Decimation and interpolation problem

[jimmylin0226]

decimation and interpolation/

When creating interpolation simulink model,
I got two different result from these two models.

Assume I want to create upsampling 2x system.
The first system, upsampling 2x then halfband FIR filtering.
The second system, only use one polyphase IIR filter.

I think these two systems should be the same amplitude after filtering.
But I got half output amplitude in the system one.
Ex, in system one, 1 Vpp sin wave in , and get 0.5 Vpp sin wave out.
But in system two, 1 Vpp sin wave in , and get 1 Vpp sin wave out.

The same situation doesn't happen in the decimation model.

Does anyone can explain this case ?

 

[hobgoblin]

decimation and interpolation problems

I don't know the details of polyphase IIR operation but the result of the first method seems to be correct. When You upsample signal by 2 you add one zero between each pair of samples. It gives gives the reduction of amplitude at the output of the HB filter in two. Try to upsample by 4 and you'll get 4x reduction.