The question everyone asks second
Is it my heartbeat or the baby
Almost always, the first thing you find is you. Not because you did it wrong, but because your circulation is the loudest thing between the phone and the baby.
The second recording in the tool is a mother's own heart. Watch the engine name it instead of doubling it.
Firstsong is not a medical device, not a doppler and it does not use ultrasound. It cannot tell you that your baby is well. If you notice fewer movements than usual, call your midwife or maternity unit now.
The short answer
If what you are hearing is slow, heavy, easy to find and lands like a boot in a dryer, it is almost certainly yours. A baby's heart is faster, lighter and much quieter, and most people say it sounds like it is happening in the next room. The way to settle it is not to listen harder, it is to put two fingers on your own wrist and compare.
Your pulse is the loudest thing in your belly
The sound you are hunting for starts in a heart the size of a walnut, then crosses amniotic fluid, the uterine wall, your abdominal wall and your own tissue before it reaches a piece of glass on your skin. Most of its energy is gone by then.
Your own circulation has no such journey. The large vessels in your abdomen sit inches from that same glass, and a serious share of your cardiac output moves through the placenta, which adds a low whoosh of its own. It is not a close contest. That would not matter if the two rhythms were far apart. They are not, and that is the trap.
Doubling is what makes the wrong answer convincing
A heartbeat is not one sound, it is two: the lub and the dub. Anything hunting for repeating structure can lock onto the gap between the two sounds of one beat instead of the gap between beats. The rate it then reports is exactly twice the real one, and it reports it confidently, because the pattern really is repeating.
Now do the arithmetic on your own resting pulse. Anywhere between 55 and 90, doubled, lands between 110 and 180, which is the whole band a baby's heart lives in. That is not bad luck, it is a structural trap sitting in the middle of the problem, and it is why the same two sentences dominate the reviews of this category: it did not work, and I think that was my own heartbeat. In a corpus of 1,995 reviews, 171 describe the second one.
| If your resting pulse is | Doubled it becomes | Where that lands |
|---|---|---|
| 62 | 124 | comfortably inside the fetal band |
| 71 | 142 | the exact number these apps love to show |
| 84 | 168 | still inside the band, still wrong |
The three defences the engine has against you
This failure was the thing to beat before any screen was designed. Three defences at three stages, because one check is one point of failure.
It knows the two bands apart
A rhythm sitting in the adult range is not quietly dropped and not quietly doubled. It is reported, in words, as an adult rhythm. Naming it out loud is what separates an honest app from a silent one.
It refuses exact integer multiples
Two real hearts are never locked to each other. A baby at 140 and her mother at 72 sit at a ratio of 1.944, not 2. So a candidate that is an exact whole number multiple of the loud rhythm is treated as an artefact of the maths, not a second heart.
In the app, it checks your real pulse
The one the browser demo cannot do. Your iPhone reads your real pulse from your Apple Watch, or the camera, and refuses anything matching it or its harmonics. A second opinion, from somewhere else.
Watch it happen on clip two
The second recording in the tool is a mother's own heart, from the same public database as the baby in clip one. Watch the engine reach the maternal verdict instead of doubling its way into a fetal one.
Runs in your browser. Nothing uploaded, no account, no cookie.
How to settle it yourself in thirty seconds
- Count your own pulse first. Two fingers on the thumb side of your wrist. Fifteen seconds, times four, written down before you listen so you are not tempted to make it fit afterwards.
- Keep your fingers there while you listen. The comparison is far easier while both signals arrive at once.
- Compare the timing, not the numbers. If the sound in your ears lands on the same beat as the pulse under your fingers, it is you. Ears are very good at hearing whether two rhythms are in step.
- Check the ratio. A rate close to exactly twice the pulse you counted is yours until something proves otherwise. Roughly twice as fast, out of step, and not a tidy multiple of yours is the pattern you actually want.
The tells, side by side
| What you notice | Almost certainly you | Possibly the baby |
|---|---|---|
| Loudness | Loud and close, there in seconds | Faint, needs a quiet room |
| Character | A heavy double thump with weight behind it | Lighter and quicker, closer to a fast tick |
| Where you find it | Almost anywhere, strongest over the vessels | One small spot, lost if you move |
| Against your counted pulse | The same, or exactly double it | Faster, and not a whole number multiple |
What we measured on exactly this failure
The engine was run against public databases on PhysioNet before any interface existed, including a set of adult heart recordings used deliberately as a trap.
The limit, stated plainly. Those recordings came from research sensors on the abdomen, not from an iPhone microphone against real skin. How much survives that move is the one number we do not have yet.
And if the answer turns out to be you: that is the expected result, it happens to almost everybody once, and it says nothing about your baby. The reassurance you wanted is in movement. The NHS guidance is here, and Kicks Count exists for that reason.
One email, the day it opens
Firstsong is not on the App Store yet. Leave an address and you get exactly one message when it is, with the price on the first line. No newsletter, no drip, no partner offers. You can reply to that email and a person reads it.
Stored on our own server, never sold, never shared. Your IP is hashed, not kept. Nothing about your pregnancy is asked for or stored.
Common questions
Can a baby's heart rate ever be the same as mine?
They can come close at the edges, usually when yours is up from moving around. At rest they rarely overlap, which is why counting your own pulse first is worth fifteen seconds. If they are genuinely close, no method based on sound alone separates them.
My app showed 150 and I did not move. Was it real?
It might have been, since 150 sits inside the fetal band. Ask two things first. What was your own pulse at that moment, and was 150 close to double it? A pulse of 75 doubles to 150, and that is the most ordinary wrong answer there is. Then ask whether that app has ever told you no.
Why does the wrong number sound so plausible?
Because it is not random. It is your own rate multiplied by two, so it is as steady as you are and it sits in the range you expected. A wrong number that looked wrong would have been caught years ago.
Is the mother clip in the tool really a mother?
Yes. A maternal recording from the Shiraz University Fetal Heart Sounds Database on PhysioNet, labelled as such by the researchers who collected it.
What if I can hear two rhythms at once?
That is the good case. One slower and heavier underneath, one faster and lighter on top, drifting in and out of step because they are unrelated. Two hearts, rather than one heart counted twice.
Keep reading
- Hear the difference. clip two is this exact problem, decided in front of you
- The long version. the signal processing behind the error
- Anterior placenta. the other reason a phone finds you and not the baby
- Do baby heartbeat apps work. the sceptical question, answered straight
Last checked 2026-08-23