Calculation of Signal : Noise Ratio

[mike]

Im an absolute beginner in GC Chromotography and need to know..

When calculating the signal to noise ratio, are you comparing the HEIGHT (Intensity) of the respective peak with the average HEIGHT (Intensity) of the noise in the chromatogram. or are you comparing the AREA of the respective peak with the AREA of the noise?

If its comparing areas, how do you measure the AREA of noise, as we have some strange situation here with root mean squares and peak areas!

Please explain a little as Im fairly clueless about this!

Mike

[tom jupille]

It's based on height. The height of the peak above the average baseline divided by the peak-to-peak noise of the baseline just before or after the peak.

[mike]

Well I will be arguing the case for peak heights soon, but still someone has said to me recently that they compare peak areas!

They measure the peak area manually and make measurments of random 'blips' on the noise peak to generate 'noise area'. They then make an average of the 'noise area' and divide the signal area and noise area to generate a signal:noise ratio.

This method seems really inaccurate. They say it is the best way to keep all the units the same. i e all data is measured as unit area.

If you think it is still ok to compare heights (which seems to be easier) then tell me and I will arue this with my masters!

Mike

[MG]

Mike, I've been doing LC/MS for 7 years and have never heard of using areas for S/N as you describe. You could challenge your boss to show you a journal article where it is done this way. If your boss shows you one, please post the reference here.

[HW Mueller]

Maybe I should go home and start enjoying Easter, possibly being to tired for this, but: It seems to me that to compare something as weird as a baseline area to a peak would require that the frequency of the baseline fluctuations would have to be flow dependent. The hight of the fluctuations might be flow dependent, but the frequency???