that is clearly wrong. I suspect
>some regression in the delayw, delayr, or deltap3 opcodes. CPU use
>also seems high but not sure what it used to be.
>
>Regards,
>Mike
>
>-----------------------------------------------------
>Michael Gogins
>Irreducible Productions
>http://michaelgogins.tumblr.com
>Michael dot Gogins at 

The nly warble I ear is wat I expect.  But I do not know te piece well.

Can you detect a time we it was as you expect?  Te we can bisect, but i am 
no sure for what I am looking.  Or a wav file of it correct?

I ave extracted instr 1 apart from the rest.

On Sat, 28 Oct 2017, Michael Gogins wrote:

> Audible when rendering xanadu.csd to dac on the current develop branch
> on Ubuntu 16.04. The first two instruments produce a warbling sound
> that did not used to be audible and that is clearly wrong. I suspect
> some regression in the delayw, delayr, or deltap3 opcodes. CPU use
> also seems high but not sure what it used to be.
>
> Regards,
> Mike
>
> -----------------------------------------------------
> Michael Gogins
> Irreducible Productions
> http://michaelgogins.tumblr.com
> Michael dot Gogins at gmail dot com

Only change I can see are variants on 2017 Feb 16

-    tap = q->curp - (int32) (FL(0.5) + *p->xdlt * csound->esr);
+    tap = q->curp - MYFLT2LRND(*p->xdlt * csound->esr);
      while (tap < (MYFLT *) q->auxch.auxp)

I reverted this change locally ad at least instr1 generates identical wav 
files.  So must be other than inst 1


On Sun, 29 Oct 2017, Michael Gogins wrote:

> Thanks, I'll look at this.
> 
> On Oct 29, 2017 1:01 PM, "jpff"  wrote:
>       Only change I can see are variants on 2017 Feb 16
>
>       -    tap = q->curp - (int32) (FL(0.5) + *p->xdlt * csound->esr);
>       +    tap = q->curp - MYFLT2LRND(*p->xdlt * csound->esr);
>            while (tap < (MYFLT *) q->auxch.auxp)
>
>       Are these not the same?
> 
> 
>

The warbling sound in instr 1 of xanadu.csd went away after I made the
following change:

;kvib        poscil      1/120, ipch/50, 1       ;vibrato
kvib        poscil      0.008333333, ipch/50, 1       ;vibrato

This indicates that expression parsing or parameter parsing has
changed in a breaking way.

I will do an extremely clean pull and rebuild and let you know if the
problem persists.

Best,
Mike

-----------------------------------------------------
Michael Gogins
Irreducible Productions
http://michaelgogins.tumblr.com
Michael dot Gogins at gmail dot com


On Sun, Oct 29, 2017 at 2:31 PM, jpff  wrote:
> I reverted this change locally ad at least instr1 generates identical wav
> files.  So must be other than inst 1
>
>
>
> On Sun, 29 Oct 2017, Michael Gogins wrote:
>
>> Thanks, I'll look at this.
>>
>> On Oct 29, 2017 1:01 PM, "jpff"  wrote:
>>       Only change I can see are variants on 2017 Feb 16
>>
>>       -    tap = q->curp - (int32) (FL(0.5) + *p->xdlt * csound->esr);
>>       +    tap = q->curp - MYFLT2LRND(*p->xdlt * csound->esr);
>>            while (tap < (MYFLT *) q->auxch.auxp)
>>
>>       Are these not the same?
>>
>>

OK; tis is cnstant folding not beig correct i division case
I have bee working on it so up to me now


On Sun, 29 Oct 2017, Michael Gogins wrote:

> The warbling sound in instr 1 of xanadu.csd went away after I made the
> following change:
>
> ;kvib        poscil      1/120, ipch/50, 1       ;vibrato
> kvib        poscil      0.008333333, ipch/50, 1       ;vibrato
>
> This indicates that expression parsing or parameter parsing has
> changed in a breaking way.
>
> I will do an extremely clean pull and rebuild and let you know if the
> problem persists.
>
> Best,
> Mike
>
> -----------------------------------------------------
> Michael Gogins
> Irreducible Productions
> http://michaelgogins.tumblr.com
> Michael dot Gogins at gmail dot com
>
>
> On Sun, Oct 29, 2017 at 2:31 PM, jpff  wrote:
>> I reverted this change locally ad at least instr1 generates identical wav
>> files.  So must be other than inst 1
>>
>>
>>
>> On Sun, 29 Oct 2017, Michael Gogins wrote:
>>
>>> Thanks, I'll look at this.
>>>
>>> On Oct 29, 2017 1:01 PM, "jpff"  wrote:
>>>       Only change I can see are variants on 2017 Feb 16
>>>
>>>       -    tap = q->curp - (int32) (FL(0.5) + *p->xdlt * csound->esr);
>>>       +    tap = q->curp - MYFLT2LRND(*p->xdlt * csound->esr);
>>>            while (tap < (MYFLT *) q->auxch.auxp)
>>>
>>>       Are these not the same?
>>>
>>>
>>

I think it is fixed in git.  Thanks for finding this -- it was an oversight 
of mine in reconstructing the parse tree.

Thank you, that fixed it.

Best,
Mike

-----------------------------------------------------
Michael Gogins
Irreducible Productions
http://michaelgogins.tumblr.com
Michael dot Gogins at gmail dot com


On Sun, Oct 29, 2017 at 4:31 PM, jpff  wrote:
> I think it is fixed in git.  Thanks for finding this -- it was an oversight
> of mine in reconstructing the parse tree.
>