What Happened ? (examining and correcting past processing)

It can easily happen that you try to open a FID with iNMR and you see a transformed spectrum instead. The program has attempted to anticipate your will and has already transformed the spectrum. In most cases the processing is so fast that you don't realize that many complex calculations have been performed. If you don't like this sort of automatic behaviour, you can turn on the option “Show the FID when Opening a Document” (the last option into the Preferences dialog). This is not practical, however, because you will be forced to manually repeat your processing every time. Independently from the setting of your preferences, you can always use the command File > Reload. In this case you will see the spectrum as it is stored on disc.

Before using Reload, you have the chance of examining the processing that has been performed. Use the command Edit > Copy > Processing, then the command Paste. After reloading, this operation may not work, because the spectrum can be in time domain. The processing parameters are still there, but you have to search them one by one. The command File > Revert to Saved can undo the effects of the command Reload (because it undoes anything the user does).

iNMR has been built around a very simple model to manage the processing work-flow. If you know this model you will better understand what happens under the graphic interface. For each dialog there is a set of parameters. For example, there is a set of parameters for the baseline correction and another set for binning. When you open a dialog or click an icon (for example: the icon for automatic baseline correction), iNMR looks for the corresponding set. Many times the set is not found (all the documents are created without processing parameters). In such cases, a new set is created with default values. Processing parameters are always stored if they are used, and many times they are stored even if they are not used. There is no space for duplicated sets. You can apply 3 different baseline corrections, cumulatively, but only the last set of parameters is stored. If you move from an extract to the whole matrix, or vice versa, you will find the same parameters. There is no mixing, however, among different dimensions: if a spectrum has N dimension, its document will contain N times more parameters. Phase correction is managed in a slightly different way. New documents are created with their phase values already in place (initially set to zero). After performing a Fourier Transform, iNMR always applies a phase correction, using the current parameters.

Whenever you apply an operation on your spectrum, iNMR updates its processing status. iNMR does not remember the order of your operations nor how many times thay have been applied, it only remembers which operations have been performed and which not. When you close the document, this processing status is registered into the .inmr file. The next time you open the same document, unless you have already chosen the option “Show the FID when Opening a Document”, iNMR checks the processing status and applies the same operations. In which order? The order of operation is built-in and immutable:

Never are all these operations necessary. The actual lists of performed operations are much shorter. The first block (up to phase correction) is repeated for each dimension of the spectrum. It is highly advisable that you follow the same order of operations in your processing, otherwise the reprocessed spectrum could be slighlty different. Actually, it comes naturally to always follow this order, because it is the most sensible and natural order. Your freedom is already restricted: the main block, from Drift Correction to FT, can't be reordered. The first operations can only be performed in time domain, the last operations can only be performed on real data in frequency domain and phase correction requires complex data. In the unlikely case you need to follow a different order (or to repeat an operation), the solution is to write a script with the console and store it into the document. The script will take the precedence over standard processing.

A few algorithms have not been inserted into the above list. They are Reference Deconvolution and Symmetrization of INADEQUATE experiments. If they become routine operations they will certainly be added to the workflow.

If you need to change a single parameter (or the whole processing), you have to reload the FID with the command File > Reload and start over as the spectrum has never been processed before. The only difference is that this time you will not find the default parameters but your old settings. The command Reload is a good friend because you know that you are really starting anew, with the original data that came from the spectrometer; at the same time your processing settings are preserved, in case you need them. If you spend a lot of time to optimize your processing, you are going to invoke this command with frequency. When you learn to drive a bycicle or a car, you start from: “Where are the brakes?”. When you learn iNMR, the best starting point is the command Reload; it is the equivalent of the brake. It is better than the command Undo.


 
Copyright © 2005-2017 nucleomatica
Valid XHTML and CSS. UTF-8 encoding.