Talk:Spectrogram View


 * Bill 11Nov2014: Here is a start on detailed documentation of spectrogram view. I intend to continue with this, showing the log(f) view, then concluding with some sort of comparison that shows that different settings (window size, linear vs. log(f), min and max frequency) are better for different source material. I also need to demonstrate when the Frequency Gain setting can be useful.
 * Bill 22Dec2014: Added Spectrogram log(f) with examples. Still need example of frequency gain.
 * Bill 22Dec2014: Needs review.
 * Steve 23Jan15: Looks good, but as part of the P2 I think it needs examples of spectral selection (assuming that spectral selection is included in 2.1)
 * Peter 04Feb15: I really don't agree. We already have a page called Spectral Selection which documents that. Furthermore the Waveforn view page doesn't carry an image with a selction.  We could consider a link to Spectral Selection. Removed the P2.
 * Steve 04Feb15: As
 * Spectral Selection is on by default,
 * Even when you "switch it off" with the Q button, it comes back on as soon as you make a new selection (so there really is no way to "turn it off"),
 * The user is more likely to have a selection than not have a selection,
 * We have a section here called What the Colors Mean,
 * I think users will, in the absence of any explanation, find it confusing that the will see colors that different to what are shown here. All that I think we need is one image of a Spectral Selection, saying: "This is how the track appears in spectrogram view when there is a selection. (link to Spectral Selection").


 * Gale 05Feb15: I have not seen this page before now, but I'm going to go further and call it P1 for an image showing spectral selection (or possibly two, one with selection with no bandwidth, and one for a selection with defined bandwidth). Even making a selection with no bandwidth draws a horizontal line in the selection. What is it assumed to be, without explanation? Is it another kind of zoom band? And as Steve says, if you drag down, what is the color overlay?
 * Peter 05Feb15: I have cribbed material from Spectral Selection for image and text and added a link. Please review.
 * Peter 05Feb15: Steve's point 2 above "Even when you "switch it off" with the Q button, it comes back on as soon as you make a new selection (so there really is no way to "turn it off")" is really what makes it so potentially dangerous imo. I really believe that there should be a mechanism for turning Spectral Selection on/off with off being the default setting, possibly an item in the Tracks menu or maybe the Edit menu.  The "Q" toggle is insufficient for the reasons that Steve states. But we can't really consider doing this before 2.1.0 sadly.
 * Peter 07Feb15: Looking at this page I note that we don't really show the user what a normal recording looks like in Spectrogram view - rather we dive off into "what the colors mean", "smearing" etc. We do show them a Spectrogram log(f) looks like but only way down the page.
 * Bill 07Feb2015: Added a "note" to the intro box about spectral selection being active. Added "intro" images for waveform versus spectrogram view of the same audio.
 * PRL 10Jun2015: I did not realize this page existed when I began writing this other page: http://manual.audacityteam.org/man/Tutorial_-_Understanding_spectrograms#Window_size For 2.1.1 I have added Zero Padding to spectrogram preferences, and I hope that the Reassignment algorithm will become a reality in version 2.1.2. Illustrations are needed to demonstrate what zero padding does and why it is useful.  I also wanted to add my own explanations of other things like the color scale.  I have outlined my explanations and given recipes for generating some theoretically illustrative sounds, though they are not real-world.  Are my examples reproducible easily enough?  Should the two pages be combined?  Does anyone want or insist on the job of filling out my outline with more prose and pictures?
 * PRL 13Jun2015: I like Bill's updates. But I hesitate over the term "frequency resolution."  I am not sure that is the right term for what we have in the Blackman-Harris window.  Perhaps we should call it "clearly separated overtones."  It is actually rectangular that has the best "resolution" though also the worst leakage, and B-H that has slightly inferior resolution to Hann.  Have you seen this graph: https://en.wikipedia.org/wiki/Window_function#Comparison_of_windows ?  Look closely at the black, red, and yellow curves for rectangular, Hann, and Blackman (Blackman-Harris is not included).  Notice that they differ in the location of the first zero but also in decay of the side lobes.  That corresponds to the narrowest bands for the overtones in the rectangular window picture, and widest for Blackman-Harris, yet side lobe decay is another matter accounting for the cleanness of the picture between the frequency peaks.
 * Bill 13Jun2015: Paul, please edit the relevant sections so they are correct.
 * PRL 14Jun2015: I am glad to see the chromatic scale example. The example for zero-padding (or whatever we end up naming it) does not seem compelling.  I admit we are still in search of a very compelling example.  But, did you see the examples in my email of vocal overtones?  The valleys between peaks of the spectrum may be better resolved with zero padding, while not increasing the temporal smearing.  I also replicated your example of a pluck with default settings, then increased zero padding with no other change.  That too made some of the overtones a bit more distinct.
 * PRL 14Jun2015: I suggest something like: "Increased zero padding may improve the display of marginally resolved overtones without increasing the time smearing." With an appropriate example, but not zoomed in to a single overtone as here.
 * Bill 14Jun2015: "marginally resolved overtones" is too technical IMO. I am still in search of a "compelling" example. Please suggest a way to generate one, or point me to an audio file I can import and play with.
 * PRL 15Jun2015:
 * Say "nearly" or "almost" instead of marginally. Have you simply tried the Pluck with default settings, then view with window size 512 and the rest defaults, then increase smoothing to 2 or 4?  You could put a click in it too to demonstrate that the time smearing does not increase.
 * I like the improved color example, but now the "what the colors mean" section is discussing more than color. -80 is what I called "pale blue."  There are really only five colors, not six, that define the rest of the palette by blending.  In future you will be able to adjust those five in custom palettes.
 * There is no example for Frequency gain yet. Frequency gain is implemented so that it is neutral at 1000 Hz, brightens what is above that and dims what is below.