Does an MPC sound different from the sampler already in your DAW, or does it just feel louder, punchier and more fun to play? Casual comparisons rarely settle it. The louder pad wins, the hardware gets a workflow halo, and you come away convinced of something you never measured.
This exercise fixes that. You will trigger the same drum break and one-shots through a hardware MPC (if you own or can borrow one) and a software MPC-style sampler in your DAW. Then you will level-match the results and judge them blind.
A quick note on the rumor mill
A forum thread titled "Akai MPC 37 hardware?" started in July 2024 and carries only a one-word tag. It contains no answer, specification or confirmation, so treat "MPC 37" as unconfirmed and don't plan a purchase around it.
The same thread lists a related post titled "Hardware Akai MPCs can now run VST plugins". Only the headline appears, so verify the details elsewhere before relying on it. If hardware can host plugins, the line between "hardware sound" and "software sound" blurs, which makes a controlled test more useful.
The test below works with any MPC-style unit you can get hold of. It also works with two software samplers if you only want to compare those.
Step 1: Gather your material
Pick one drum break and three to five one-shots: a kick, a snare, a closed hi-hat, and something with low-end weight such as an 808 or a sub hit. Use the same source files in both samplers. Use uncompressed WAV, and avoid files that have already been through a limiter.
If the hardware needs files on a card or drive, copy the exact same WAVs. Do not resample or convert them along the way.
Step 2: Set up the project
Work through this checklist before you record anything.
- Set the DAW project to one sample rate and keep it there. Check your hardware's manual for its native rate and match it, so neither path forces a conversion. Common choices are 44.1 kHz and 48 kHz.
- Record at 24-bit. Do not dither, and do not let the DAW convert anything on the way in.
- Bypass or remove every master-bus plugin: no limiter, no EQ, no saturation, no "glue" compressor.
- Set the same tempo for both sources and time-stretch nothing. If the break has a native tempo, use it.
- Program one MIDI clip in the DAW with fixed velocities and use it to trigger both samplers. The software sampler receives it internally. The hardware receives it over MIDI out from your interface or over the MPC's USB connection. This removes finger-drumming variation from the test.
- Turn off velocity-to-volume tricks, swing, humanize, and any built-in effects or "vintage" modes on both sides for the first pass.
- Set both samplers' pad or channel levels to unity. Aim for peaks around -6 dBFS on the recorded tracks, with no clipping anywhere.
Match the signal path as closely as you can
The hardware's output passes through your interface's converters before it reaches the DAW. The software's output does not. To even this out, you can route the software sampler out of one interface output and back into an input, so both signals go through the same conversion. Then you are comparing samplers, not converters.
Record each source to its own audio track. Zoom in and align the first transient of each file by waveform, because round-trip latency will offset them. Even a few milliseconds of misalignment can change how snappy a hit feels.
Step 3: Level-match both signals
Louder sounds better, almost every time. Remove that variable before you listen. For more on this, see related: srv-330 reverb emulation vs stock reverb: a/b test guide.
Bounce the same section from each track, with identical start and end points. Run each bounce through a meter that reports integrated LUFS and note the numbers. If your DAW has no meter, a free metering plugin will do. Use RMS if LUFS is not available.
Trim the louder bounce with clip gain or a plain gain stage until the readings agree within about 0.5 dB. Do not use a limiter or compressor to make them match, because that changes the very thing you are testing. Match the loop, then compare hit by hit.
Two cautions apply:
- Short one-shots: A very short one-shot may give an unreliable LUFS reading. Match using the full break or a loop of repeated hits.
- Sharp transients: A sound with sharper transients can read quieter than it feels. Also check peak levels, and note any large difference on your scorecard.
Keep one unmatched pass for the noise floor
Trimming gain also moves the noise floor, so test noise on a separate, unmatched recording. Record several seconds of silence from each source with the same output level and the same interface gain. Then compare the two silent clips at a safe monitoring level.
Step 4: Run the blind A/B
Create two tracks called A and B. Ask a friend to assign the hardware and software bounces to them at random, or use a DAW feature or plugin that hides which is which. If nobody can help, rename the files with random codes and write the key on paper without looking at it.
Follow this process:
- Listen to a short section of A, then B, then A again. Keep each section under 10 seconds so you can remember what you heard.
- Write your choice on the scorecard before you check the answer.
- Randomize the order and repeat for at least 10 trials, using different hits and different sections of the break.
- Use the same headphones or monitors, and keep the volume knob in the same position throughout.
After 10 trials, count how many times you correctly identified the hardware. Pure guessing gets about half right. Getting 8 or more right happens by chance only about 5% of the time, so that result suggests you can hear a real difference. A score near 5 means any difference is smaller than your ears or monitors can reveal.
Step 5: Listen for specific qualities
Impressions like "warmer" or "fatter" are hard to act on. Judge four things separately. See a closer look at are universal audio plugins worth it? a level-matched test for additional background.
Transient snap
Focus on the first few milliseconds of the snare and hi-hat. Does one source sound sharper or rounder at the attack? Check the waveform too, since a softened peak on screen supports what you hear.
Low-end weight
Play the kick and the sub hit. Listen for differences in length, tightness and perceived sub energy. A spectrum analyzer can confirm what you think you heard, but your ears get the final say.
Noise floor
Use the unmatched silent recordings. Note any hiss, hum, fan noise or digital whine, and whether it changes when the sampler is idle versus playing.
Filter or saturation character
Turn on the sampler's own filter, bit-reduction or vintage mode on both sides and set them as closely as you can. Compare how each behaves as you push it. Re-match levels after each change, since saturation often raises loudness.
Step 6: Fill in the scorecard
Copy this table and complete one row per quality after you reveal the answers.
| Quality | Heard a difference? (Y/N) | Preferred source | Confidence (1-5) | Notes | |---|---|---|---|---| | Transient snap | | | | | | Low-end weight | | | | | | Noise floor | | | | | | Filter/saturation | | | | | | Overall blind score (out of 10) | | | | |
Also read: see how to compare granular synths in your daw: 6-step test
Step 7: Separate sound from workflow
You now have a sound result. Workflow is a separate question, and mixing the two is how people overpay. Ask yourself two things.
Would you still prefer the hardware if it sounded identical? If yes, you are buying a way of working. That is a valid reason, but a different one. Would you still prefer it if it were harder to use? If yes, the sound difference probably matters to you.
Buy-or-skip guide
- Buy for sound if you scored clearly above chance, you prefer the hardware in at least one category you care about, and you would use that character on real tracks.
- Buy for workflow if the blind test showed no difference but you finish more beats away from the computer.
- Skip if your score sat near chance, you can't name a quality that changed your decision, and the main appeal is the "hardware sound" reputation. A free or low-cost software sampler in your DAW will serve you well.
If the result is close, repeat the test with your own samples and your usual processing. Borrowing a unit from a friend or trying one in a shop costs far less than discovering the difference was imaginary after you buy.
What to expect next
Your first run will likely expose problems in your own setup, such as a level you forgot to match or a latency offset you missed. Fix them and run another pass before you draw conclusions. If new MPC hardware is announced, the same exercise applies, and you can judge it by ear instead of by forum rumor.
