January 10, 2012
by Arvin Moser, Team Manager, Application Scientists, ACD/Labs
For each increment of zero filling, the number of data points defining a line shape is increased. The trade-off is the size of the FID increases proportionately. An FID with 1K = 1024 points will produce a file of ~ 50 kb (including data parameters). The goal is to balance the file size with the amount of gain from applying zero filling.
Spectra A and B appeal aesthetically due to the higher level of smoothness across the signals. Whereas the original data, spectrum E, is jagged and presents less than ideal line shapes.
Analysts typically apply 1 to 2 zero fillings; this equates to adding respectively 1 to 2 increments of zeroes equal to the number of data points.
So the only downside to zero filling is hard disk space? So then A would be the best spectra, ignoring disk space requirements (which become cheaper and cheaper each year)?
Hello Canageek,
Thank you for taking the time to contribute.
I think there are several disadvantages to overestimating the amount of zero filling for an NMR spectrum. Here are a few issues to consider:
Yes, the hard drive space can fill up quickly with large files built up by excessive zero filling. This becomes quite evident with 2D and 3D NMR datasets.
Large files may also impact data storage systems that house the data. Library search algorithms may be affected by bulky files.
Software used to process and report the spectral data can be bogged down by larger files.
Please note that I moved your comment from Logic Puzzle #26 to #25. I find it more appropriate with the latter post.
Regards,
Arvin