[PD] Scaling values in pd

[Mario Mey]

El 03/09/13 10:11, Lorenzo Sutton escribió:
> On 03/09/2013 14:06, Mario Mey wrote:
>> IOhannes, you are right only in these cases:
>>
>> 0 127 0 500
>> 0 300 0 1
>> ...
>>
>> But, if I need:
>>
>> 50 10 0 500|
> (assuming you want to map have 50 mapped to the 'minimum'):
> |
> [- 10]
> |
> [t b f]
> |   /
> [-  ]
> |
> [* 12.5]
> |
>
>> 3000 -3000 0.5 0.6
> [* -1]
> |
> [+ 3000]
> |
> [/ 60000]
> |
> [+ 0.5]

Yes, you can make this math every scale you need... or use an abstract 
that does the same automatically. In my patch, I use 284 lin-eq-conv 
objects. I didn't want to think how to make that math... and change 
everytime (normally, I create a lin-eq-conv and change its values a lot 
of times).

Best.


>
>> ...
>>
>> I will need a "linear equation conversion". As I wrote in last mail, 
>> I was needing something like this, first in ActionScript... then in 
>> Python... but I never could did it. Now, I needed again in Pd... so, 
>> I made lin-eq-conv.pd with extrapolation and lin-eq-conv-clip.pd for 
>> clipped values. I made it as neat as I could, to see how it works. 
>> Using x0-x1 and y0-y1, it uses expr to get "a" and "b" at load. Then, 
>> it only computes "aX + b = Y".
>>
>> I attach the lin-eq-conv.pd, lin-eq-conv-clip.pd and 
>> lin-eq-conv-help.pd.
>>
>> Also, I have some issues using [autoscale]. I start giving values and 
>> it outputs only 1. Then, I start to down the input and, then, it 
>> shows the real output. Maybe it's about this version (0.43.4 
>> Pd-Extended 64bits).
>>
>> PD: translated to Python:
>>
>> |def lin_eq_conv(x, x0, x1, y0, y1):
>>    a = (y0 - y1) / (x0 - x1)
>>    b = (a * x0) + y0
>>    return a * x + b|
>>
>>
>>
>>
>> El 03/09/13 03:32, IOhannes zmölnig escribió:
>>> On 09/02/2013 06:17 PM, hghoyer wrote:
>>>> Hi,
>>>>
>>>> in Max/MSP there is an object for simple scaling.
>>>>
>>>> If you create in MAX these object with this arguments:
>>>>
>>>> [scale 0 127 0 500] incomming messages from 0 to 127 are automatically
>>>> scaled from 0 to 500...
>>> honestly i'm of the firm conviction that you should learn how scaling
>>> works: it really is only a matter of adding, multiplying, dividing,
>>> subtracting - stuff you should heave learned in primary school.
>>>
>>> as frank pointed out, this should do for you:
>>>
>>> |
>>> [/ 127]
>>> |
>>> [* 500]
>>> |
>>>
>>> if you find it too tedious to do the maths over and over again, you
>>> might want to create an abstraction.
>>>
>>>
>>> being able to solve trivial problems like this will surely empower you
>>> to solve more complex problems :-)
>>>
>>> gamsdr
>>> IOhannes
>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> Pd-list at iem.at  mailing list
>>> UNSUBSCRIBE and account-management 
>>> ->http://lists.puredata.info/listinfo/pd-list
>>
>>
>>
>> _______________________________________________
>> Pd-list at iem.at mailing list
>> UNSUBSCRIBE and account-management -> 
>> http://lists.puredata.info/listinfo/pd-list
>
>
> _______________________________________________
> Pd-list at iem.at mailing list
> UNSUBSCRIBE and account-management -> 
> http://lists.puredata.info/listinfo/pd-list