;;+ ;FUNCTION: mvn_lpw_r_block16_byte ;PURPOSE: ; Decomutater for the compression for block=16 ; This routine is called by the mvn_lpw_r_header_l0.pro ;-------------- from SECTION 8.9 data compression explination --------------- ;32 blocks of 16 bits or 16 blocks of 32 bits (20 bits sign estended) ;first value is the absolute the following points are the fluctuation, PADING might have been used ;EUV and AVG have both 16 and 32 sample blocks (swp_i 16 and swp_v 32) ;broken into multiple 64 byte compression blocks ;16 block ;first 16 block starts with 20 bit sample followed ND as 9 bits then 16 deltas of length ND ;the 9 bits are P3P2P1P0D4D3D2D1D0 EDACbits then ND-1bits P3=D4 and P2 =D1xorD2xorD3; P1=D0xorD2xorD3; P0=D0xorD1xorD3xorD4 ;second 16 block with 20 bit sample folloewd with 9 bit ND ND - 20 delta is disable ; ; ;For this one it is as follows: ;I1(0-15) I1(16-31) I1(32-47) I1(48-63) I1(64-79) I1(80-95) I1(96-111) I1(112-127) V2(0-31) V2(32-63) V2(64-95) V2(96-127) ; ;USAGE: ; mvn_lpw_pkt_r_block16_byte,data,ptr,nn_e,mask8,bin_c,index_arr,arr_size,edac_on ; ;INPUTS: ; data this should be the unsigned array of the read in file ; ptr this is which element to start with in the data array and updated ptr ; nn_e this is which bit to start with based on element 'ptr' and array 'data' and update nn_e ; mask8 this is to convert the value in the array 'data' to a 8-bit binary array ; bin_c this is an index array to convert the bit's to a integer ; index_arr this is either 16 or 32. Set fix when calling on the routine ; edac_on: constains information if errors was found in the de-compression ;OUTPUTS ; output: the result of the read resulting in an array with values form the array 'data' ; ;KEYWORDS: ; ; ;CREATED BY: Laila Andersson 17 august 2011 ;FILE: mvn_lpw_pkt_r_block16_byte.pro ;VERSION: 1.1 ; Changes: this routine originally was for a word-array and modified for L0-data 2013 May 10 by Laila Andersson ;LAST MODIFICATION: 05/16/13 ;- function mvn_lpw_r_block16_byte, data, ptr, nn_e, mask8, bin_c, index_arr, arr_size, edac_on edac_on=0 n_loops=arr_size/16 ; how many time to do the loop for; examples for EUV 16/16 =1 and AVG 128/16=8 output=fltarr(arr_size) reject=0 ;check to see everything went smoothly temp_16=fltarr(16) ; this is the number of points in this 16*1=16 or 16*8 = 128 total number of values in each sweep ; the maxumum size length in bits:20+9+20*15 (i.e. 16 values) tmp_a=intarr(8) ptr_size=(20+9+20*15)/8 < ((n_elements(data)-ptr)-1) ; the maxumum size length in bytes to request data from and not longer than the data array IF (ptr+ptr_size)/n_loops LE n_elements(data) THEN $ ; make sure that there is enough of data in the data_array to meet arr_size for nii = 0,n_loops-1 do begin ; 1 or 8 sets based on arr_size bin=intarr(8) ; newfile_byte is 8 bit array, bin needs to be reset for each 16 block values ptr_end=long(ptr+ptr_size) ; grab a 16 block at the time ; < (n_elements(data)-1) ;(length2[i]-len_offset-4)/8 +1 ; this should be based on the (length2[i]-len_offset-4) tmp=data[ptr:ptr_end] ; get all the bytes that is in the array, ptr_end should be large enough to garantee this bin[*]=mask8[tmp[0],*] ; take the first byte and create the start array of 8 bits in bin-array for ui=1,ptr_size-1 do begin ; loop over al bytes and put them as bits into the bin-array - from here bin is independent on how the data was read tmp_a[*]=mask8[tmp[ui],*] bin=[bin,tmp_a] endfor ;the following should be independen how the data was read in: