ah haaa! (out loud... a number or times...)
this is great - i was so stuck trying to figure out wft was going on -
thanks for that sam!
this isnt mentioned here is it?:
http://savonet.sourceforge.net/doc-svn/language.html
because i stared at the section on "types" for some time scratching my head
thanks again sam - now i can move forward!
rob
On 19/03/13 13:59, Samuel Mimram wrote:
> Hi Rob,
>
> Quite a subtle question as usual :) There are basically two kinds of
> encoding of floats used in Liquidsoap. The first is for constants (as
> in del=300.) which are simply of type float. The other one is used for
> floats which can vary over time such as those produced by osc.float:
> osc.float("/sc/topoverdel", 300.) has type ()->float which mean that
> it is a function, which takes an empty list of arguments and returns
> the current value of the float. You could define a function of this
> type manually, for instance by
>
> def x() = 300. end
>
> or even
>
> {300.}
>
> the second being a short notation for the first one. There is a little
> trick used in Liquidsoap, which makes that most functions can either
> take a constant float (a float) or a varying float (a ()->float);
> Liquidsoap chooses the right type for you. You can see this in the
> type:
>
> $ ./liquidsoap -h ladspa.bandpass_a_iir
>
> Type: (?id:string,?bandwidth:'a,?center_frequency:'b,
> source(audio='#c,video='#d,midi='#e))->
> source(audio='#c,video='#d,midi='#e)
> where 'a, 'b is either float or ()->float
>
> * bandwidth : anything that is either float or ()->float (default:
> 295.831787109)
> Bandwidth (Hz) (4.41 <= <code>bandwidth</code> <= 19845).
>
> [...]
>
> This is the meaning of "either float or ()->float" above. However, the
> way you coded this forces deltime to be a float (and not a ()->float)
> because you multiply it by 1.5 and multiplication can only be
> performed between floats (and not ()->float). You can check this by
>
> liquidsoap -i test.liq
>
> pandelay :
> (source(audio=2,video=0,midi=0),float,'a)->source(audio=2,video=0,midi=0)
> where 'a is either float or ()->float
>
> We could extend multiplication for this case to work, but this is not
> done for now. So, if you want a ()->float as deltime, you have to
> handle the function explicitly, i.e. the following should be fine:
>
> def pandelay (z,deltime,f)
> z = ladspa.lcrdelay(id="", dry_wet_level=1., l_level=50.,
> c_level=50., r_level=50., l_delay=deltime, delay={deltime()*1.5},
> r_delay={deltime()*1.25}, low_damp=0., high_damp=0.,
> feedback=90.,spread=50., z)
> z=ladspa.tap_autopan(frequency=f, depth=80., z)
> z
> end
>
> I hope this helps!
>
> Cheers,
>
> Sam.
>
> PS : thanks for the typo on infer_r_ed, its now corrected!
>
> On Tue, Mar 19, 2013 at 1:37 PM, Rob Canning <[email protected]> wrote:
>> hi,
>> maybe i didn't give enough information...
>> perhaps it is something the way types need to be defined in function
>> arguments?
>> or i somehow need to evaluate the <fun> before passing it as an argument??
>>
>> On 15/03/13 15:34, Rob Canning wrote:
>>> hi!
>>>
>>> not sure why the value del here causes an error when uncommented while
>>> freqx work fine
>>> del is being passed on one function further but not sure why this causes
>>> a problem as its fine when i hardcode it with del=300.
>>>
>>> del=300.
>>> #del=osc.float("/sc/topoverdel", 300.)
>>> freqx=osc.float("/sc/topoverfreq", 11110.)
>>>
>>> topover =
>>> pandelay(ladspa.bandpass_a_iir(center_frequency=freqx,bandwidth=50.,
>>> topover), del, 0.3)
>> my function is like this:
>>
>> def pandelay (z,deltime,f)
>> z = ladspa.lcrdelay(id="", dry_wet_level=1.,
>> l_level=50.,c_level=50.,r_level=50.,
>> l_delay=deltime, c_delay=deltime*1.5,r_delay=deltime*1.25,
>> low_damp=0.,high_damp=0., feedback=90.,spread=50.,
>> z)
>> z=ladspa.tap_autopan(frequency=f, depth=80., z)
>> z
>> end
>>
>> the message from osc is created the exact same way as freqx so is
>> definitely a float
>>
>> any ideas?
>>
>> thanks,
>>
>> rob
>>> thanks
>>>
>>> rob
>>>
>>> At line 71, char 93:
>>> this value has type
>>> ()->_ (infered at line 60, char 14-37)
>>> but it should be a subtype of (the type of the value at line 10, char
>>> 24)
>>> float (infered at line 18, char 19-21)
>>>
>>>
>>> btw. inferred is spelt inferred not infered
>>>
>>>
>>
>> ------------------------------------------------------------------------------
>> Everyone hates slow websites. So do we.
>> Make your web apps faster with AppDynamics
>> Download AppDynamics Lite for free today:
>> http://p.sf.net/sfu/appdyn_d2d_mar
>> _______________________________________________
>> Savonet-users mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/savonet-users
> ------------------------------------------------------------------------------
> Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics
> Download AppDynamics Lite for free today:
> http://p.sf.net/sfu/appdyn_d2d_mar
> _______________________________________________
> Savonet-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/savonet-users
------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_mar
_______________________________________________
Savonet-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/savonet-users
