i did some more tests on array semantics; in particular i compared 
i-arrays and tables.  as the usage of tables is so old in csound, it 
might be like a model for array usage.

all good so far, but what did not work is to set an i-array at k-time. 
the line
   giArray[kIndx] = -2
leads to the error message:
   Cant set i-array at k-rate
	##array_set.e	giArray	#i0	kIndx	

according to http://github.com/csound/csound/wiki/Array-Semantics i 
think it should work.  and it would be good to see it working — if 
nothing is against — because in function tables we do it regularily.

the complete example / comparison is pasted below.

	joachim




-n -m0



sr = 44100
ksmps = 32
nchnls = 2
0dbfs = 1

giTable ftgen 0, 0, 7, -2, 1, 2, 3, 4, 5, 6, 7
giArray[] fillarray 1, 2, 3, 4, 5, 6, 7


instr Read_i ;get i-val from i-table or i-array
iIndx = 2
iTabEl table iIndx, giTable
iArrEl = giArray[iIndx]
print iTabEl, iArrEl
endin

instr Read_k ;get k-val from i-table or i-array
kIndx init 2
kTabEl table kIndx, giTable
kArrEl = giArray[kIndx]
printks "kTable = %.3f  kArrEl = %.3f\n", 0, kTabEl, kArrEl
turnoff
endin

instr Write_i ;change table at i and call Read-i/k
iIndx = 2
tablew -2, iIndx, giTable
giArray[iIndx] = -2
schedule "Read_i", 0, 1
schedule "Read_k", 0, 1
endin

instr Write_k ;change table at k and call Read-i/k
kIndx init 2
tablew -2, kIndx, giTable
giArray[kIndx] = -2
schedule "Read_i", 0.1, 1
schedule "Read_k", 0.1, 1
turnoff
endin



i "Read_i" 0 1
i "Read_k" 1 1
i "Write_i" 2 1
i "Write_k" 3 1

According to what we discussed, i-rate arrays cannot be set at perf-time.

Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy
Maynooth University
Ireland

> On 4 Nov 2016, at 17:51, joachim heintz  wrote:
> 
> i did some more tests on array semantics; in particular i compared i-arrays and tables.  as the usage of tables is so old in csound, it might be like a model for array usage.
> 
> all good so far, but what did not work is to set an i-array at k-time. the line
>  giArray[kIndx] = -2
> leads to the error message:
>  Cant set i-array at k-rate
>    ##array_set.e    giArray    #i0    kIndx    
> 
> according to http://github.com/csound/csound/wiki/Array-Semantics i think it should work.  and it would be good to see it working — if nothing is against — because in function tables we do it regularily.
> 
> the complete example / comparison is pasted below.
> 
>    joachim
> 
> 
> 
> 
> -n -m0
> 
> 
> 
> sr = 44100
> ksmps = 32
> nchnls = 2
> 0dbfs = 1
> 
> giTable ftgen 0, 0, 7, -2, 1, 2, 3, 4, 5, 6, 7
> giArray[] fillarray 1, 2, 3, 4, 5, 6, 7
> 
> 
> instr Read_i ;get i-val from i-table or i-array
> iIndx = 2
> iTabEl table iIndx, giTable
> iArrEl = giArray[iIndx]
> print iTabEl, iArrEl
> endin
> 
> instr Read_k ;get k-val from i-table or i-array
> kIndx init 2
> kTabEl table kIndx, giTable
> kArrEl = giArray[kIndx]
> printks "kTable = %.3f  kArrEl = %.3f\n", 0, kTabEl, kArrEl
> turnoff
> endin
> 
> instr Write_i ;change table at i and call Read-i/k
> iIndx = 2
> tablew -2, iIndx, giTable
> giArray[iIndx] = -2
> schedule "Read_i", 0, 1
> schedule "Read_k", 0, 1
> endin
> 
> instr Write_k ;change table at k and call Read-i/k
> kIndx init 2
> tablew -2, kIndx, giTable
> giArray[kIndx] = -2
> schedule "Read_i", 0.1, 1
> schedule "Read_k", 0.1, 1
> turnoff
> endin
> 
> 
> 
> i "Read_i" 0 1
> i "Read_k" 1 1
> i "Write_i" 2 1
> i "Write_k" 3 1
> 

in the wiki which i quoted is written:
"isig[k] - outputs k-val"

if there are reasons not to allow this, ok.  otherwise i'd think it 
would be useful and would fit to the usage of tables.