nd_1=bin[24+nn_e]*16+bin[25+nn_e]*8+bin[26+nn_e]*4+bin[27+nn_e]*2+bin[28+nn_e] ; should be first 4 EDAC bits then 5 bit which is ND-1 ; bin xor bin xor bin == (bin+bin+bin) mod 2 if bin[20+nn_e] NE bin[24+nn_e] AND $ ;P3 = D4 bin[21+nn_e] NE (bin[27+nn_e]+bin[26+nn_e]+bin[25+nn_e]) mod 2 AND $ ;P2 = D1 xor D2 xor D3 bin[22+nn_e] NE (bin[28+nn_e]+bin[26+nn_e]+bin[25+nn_e]) mod 2 AND $ ;P1 = D0 xor D2 xor D3 bin[23+nn_e] NE (bin[28+nn_e]+bin[27+nn_e]+bin[25+nn_e]+bin[24+nn_e]) mod 2 THEN edac_on=1 ;P0 = D0 xor D1 xor D3 xor D4 ; Print,' Warning the EDAC and ND values did not agree in this 16 block compression ',bin(nn_e+0) ;use this later to fluch this package if bin[20+nn_e] NE bin[24+nn_e] AND $ ;P3 = D4 bin[21+nn_e] NE (bin[27+nn_e]+bin[26+nn_e]+bin[25+nn_e]) mod 2 AND $ ;P2 = D1 xor D2 xor D3 bin[22+nn_e] NE (bin[28+nn_e]+bin[26+nn_e]+bin[25+nn_e]) mod 2 AND $ ;P1 = D0 xor D2 xor D3 bin[23+nn_e] NE (bin[28+nn_e]+bin[27+nn_e]+bin[25+nn_e]+bin[24+nn_e]) mod 2 THEN reject=reject+1 ;P0 = D0 xor D1 xor D3 xor D4 ; 2^n calculated where bin_c(n)=2^n -- bin is the long array with 1's and 0's -- Uses the first 20 to get the value - should the first be the sign? ; to flip a bit == (1-a)*(1+1) temp_16[0] = (1-bin[0+nn_e])* total( bin[1+nn_e:19+nn_e] *bin_c[18-index_arr[0:18]])- $ ; if the value is positive bin[0+nn_e] *(total((1-bin[1+nn_e:19+nn_e])*(1+bin[1+nn_e:19+nn_e])*bin_c[18-index_arr[0:18]])+1) nd=nd_1+1 ; the recorded information is ND-1 nn_g=9+20+nn_e ; where the second number (the delta) starts with respect to ptr IF nn_g+14*nd+nd-1 GE n_elements(bin) then nd=(n_elements(bin)-nn_g)/15 IF nd EQ 1 then $ ; for ii=0,14 do temp_16[ii+1]=temp_16[ii]+ $ ; first value is the sign, then sum over the other nd-1 elements 0 - $ ; if it is positive bin[nn_g+0+ii*nd] $ ; if it is negative ELSE $ ;nd EQ 20 IF nd EQ 2 then $ for ii=0,14 do temp_16[ii+1]=temp_16[ii]+ $ ; first value is the sign, then sum over the other nd-1 elements (1-bin[nn_g+0+ii*nd])* bin[nn_g+ii*nd+1] -$ ; if it is positive bin[nn_g+0+ii*nd] *((1-bin[nn_g+ii*nd+1])*(1+bin[nn_g+ii*nd+1])+1) $ ; if it is negative ELSE $ ;if nd is not ==1,2 IF nd EQ 20 then $ ; ND == 20 decoding disable for ii=0,14 do temp_16[ii+1]= $ ; first value is the sign, then sum over the other nd-1 elements (1-bin[nn_g+0+ii*nd])* total( bin[nn_g+ii*nd+1:nn_g+ii*nd+19] *bin_c[18-index_arr[0:18]] ) -$ ; if it is positive bin[nn_g+0+ii*nd] *(total((1-bin[nn_g+ii*nd+1:nn_g+ii*nd+19])*(1+bin[nn_g+ii*nd+1:nn_g+ii*nd+19])*bin_c[18-index_arr[0:18]] )+1) $ ; if it is negative ELSE $ ; the rest not special cases for ii=0,14 do temp_16[ii+1]=temp_16[ii]+ $ ; first value is the sign, then sum over the other nd-1 elements (1-bin[nn_g+0+ii*nd])* total( bin[nn_g+ii*nd+1:nn_g+ii*nd+nd-1] *bin_c[nd-2-index_arr[0:nd-2]] ) -$ ; if it is positive bin[nn_g+0+ii*nd] *(total((1-bin[nn_g+ii*nd+1:nn_g+ii*nd+nd-1])*(1+bin[nn_g+ii*nd+1:nn_g+ii*nd+nd-1])*bin_c[nd-2-index_arr[0:nd-2]] )+1) ; if it is negative nn_e=(nn_g+nd*15) mod 16 ; ptr=ptr+floor(1.0*(nn_g+nd*15)/16)*2 ; 20+9+nd*15 bit should one block contain and newfile_byte is a 8 bit array ptr_size=(20+9+20*15)/8 < ((n_elements(data)-ptr)-1) ; as ptr is moved calculate the new ptr_size - same line as before the loop output[nii*16+0: nii*16+15] = temp_16 ; this is just 1/8 of the points endfor ELSE ptr= n_elements(data) ; change ptr if end array has been reached "IF (ptr+ptr_size)/8 LE n_elements(data) THEN" If reject GT 0 then output=output*0 ; there was an issue during this packet return,output end