A leaked Teenage Engineering CD player has a VARISPEED control, and it raises a fair question for anyone who samples: is that just time-stretching under another name? It isn't. The two methods change a sample in different ways, and you can hear the difference with a track you already own.
About the leak (unconfirmed)
The gearnews report describes a leaked Teenage Engineering EP BC-1 CD player with an open-air design, a VARISPEED control (what older DJs call pitch control), cue points and USB connectivity. It suggests the player could feed an EP-133 K.O. II sampling workflow. Teenage Engineering has not confirmed any of this. No specs, price or ship date are official, and the report's "Hopefully in 2027!" is a hope, not an announcement.
The leak is only the reason to ask the question. You don't need the hardware to answer it, because your DAW can do both processes today.
Varispeed vs time-stretch: the difference in plain terms
Varispeed changes playback speed, and pitch moves with it. It works like a turntable, a tape machine or a CD player's pitch control. Speed a sample up and everything gets higher and faster together, including the drum tones and the vocal character.
Time-stretching changes duration without changing pitch. The algorithm slices the audio into small pieces, rearranges them and smooths the joins. That's why it can leave artifacts, and why quality depends on the mode you pick.
The tempo math is the same either way. Moving 90 BPM to 100 BPM is a ratio of 100/90, or about 1.111, so the audio runs roughly 11.1% faster. With varispeed, that ratio also raises pitch by about 1.82 semitones (roughly 182 cents). With time-stretch, pitch stays put.
Repitching a sample in a sampler is usually varispeed-style, since speed and pitch move together. Check your own sampler's manual to confirm how it handles this. See related: how to compare an analog synth plugin to hardware fairly for additional background.
What you need
Pick one track you own. A loop with clear drums, a bass line and some melodic or vocal content works best, because it shows each method's weak spots. Choose a section that starts on a downbeat and runs a whole number of bars, such as four bars at 90 BPM.
You also need a loudness meter, either your DAW's own or a free plugin that shows LUFS or RMS. A friend or a shuffled file list helps with the blind test.
The exercise, step by step
-
Set up the reference (render A). Import the loop into a new project set to its original tempo, 90 BPM in this example. Turn off any warping or stretching. Check the length: four bars at 90 BPM should last about 10.67 seconds. Render it as a 24-bit WAV with peaks below -1 dB. This is your unaltered anchor.
-
Render the varispeed version (render B). Set the project to 100 BPM. Disable warp/stretch on the clip, then speed the audio up by the ratio 1.111 using the playback rate or resample function. In a sampler-style device, a transpose of +1 semitone and +82 cents gets you close. The same four bars should now last about 9.6 seconds. If they don't, your ratio is off. Render it.
-
Render the time-stretch version (render C). Go back to the original audio and set the project to 100 BPM. Enable warp/stretch so the loop locks to the grid with its pitch unchanged. For full mixes, start with a high-quality mode such as Complex or Elastique Pro (names vary by DAW). Leave formant and pitch settings at their defaults. Render it and confirm it's also about 9.6 seconds.
-
Optionally, repeat with a second setting. Render a C2 with a different stretch mode, for example a rhythmic or transient-focused one. This shows how much of what you hear comes from the method and how much from the algorithm.
-
Level-match the renders. Loop the same section of each file and read its loudness on your meter. Adjust each file's gain until all of them sit within 0.5 dB of one another on integrated or short-term LUFS. With no meter, match by ear and stay inside about 0.5 dB. Louder almost always sounds better, so skip this step and the comparison means nothing. Varispeed shifts the frequency balance, so expect to trim it a little.
-
Randomize the labels. Rename the three files with neutral labels (1, 2, 3), or ask a friend to do it so you don't know the order. Write the key down somewhere you won't see it. Then listen in random order, at least twice, before you peek.
-
Score each file with the checklist below. Listen on speakers and, if you can, on headphones. Spend a few passes on each element: drums first, then bass, then any vocal or lead.
Listening checklist
The table describes what to listen for and where each method tends to cause trouble. It isn't a set of test results. Your ears and your source material decide. We cover related ground in polyphonic vs channel aftertouch: test it in your daw in depth.
Also read: related topic: create custom nks presets and verify recall in your daw
| Criterion | Reference (A) | Varispeed (B) | Time-stretch (C) | |---|---|---|---| | Transient smearing | Your baseline for how sharp the kicks and snares are | Attacks keep their shape but arrive faster. Listen for any change in snap | Check whether drum hits blur, double or lose their click, especially snares and hats | | Formant / pitch shift | Original key and voice character | Everything rises about 1.82 semitones. Check whether voices sound smaller or thinner | Key and voice character should match A. Listen for unnatural vocal tone | | Graininess | Clean source for comparison | Little processing artifact. Any grit comes from the source | Listen for warbling, metallic ringing or fluttering on sustained notes and cymbal tails | | Low-end change | Reference kick and bass weight | Bass notes move up in pitch and decay faster. Check whether the low end feels lighter | Check whether the bass wobbles, thins or loses punch. Stretched bass can feel unstable | | Overall vibe | Original groove | Does it feel like a record sped up, with a lively, tape-like energy? | Does it feel natural and tidy, or processed and slightly artificial? |
Which method fits which sample
Once you've scored the files, use these guidelines as a starting point. They're inferences from how the processes work, so trust your ears over them.
Drums. Varispeed is the safer bet for single hits and breaks, because it doesn't slice the transients. If the pitch shift bothers you on a kick or snare, try a transient-focused stretch mode and compare the snap against render A.
Vocals. Even a small varispeed change can alter a voice's character, so check B for a thinner or smaller sound. Time-stretch keeps the voice in place but can add warble, especially on long vowels. If the key shift sounds musical to you, varispeed is fine. It's a classic sampling move.
Full-mix loops. These expose the most. Stretching a dense mix can make the kick, bass and vocal behave differently from each other. Varispeed keeps the mix coherent but changes its key, which matters if you want to play over it. If the key shift clashes with your other parts, stretch. If it doesn't, varispeed may feel more natural.
Small tempo moves are the most forgiving for both methods. A big jump, say 90 to 120 BPM, will stress either approach, so repeat the exercise at your real target tempo.
What to do next
After the blind listen, check your notes against the key. You may prefer different methods for different parts of the same track, so try the exercise once with drums alone and once with a vocal. If the BC-1 gets confirmed and details emerge, render B will give you a rough idea of what a pitch fader would do to your CD samples. Until then, the render files are your reference.
