Back to the main page.
Bug 464 - ft_freqbaseline looks outdated and inefficiently implemented
| Status | CLOSED FIXED |
| Reported | 2011-02-01 10:04:00 +0100 |
| Modified | 2011-03-23 10:09:06 +0100 |
| Product: | FieldTrip |
| Component: | core |
| Version: | unspecified |
| Hardware: | PC |
| Operating System: | Mac OS |
| Importance: | P1 normal |
| Assigned to: | Eelke Spaak |
| URL: | |
| Tags: | |
| Depends on: | |
| Blocks: | |
| See also: |
Jan-Mathijs Schoffelen - 2011-02-01 10:04:07 +0100
make new improved implementation of this function
Jan-Mathijs Schoffelen - 2011-02-25 10:41:48 +0100
Hi Eelke, Let's do this one together. Could you make a plan which we can discuss? I'll supervise ;o)
Eelke Spaak - 2011-03-01 12:19:28 +0100
Some simple suggestions for improvement: - Combine the three computing subfunctions into one - Remove the loops over frequencies and channels, replace with vector-based code I am not entirely sure about the relation between the dimord and the possibility to normalise a coherence spectrum. If the dimord is 'rpt_chan_freq_time', the function refuses to normalise a coherence spectrum. This is because coherence within a single trial does not make sense? I guess we should discuss this.