On 04/11/16 19:52, Victor Lazzarini wrote:
> According to what we discussed, i-rate arrays cannot be set at perf-time.
>
> Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy
> Maynooth University
> Ireland
>
>> On 4 Nov 2016, at 17:51, joachim heintz  wrote:
>>
>> i did some more tests on array semantics; in particular i compared i-arrays and tables.  as the usage of tables is so old in csound, it might be like a model for array usage.
>>
>> all good so far, but what did not work is to set an i-array at k-time. the line
>>  giArray[kIndx] = -2
>> leads to the error message:
>>  Cant set i-array at k-rate
>>    ##array_set.e    giArray    #i0    kIndx
>>
>> according to http://github.com/csound/csound/wiki/Array-Semantics i think it should work.  and it would be good to see it working — if nothing is against — because in function tables we do it regularily.
>>
>> the complete example / comparison is pasted below.
>>
>>    joachim
>>
>>
>> 
>> 
>> -n -m0
>> 
>> 
>>
>> sr = 44100
>> ksmps = 32
>> nchnls = 2
>> 0dbfs = 1
>>
>> giTable ftgen 0, 0, 7, -2, 1, 2, 3, 4, 5, 6, 7
>> giArray[] fillarray 1, 2, 3, 4, 5, 6, 7
>>
>>
>> instr Read_i ;get i-val from i-table or i-array
>> iIndx = 2
>> iTabEl table iIndx, giTable
>> iArrEl = giArray[iIndx]
>> print iTabEl, iArrEl
>> endin
>>
>> instr Read_k ;get k-val from i-table or i-array
>> kIndx init 2
>> kTabEl table kIndx, giTable
>> kArrEl = giArray[kIndx]
>> printks "kTable = %.3f  kArrEl = %.3f\n", 0, kTabEl, kArrEl
>> turnoff
>> endin
>>
>> instr Write_i ;change table at i and call Read-i/k
>> iIndx = 2
>> tablew -2, iIndx, giTable
>> giArray[iIndx] = -2
>> schedule "Read_i", 0, 1
>> schedule "Read_k", 0, 1
>> endin
>>
>> instr Write_k ;change table at k and call Read-i/k
>> kIndx init 2
>> tablew -2, kIndx, giTable
>> giArray[kIndx] = -2
>> schedule "Read_i", 0.1, 1
>> schedule "Read_k", 0.1, 1
>> turnoff
>> endin
>>
>> 
>> 
>> i "Read_i" 0 1
>> i "Read_k" 1 1
>> i "Write_i" 2 1
>> i "Write_k" 3 1
>> 
>> 

Outputs -- means you can read it at k-time,
but not write to it.

Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy
Maynooth University
Ireland

