I am new to the CMSIS DSP libraries and was going through the convolution example www.keil.com/.../arm_convolution_example_f32_8c-example.html. I understand the flow and my only confusion is the output generated and stored in testRefOutput_f32[].
If I do a direct convolution using arm_conv_f32(), I get a different output (see below ConvOutput[]) which matches what I get in Matlab using the conv() function or doing an fft convolution (z=ifft(fft(x,length(x)+length(y)-1).*fft(y,length(x)+length(y)-1))) with the 2 input arrays testInputA_f32[] and testInputB_f32[].
I am a bit confused as to the different outputs and any insight would be very helpful.
ConvOutput[]={-0.755308018,1.229485034,0.10131416,1.016604633,3.502613814,-1.963654563,6.472972163,3.478031305,1.954949486,6.648379937,-3.996195718,1.792012056,-0.586438574,-7.447816242,-1.3 84041433,-10.65744205,-6.240046558,-6.524235584,-6.546002989,-3.739002659,-3.087535502,0.184150477,2.158228464,3.786510619,0.834278033,5.400613305,2.731501134,7.403361883,4.647 78187,2.080267431,-4.808539835,2.749673099,-7.002291217,2.748762471,-4.448640435,0.981555193,0.704749604,0.552720292,-3.828611537,4.703986428,-6.858765161,8.759337937,-10.91725 001,-3.994683449,-9.327700559,-5.014144908,-8.460611773,1.599489267,-15.68536376,-0.39188456,7.548501968,-2.311238772,2.634787105,1.761418676,0.969400228,10.43300656,-1.8420863 56,4.297158491,3.565491908,3.209464906,-3.744203172,5.569915527,-0.109928189,5.582024558,13.78780707,0.642280568,14.36921433,1.686617503,0.212125507,1.343078987,6.937862136,1.0 99071806,-0.251404117,-0.173721376,-7.797617973,6.528383368,-8.177245377,1.726289794,2.163856296,11.02343867,-1.56735296,-4.690185416,2.686433843,3.9438**141,11.81517545,5.1683 66058,2.191606134,1.55918133,4.009872096,2.852416683,1.00816686,-6.369077828,-1.809436282,11.83796165,-4.108340361,4.469680087,3.357955135,-2.674046562,8.960443814,-6.809969954 ,-0.914164719,9.112410682,5.461497019,-1.336679247,-2.214374312,-1.564298107,-6.913312679,8.904012238,-3.878144422,6.878010209,-5.203537038,3.369907792,0.845670751,-2.007356641 ,-1.06352301,1.370684564,1.200765603,-6.276606386,3.905318469,-3.352637676,1.498066762,0.902210557,0.602243835,-1.40873175,0.312370773,0.34844076,-0.032509683}
[Some of the numbers might have '*' as the robot detects some numbers as spam]