[ Oops, switched out of the list with last answer, back to it in case it
is of interest for the community ;) ]
Bart, yes I am linking that to the smallest number question, as it
allows you to get not only *small numbers* but numbers *as close to
zero* as you want, without dividing by zero. You can eventually set a
max clipping value to the inverse parameter to avoid the
rounding-to-zero-when-number-is-really-too-close-to-zero. Did I miss
your point ?
Le 13/06/2014 21:03, Bart Brouns a écrit :
> Hi Vincent,
>
> On Fri, 2014-06-13 at 20:54 +0200, Vincent G. Listes wrote:
>> Hi Bart,
>>
>> I do not know if such an expression exists, but as I experienced it,
>> this situation often happens when you have a division downstream, and
>> you obviously do not want to divide by zero.
> Yeah, that's a good example of what I mean.
>
>
>> In such cases, I found it
>> sometimes relevant to take the inverse parameter. E.g. a "step size"
>> parameter can be turned into a "number of steps" parameter.
>> My 2 cents, not a proper answer to your question though.
> I can see how you could make the step size 1/nrSteps.
> What I dont follow ia: are you linking that to the smallest number
> question, or is it unrelated other than "you can get small numbers by
> dividing..."
>
> Thanks,
>
> Bart.
>
>> Vincent
>>
>>
>> Le 11/06/2014 11:47, Bart Brouns a écrit :
>>> Right. Thanks.
>>> Fixed it, so finally everyone can have a listen, and tell me what they
>>> think ;)
>>>
>>> That reminds me of another question: sometimes it is useful to have an
>>> expression denoting the smallest possible float bigger than 0.
>>> Maybe this case is not the best example, but still.
>>>
>>> I assume it's value depends on the compiler options, so it wouldn't be
>>> practical to have it as a constant number.
>>>
>>> Does such an expression exist?
>>>
>>> Cheers,
>>> Bart.
>>>
>>> On Wed, 2014-06-11 at 09:51 +0200, Sarah Denoux wrote:
>>>> Ok I see want you mean. The problem is that most architectures don't
>>>> support null steps. We are discussing what to do with them, but for now, I
>>>> think you should better specify a non-null step.
>>>>
>>>> Cheers,
>>>> Sarah
>>>>
>>>> Le 11 juin 2014 à 09:34, Bart Brouns a écrit :
>>>>
>>>>> Hi Sarah,
>>>>>
>>>>> What I mean is: I don't care how small the steps are.
>>>>> It seems to work for faust2jack.
>>>>> Does it give you problems on other architectures?
>>>>> Are there other reasons for not doing this?
>>>>>
>>>>> I did notice nobody else is doing it...
>>>>>
>>>>> Cheers,
>>>>> Bart.
>>>>>
>>>>> On Wed, 2014-06-11 at 09:27 +0200, Sarah Denoux wrote:
>>>>>> Hi Bart,
>>>>>>
>>>>>>
>>>>>> What I meant is in the slider declaration. You have for the factor :
>>>>>> factor = qompanderGroup(hslider("[0] factor[unit::1]",3, 0.8, 8,
>>>>>> 0):smooth(0.999));
>>>>>> The last parameter of the slider is 0 which is supposed to be the
>>>>>> step. What do you mean by step is zero ?
>>>>>>
>>>>>>
>>>>>> Cheers,
>>>>>> Sarah
>>>>>
>>>>>
>
>
------------------------------------------------------------------------------
HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions
Find What Matters Most in Your Big Data with HPCC Systems
Open Source. Fast. Scalable. Simple. Ideal for Dirty Data.
Leverages Graph Analysis for Fast Processing & Easy Data Exploration
http://p.sf.net/sfu/hpccsystems
_______________________________________________
Faudiostream-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/faudiostream-users