> On 4 Nov 2016, at 20:25, joachim heintz  wrote:
> 
> in the wiki which i quoted is written:
> "isig[k] - outputs k-val"
> 
> if there are reasons not to allow this, ok.  otherwise i'd think it would be useful and would fit to the usage of tables.
> 
> 
>> On 04/11/16 19:52, Victor Lazzarini wrote:
>> According to what we discussed, i-rate arrays cannot be set at perf-time.
>> 
>> Victor Lazzarini
>> Dean of Arts, Celtic Studies, and Philosophy
>> Maynooth University
>> Ireland
>> 
>>> On 4 Nov 2016, at 17:51, joachim heintz  wrote:
>>> 
>>> i did some more tests on array semantics; in particular i compared i-arrays and tables.  as the usage of tables is so old in csound, it might be like a model for array usage.
>>> 
>>> all good so far, but what did not work is to set an i-array at k-time. the line
>>> giArray[kIndx] = -2
>>> leads to the error message:
>>> Cant set i-array at k-rate
>>>   ##array_set.e    giArray    #i0    kIndx
>>> 
>>> according to http://github.com/csound/csound/wiki/Array-Semantics i think it should work.  and it would be good to see it working — if nothing is against — because in function tables we do it regularily.
>>> 
>>> the complete example / comparison is pasted below.
>>> 
>>>   joachim
>>> 
>>> 
>>> 
>>> 
>>> -n -m0
>>> 
>>> 
>>> 
>>> sr = 44100
>>> ksmps = 32
>>> nchnls = 2
>>> 0dbfs = 1
>>> 
>>> giTable ftgen 0, 0, 7, -2, 1, 2, 3, 4, 5, 6, 7
>>> giArray[] fillarray 1, 2, 3, 4, 5, 6, 7
>>> 
>>> 
>>> instr Read_i ;get i-val from i-table or i-array
>>> iIndx = 2
>>> iTabEl table iIndx, giTable
>>> iArrEl = giArray[iIndx]
>>> print iTabEl, iArrEl
>>> endin
>>> 
>>> instr Read_k ;get k-val from i-table or i-array
>>> kIndx init 2
>>> kTabEl table kIndx, giTable
>>> kArrEl = giArray[kIndx]
>>> printks "kTable = %.3f  kArrEl = %.3f\n", 0, kTabEl, kArrEl
>>> turnoff
>>> endin
>>> 
>>> instr Write_i ;change table at i and call Read-i/k
>>> iIndx = 2
>>> tablew -2, iIndx, giTable
>>> giArray[iIndx] = -2
>>> schedule "Read_i", 0, 1
>>> schedule "Read_k", 0, 1
>>> endin
>>> 
>>> instr Write_k ;change table at k and call Read-i/k
>>> kIndx init 2
>>> tablew -2, kIndx, giTable
>>> giArray[kIndx] = -2
>>> schedule "Read_i", 0.1, 1
>>> schedule "Read_k", 0.1, 1
>>> turnoff
>>> endin
>>> 
>>> 
>>> 
>>> i "Read_i" 0 1
>>> i "Read_k" 1 1
>>> i "Write_i" 2 1
>>> i "Write_k" 3 1
>>> 
>>> 

i have written now a new section in 
http://write.flossmanuals.net/csound/e-arrays/ called "More on Array 
Rates".  i put it here below and appreciate any corrections.

while writing, i was wondering whether i(kArr[i]) should perhaps better 
yield an error, instead of always returning zero, for instance:

kArray[] fillarray 1, 2, 3
iFirst = i(kArray[0])
print iFirst

returns iFirst as zero which is obviously nonsense from the user's point 
of view, but can be hidden as a wrong result because it passes silently.

what do you think?

	joachim


#################################

More on Array Rates

Usually the first character of a variable name in Csound shows whether 
it is i-rate or k-rate or a-rate. But for arrays, we have actually two 
signifiers: the array variable name, and the index type. If both 
coincide, it is easy:

    *  i_array[i_index] reads and writes at i-time
    *  k_array[k_index] reads and writes at k-time
    *  a_array[a_index] reads and writes at a-time

But what to do if array type and index type do not coincide? In general, 
the index type will then determine whether the array is read or written 
only once (at init-time) or at each k-cycle. This is valid in particular 
for S_arrays (containing strings) and f_arrays (containing f-data). 
Other cases are:

    *  i_array[k_index] reads at k-time; writing is not possible (yields 
a runtime error)
    *  k_array[i_index] reads and writes at k-rate
    *  a_array[i_index] reads and writes at a-rate

For usual k-variables, you can get the value at init-time via the 
expression i(kVar), for instance:

instr 1
  gkLine linseg 1, 1, 2
  schedule 2, .5, 0
endin
instr 2
  iLine = i(gkLine)
  print iLine
endin

will print: iLine = 1.499.

