Very long time load FIR table big coefficients
Hallo,
I try to load in FIR filter around 9000 float 32bit coefficients for ADAU1452 under Sigma Studio 3.12.3.
I use "Load table value from file..." . It work, but it take for 9000 taps .....a few hours .
P.S. My PC : last core i7 with SSD
Perhaps someone could have a quicker solution.....
Audio Signal Processors
adau1452
SigmaDSP Audio Processors
Greetings Dimitri,
I was unable to reproduce your case. Are you able to share your SigmaStudio project, or a step by step procedure of how to reproduce this issue? It will also help to see some screen shots.
Thank you.
Wilfrido Sierra
Hi Wilfrido,
Thank you for help.
You can try open joined file for test.
I find an issue : transform this values in double float for load .
Regards
Dimitri
Hi,
I experience the same problem when using big FIR filters. To me it looks like SigmaStudio hangs after loading the coefficients when I try to close the editor window.
It's possible to set up to 10000 taps and by default all are set to zero. There is no problem to close the Table Editor window with default coefficients but after having loaded other values from a different filter design software it hangs. I'm using 32 bit float.
- Is there a limitation in the representation?
In the SigmaStudio Help file it says there is a limitation of 800 taps.
- Where is that coming from?
Hi,
Further testing has shown that there is possible to load bigger FIR filters. However it takes very long time to close the Table editor window. The time looks approx proportional to the number of FIR taps.
Example using SigmaStudio 3.12.3 BETA on a 32 bit computer with 4G RAM & SSD. 2000 taps takes 2 min 35 sec and 4000 taps takes 5 min 45 sec.
Testing on a 64 bit PC running Win 10 with SIgmaStudio 3.11 shows an error message (memory exception) after about 4 minutes.
What does SigmaStudio do that takes so long time when closing the Table Editor window?
Regards,
Mats
Hi,
I've seen that SigmaStudio converts float to fix numeric in some cases. I've also seen that a table in float format refuse to load.
I attach two images with error messages that I've captured and hope that someone can find the bug and correct it. It's really limiting my use of the 1452.
Both screens are taken from 3.12.3 BETA on a PC with 4GB RAM
Regards,
Mats
I made a custum a powershell script for transform numeric values in double float for load.
Load 10000 taps happens instantly !
initialy value
-0.00000000000000014706001404134733
-0.00000000000000029566666930119673
-0.00000000000000044586818932125847
-0.00000000000000059761847546011389
-0.00000000000000075101090422029943
-0.00000000000000090595607470270805
-0.0000000000000010625953208164742
-0.000000000000001220786223902057
-0.0000000000000013807235758526145
-0.0000000000000015422189668582696
......................................................................
transfomed value
-1.47060014041347E-16
-2.95666669301197E-16
-4.45868189321258E-16
-5.97618475460114E-16
-7.51010904220299E-16
-9.05956074702708E-16
-1.06259532081647E-15
-1.22078622390206E-15
-1.38072357585261E-15
-1.54221896685827E-15
I think we are a presence of a Sigma Studio bug, which is not adapted to the format 32 bit processor Adau1452
This issue is fixed in SigmaStudio 3.13.1Beta
This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.
Thank you,
EZ Admin
Please log in to reply.



