How Tab Sketcher works
One record behind everything you see
Every note you place is kept as two plain numbers: which string, and which fret. The tab, the standard staff, the playback and the exported file are all produced from that one record, so they cannot drift apart from one another. If the staff shows a G, the player is sounding that same G.
None of the steps in between are guesswork. Each one is ordinary arithmetic you can redo on paper with a guitar in your lap. This page walks through the three that matter most: how a fret becomes a pitch, what transposing calculates, and where the sound comes from.
How a fret number becomes a real note
Each open string has a fixed pitch, and moving up one fret raises that pitch by one semitone, the smallest step in Western music. So the pitch of any note is simply the open-string pitch plus the fret number. Fret 0 means the open string, sounded without pressing anything down.
In standard tuning the six strings sound E4 B3 G3 D3 A2 E2, counting from the thin first string down to the thick sixth. Strings are numbered the way a beginner is taught: string 1 is the thinnest and highest. The note name the virtual fretboard reports for a position comes from that same addition, not from a stored picture of a fretboard.
Picking a tuning preset (Drop D, half step down, whole step down, Open G, Open D or DADGAD) only changes the starting pitch of the strings; every fret is then recalculated from the new starting points. A capo setting is stored with the piece alongside the tuning and handed to the notation engine, so it travels with the file instead of living only in your head.
One deliberate simplification: note names are always written with sharps, so a piece sitting on D flat is labeled C sharp. The pitch is identical, but a teacher may well write it the other way around.
What transposing actually calculates
Transposing means moving every note by the same distance, which is why the piece keeps its shape instead of turning into a different tune. Shift by two semitones and the tool adds two to the pitch of every single note. Then it has to answer a second question: where on the neck does that new pitch live?
It tries the easy options first. The same string as before, if the new note lands within the first twelve frets. Failing that, whichever other string can reach the note at the lowest fret. Two notes of one chord are never sent to the same string, because a string can only sound one note at a time. Only when nothing fits inside the twenty-four frets of the neck does the tool move that beat by a full octave, which keeps the note names and changes how high the passage sits.
The key selector shows the key you land in and always takes the shorter way around the twelve keys, so it never proposes eleven steps up where one step down does the same job.
Two things are worth knowing. Transposing changes the fingering, so a shape that was comfortable can become awkward, and a capo is often the friendlier answer for a singer. And when an octave jump is needed, the notes are still right but the passage sits in a different register than the original.
Chord shapes and chord names
The chord library stores each shape as the same string-and-fret pairs used for ordinary notes. Strings that are not part of a shape are marked with an x, meaning they are left unplayed. When you drop a chord from the library into a beat, exactly those pairs are written into the piece, so the diagram, the notation and the sound are three readings of one stored shape.
When a chord name is filled in for you, the tool gathers the notes in the bar, reduces them to note names regardless of octave, and compares the result against a list of interval patterns. Four-note patterns (major seventh, dominant seventh, minor seventh, diminished seventh) are checked before three-note ones, so a seventh chord is not mislabeled as a plain triad. If nothing matches exactly, nothing is printed. A wrong name confuses a beginner more than no name at all.
The stored shapes are covered by an automated test for the common chords: it turns each stored fingering into actual pitches and confirms that the naming step hands back the name it started from. Only the familiar chord families are recognized, and one set of notes can honestly carry more than one name, so any name can be typed over by hand.
Where the playback sound comes from
The finished notation is handed to AlphaTab, an open-source notation engine, which both draws the score and plays it back through a built-in synthesizer. The pitches come from the numbers described above, and the timing comes from the tempo and the time signature.
The instrument tone comes from a freely distributed SONiVOX SoundFont, a small bank of recorded instrument samples that the synthesizer retunes and shapes to reach each note. What you hear is synthesized, not a recording of a guitarist. It is dependable for checking pitch, rhythm and harmony, and it is not meant to stand in for real guitar tone.
Some playing techniques are approximations by nature. A bend is written with one standard shape rather than the exact curve your finger would make, and a pull-off carries the same marking as a hammer-on because the notation format uses a single symbol for both directions. Treat playback as a way to proofread the notes, not as a recording of a performance.
Beats, bars and note lengths
A bar holds a sequence of beats, and each beat holds the notes that sound together plus one length, anywhere from a whole note down to a thirty-second. Time signatures come as presets (4/4, 3/4, 6/8, 2/4 and 12/8), and a change of signature is recorded on the bar where it happens, so every bar after it follows the new count. Tempo is stored in beats per minute and drives playback directly.
Dotted notes and triplets are not part of the model yet. Lengths snap to whole, half, quarter, eighth, sixteenth and thirty-second, and an imported file that uses dots or triplets has those lengths rounded to the nearest plain value.
What survives an import
A Guitar Pro file often carries the same music twice, once as a staff and once as a tab, and only the tab part holds fret numbers and a tuning. The importer looks through the file and picks the part that has both, which is why an imported piece arrives with playable fret numbers instead of empty bars.
The import keeps the first track and the first voice of the file. If the file holds more than one track, or more than one part, a notice tells you which one was used rather than leaving you to discover it later. And if the conversion cannot produce anything playable, the original notation stays on screen in a view-only form instead of being replaced by a blank page, so opening a file never quietly erases it.
Sources, licenses and the sample pieces bundled with the tool are listed on the about page.