This expression cannot be used for arrays:

kArray[] fillarray 1, 2, 3
iFirst = i(kArray[0])
print iFirst

will print: iFirst = 0.000, which is obviously not what could be 
expected. For this purpose, the i() expression can be used to pass the 
index as second argument:

kArray[] fillarray 1, 2, 3
iFirst = i(kArray, 0)
print iFirst

will print: iFirst = 1.000.

############################



On 04/11/16 21:34, Victor Lazzarini wrote:
> Outputs -- means you can read it at k-time,
> but not write to it.
>
> Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy
> Maynooth University
> Ireland
>
>> On 4 Nov 2016, at 20:25, joachim heintz  wrote:
>>
>> in the wiki which i quoted is written:
>> "isig[k] - outputs k-val"
>>
>> if there are reasons not to allow this, ok.  otherwise i'd think it would be useful and would fit to the usage of tables.
>>
>>
>>> On 04/11/16 19:52, Victor Lazzarini wrote:
>>> According to what we discussed, i-rate arrays cannot be set at perf-time.
>>>
>>> Victor Lazzarini
>>> Dean of Arts, Celtic Studies, and Philosophy
>>> Maynooth University
>>> Ireland
>>>
>>>> On 4 Nov 2016, at 17:51, joachim heintz  wrote:
>>>>
>>>> i did some more tests on array semantics; in particular i compared i-arrays and tables.  as the usage of tables is so old in csound, it might be like a model for array usage.
>>>>
>>>> all good so far, but what did not work is to set an i-array at k-time. the line
>>>> giArray[kIndx] = -2
>>>> leads to the error message:
>>>> Cant set i-array at k-rate
>>>>   ##array_set.e    giArray    #i0    kIndx
>>>>
>>>> according to http://github.com/csound/csound/wiki/Array-Semantics i think it should work.  and it would be good to see it working — if nothing is against — because in function tables we do it regularily.
>>>>
>>>> the complete example / comparison is pasted below.
>>>>
>>>>   joachim
>>>>
>>>>
>>>> 
>>>> 
>>>> -n -m0
>>>> 
>>>> 
>>>>
>>>> sr = 44100
>>>> ksmps = 32
>>>> nchnls = 2
>>>> 0dbfs = 1
>>>>
>>>> giTable ftgen 0, 0, 7, -2, 1, 2, 3, 4, 5, 6, 7
>>>> giArray[] fillarray 1, 2, 3, 4, 5, 6, 7
>>>>
>>>>
>>>> instr Read_i ;get i-val from i-table or i-array
>>>> iIndx = 2
>>>> iTabEl table iIndx, giTable
>>>> iArrEl = giArray[iIndx]
>>>> print iTabEl, iArrEl
>>>> endin
>>>>
>>>> instr Read_k ;get k-val from i-table or i-array
>>>> kIndx init 2
>>>> kTabEl table kIndx, giTable
>>>> kArrEl = giArray[kIndx]
>>>> printks "kTable = %.3f  kArrEl = %.3f\n", 0, kTabEl, kArrEl
>>>> turnoff
>>>> endin
>>>>
>>>> instr Write_i ;change table at i and call Read-i/k
>>>> iIndx = 2
>>>> tablew -2, iIndx, giTable
>>>> giArray[iIndx] = -2
>>>> schedule "Read_i", 0, 1
>>>> schedule "Read_k", 0, 1
>>>> endin
>>>>
>>>> instr Write_k ;change table at k and call Read-i/k
>>>> kIndx init 2
>>>> tablew -2, kIndx, giTable
>>>> giArray[kIndx] = -2
>>>> schedule "Read_i", 0.1, 1
>>>> schedule "Read_k", 0.1, 1
>>>> turnoff
>>>> endin
>>>>
>>>> 
>>>> 
>>>> i "Read_i" 0 1
>>>> i "Read_k" 1 1
>>>> i "Write_i" 2 1
>>>> i "Write_k" 3 1
>>>> 
>>>> 
>>>