This is a step by step instruction on how to get started with jDFT and also some tips for getting the most out of it. Unfortunately these screenshots and instructions are not updated for the latest version of jDFT. However everything that's mentioned here still apply to the latest version of jDFT, it's just some new features that are not mentioned.

Before we start, just a note about the dB-grid that's added in v 2.2. dB is a relative measure, it's just a way to compare two signals. If you have a volt meter with a dBu mode then the voltage (U) is compared to a reference voltage (Uref=0.7746 V), as 20*log10(U/Uref). In jDFT it's the same but Uref is unknown and will be different on different computers and depend on the Volume control, the scaling, the offset, the window function and number of samples used (maybe I missed some parameter). You can however compare two signals if you keep all the above settings fixed.

This is the opening view that you will encounter when you start jDFT 2. Simply follow the onscreen instructions!
 

The frequency spectrum of some audio signal, guess it's some music being played.
At the bottom of the window there are some options that you should take a look at...

• First of all is the frequency span, just select the desired frequency span to analyze (Note! A new simpler zoom feature is added to v2.3, just zoom with the mouse), note however that the max frequency is limited to half the sampling rate. This is a theoretical limit of the sampling process itself and is almost as fundamental as the gravity.
  
• The Inertia could maybe have been called something else, but it's simply a number that's used to decide how fast the reaction should be. If the inertia is set to X, the output on the screen is the average of the X last FFTs.
  
• Pks, meaning Peaks is a way to let peaks pass through the system. So even though you have a high inertia all peaks will be shown, and then beacuse of the high inertia they will slowly 'decay'.
  
• Length, don't bother about it. If you lower the value you will do FFT on a shorter sequence than the number of samples that you choosed in the opening screen, the remaining samples will be zero-filled. You could use it if you have a low sample rate and a high number of samples, because then you could be analyzing almost 1 s of audio and if you're looking at music you can for example not distinguish between the snare drum and the kick drum. (What you should do in this case is to lower the length to, say 150 ms.)
  
• Scale is used to scale the Y-axis.
• Offset is used in a similar way to set the reference level in some sense. Useful if you're doing stereo fft:ing and want to be able to distinguish between the left and right channel.
  
• Stereo?, self-explained.
• Window function, the input signal can be filtered through a time window to change the characteristics of main lobes and side lobes of the signal's frequency spectrum. A rectangular window gives a more narrow main lobe but you get more side lobes. Hanning and Hamming on the other hand will give you a wider main lobe but the side lobes will have a lower amplitude. This can be verified by analyzing a pure sine wave.
  
• Min X-space is used to set the X-spacing, in other words you set the minimum width of the bars. Unfortunately you must have at least 3 to see anything at all at the higher frequencies, this is sometimes called an unwanted feature. This unwanted feature is fixed in v 2.2!!

On this screen Min X-space has been set to 11.
 

This is stereo mode, as you can see there is a small difference in the offset for the left and right channel.
 

Wow, Always on top (this hides the settings as well). You enable always on top by double clicking with the right button.
 

I guess this is a 1kHz tone, as you can see the Window is set to Rectangular and therefore there is quite much side lobe effects.
 

Finally we have some pink noise!
 

Just a small note, this application is by no means perfect but I think that you can learn quite much anyway by playing around with different sounds.