Speednote playback notes
Keep speed, pitch and time connected.
One ratio can describe three changes. The connection depends on how playback is performed.
Start with the coupled model
Imagine advancing through the same recorded material at a constant forward rate. If the rate doubles without pitch correction, each original second takes half an elapsed second, and every frequency is doubled. The model therefore links speed, frequency ratio and duration. Speednote evaluates that relationship from one chosen driver. It does not load a recording, inspect a waveform or apply an effect.
Many players and editors can preserve pitch while changing speed. That behavior breaks the pitch part of this coupled relationship. Check the actual playback or processing setting before treating an interval here as a prediction of what you will hear. The result panel repeats this assumption because a rate value by itself does not establish a player's pitch behavior. There is no pitch-preservation switch in this calculator.
Distinguish two speed percentages
Let r be the multiplier relative to the original speed. A value of 1 means unchanged speed;2 means twice the speed;0.5 means half. Percent of original speed is 100 r. Signed speed-change percent is 100(r−1). Those are different labels for related numbers. At 1.25×, the speed is 125% of the original and the change is+25%. Enter 25 in the speed-change driver to describe that increase, rather than 125.
Negative speed-change values can still mean forward playback. A−20% change gives r=1−0.20=0.8, which is slower but positive. A change of−100% would give zero rate and no finite completion time for this model, so it is rejected. Reverse playback and changing-rate automation are outside this tool.
Pitch is a ratio, expressed as an interval
An octave corresponds to a frequency ratio of 2. Equal temperament divides that octave into 12 equal ratio steps. A semitone therefore multiplies frequency by 2^(1/12), and 100 cents equal one semitone. Speednote reports an interval between original and coupled playback, without needing an original note name or a reference frequency.
total cents =1200 × log₂(r)
r =2^(semitones /12) =2^(total cents /1200)
A positive interval raises pitch in this model; a negative interval lowers it. Total cents means the complete interval in cents, not a cents remainder after removing whole semitones. For example,−50 cents is−0.5 semitones. It calls for a multiplier of about 0.971531941, and a 60-second source would take about 61.758134 seconds.
Work an exact octave check
Start with a 120-second source. At 1× there is no speed, interval or duration change. At 2× the speed is 200% of original, the speed change is+100%, the interval is+12 semitones or+1200 cents, and the duration is 60 seconds. At 0.5× the speed change is−50%, the interval is−12 semitones, and the same source takes 240 seconds. These ratios are useful checks of labels and direction.
A 102-second source, raised 3 semitones
Choose Pitch interval, enter 3 semitones, and use source 1:42.000. The multiplier is 2^(3/12), approximately 1.189207115. That is about 118.9207115% of original speed, or a+18.9207115% speed change.
The predicted elapsed duration is 102 divided by that multiplier, approximately 85.771434356 seconds. The clock display rounds this to 1:25.771. It is an arithmetic result for coupled playback; it is not a processed file or a guarantee of an editor's sample boundaries.
Duration follows the reciprocal
For fixed source duration T₀, the new duration is T₀/r. The signed elapsed-time change is T₀/r−T₀. The duration-change percent is 100(1/r−1). Applying the speed-change percent directly to the duration would generally produce the wrong answer.
A 150-second source at 0.8×
Enter 0.8 in the multiplier driver and 150 for the source. Speed changes by−20%, but the duration becomes 150/0.8=187.5 seconds, a 25% increase. The clock reads 3:07.500. The coupled pitch interval is about−3.863137139 semitones, or−386.313713865 cents.
Reversing a change also uses a reciprocal. To undo a 1.2× ratio mathematically, use 1/1.2≈0.833333333×. Its signed speed change is about−16.6666667%, not−20%. Multiplying 1.2 by 0.8 would give 0.96, rather than return to 1. This says nothing about reversibility of audio quality after a rendered effect.
Fit the source into an entered target
Fit mode asks for both source and target. Its multiplier is source divided by target; the pitch result comes from that same ratio. For a 45-second source and 40-second target, r=45/40=1.125. The speed change is+12.5%, the coupled interval is about+2.039100017 semitones, and the elapsed-time change is−5 seconds. A different tool or playback setting is needed if pitch should remain unchanged.
Switching from another valid driver to Fit carries the current predicted duration into the target. The underlying numerical model is retained while the visible target is rounded to milliseconds. Editing that target uses your entered duration instead. If the current result exceeds the target-input limit, the switch is refused; the prior representation remains available. A positive source must exist before carrying a model into Fit.
Enter durations deliberately
Use decimal seconds such as 102 or 102.125, two fields such as 1:42.125 for minutes and seconds, or three fields such as 1:02:03.125 for hours, minutes and seconds. In colon forms, seconds must be below 60 and a middle minutes field must be below 60. Two fields always mean minutes:seconds, so 60:00 means one hour. Up to 3 fractional-second digits are accepted. Signs, exponents, units and grouping separators are rejected.
A blank source in the first three drivers is allowed. Speed and interval still calculate, but duration, elapsed-time change and duration-change percent say unavailable. Supplied zero or negative durations are invalid. Fit mode needs both durations. Source and target inputs are capped at 604800 seconds, or 168 hours; the rate range is 1/64× through 64×. These are this implementation's software bounds, rather than limits attributed to an audio editor.
Keep precision separate from processing
Typed duration input is parsed in whole milliseconds. Intervals and derived rates use numerical logarithms and powers; repeating duration quotients are also approximate. Representation switches retain the unrounded model rather than calculating again from a rounded displayed number. Editing a converted driver adopts its new text. Clock formatting rounds the final duration to the nearest millisecond and carries into minutes or hours correctly. A very short positive result can consequently display 0:00.000; the accompanying seconds value and report retain the numerical result.
Reports are created locally from the current valid conversion. Invalid input clears the old result and disables exports. Reset restores the original 102-second,3-semitone example. Clipboard refusal offers a selectable report. No sample-rate conversion, frame quantization, time-stretch quality, waveform analysis, hardware calibration or actual processing is assessed.
Primary references
The official Audacity manual describes coupled speed and pitch behavior. The UNSW Physics interval reference explains equal-tempered frequency ratios. The W3C Web Audio 1.1 Working Draft documents rate and detune relationships; this site does not implement that API. The WHATWG media standard explains pitch-preserving playback. These sources support definitions; the examples, presentation and implementation here are original.