Quick experiment to start. Imagine you are listening to a narrator in a quiet room, and underneath the voice there is a faint something. Not loud. Not obviously bad audio. Just a little texture.<\/p>\n
Now imagine that texture changes every twenty seconds. That little moment of checking is the problem this part is about.<\/p>\n
Not noise in the abstract. Not noise as a meter reading. Noise as attention leakage: the tiny part of the listener’s brain that keeps checking: what is that? Is it still there? Did it change?<\/p>\n
Audiobooks are odd because the bar is not “sounds fine for ten seconds”. The bar is “can someone live inside this voice for hours without being pulled out of it?”<\/p>\n
There is a story people tell themselves when they start recording: I need silence. And then reality happens.<\/p>\n
Trains exist. Motorbikes exist. Air conditioners exist. And in Thailand, certain birds have apparently decided that human speech is the perfect backing track for a solo.<\/p>\n
So silence is a myth. Or at least, it is not a production plan. The useful goal is different: make the noise boring. Make it stable. Make it non-noticeable. Or, failing that, make clean decisions quickly and keep moving.<\/p>\n
Here is what my sessions look like in practice. I am fussy, and I am fussy on purpose. If a train is going past in the background, I will often just stop and wait. Same with a motorbike. I use the pause to read ahead, reset my posture, check my mic position, and re-enter the scene.<\/p>\n
It makes recording take maybe twenty percent longer than it technically needs to, but it saves time later and keeps the performance cleaner.<\/p>\n
I would rather keep momentum in the story than keep momentum in the waveform. That sounds like a small distinction until you are forty minutes into a chapter and getting impatient.<\/p>\n
At the same time, I have learned that not all noises are equally dangerous. A distant train can be surprisingly survivable if it is low, smooth, and consistent. It may be there, but it does not always demand attention.<\/p>\n
Some noises are different. Asian koels and magpie-robins, for example, are basically recording-breakers for me. They are not just loud. They are patterned, bright, and attention-grabbing. They do not sit under narration. They compete with it.<\/p>\n
Then there is the equipment side. I disconnect the air purifier because it can shift or cycle or change character. I keep the air-con on, on its quiet setting, because it is more stable, and stable noise is easier to live with than heat, sweat, and irritation. I make worse decisions when I am uncomfortable. All of that leads to the useful distinction.<\/p>\n
Stationary noise is steady: a consistent room tone, a quiet fan, a stable air-con whoosh. Intermittent noise starts and stops, or changes: trains coming and going, birds calling, motorbikes, a neighbour’s door, a distant voice.<\/p>\n
The psychoacoustic trick is that intermittent noise often reads louder than it meters, because your brain treats change as information.<\/p>\n
A steady low hiss can vanish into the background. But an intermittent chirp, even if it is not technically very loud, grabs attention because it is new. It has edges. It has meaning.<\/p>\n
That is why bird calls are so destructive. Their whole job is to be noticed, and annoyingly, they are very good at it.<\/p>\n
The practical issue is not only the noise itself. It is what happens around the edits. A bed change is when the noise floor or ambience shifts between cuts. You remove a noise event, and suddenly the room underneath the voice changes.<\/p>\n
You might hear it as:<\/p>\n
It is one of the easiest ways to break immersion without immediately knowing why.<\/p>\n
So the question is not simply: can I remove this noise? The better question is: can I remove it without creating a bed change that is more distracting than the noise itself?<\/p>\n
I use this to keep sessions sane. First: is it intermittent and noticeable? Bird calls, motorbikes, horns, a single clunk, a voice, a slap of wind. Default answer: retake or wait.<\/p>\n
If it is noticeable once, it will probably be noticeable later. And if I try to remove it surgically, I may create bed changes or artefacts. I would rather pay for a clean retake while I am still in the performance than pay for a fiddly repair later.<\/p>\n
Second: is it stationary and low? A stable quiet air-con, very low distant rumble, consistent room tone. Default answer: usually leave it.<\/p>\n
Trying to erase it can do more harm than good. Heavy denoise can shave life off narration. It can smear consonants, mess with breath texture, or create that slightly phasey, underwater sound that screams “processed”.<\/p>\n
Third: is it predictable intermittent noise, like a train? That is the grey zone. If it is strong enough to be noticed, or if it changes the bed, I wait it out or retake. If it is far away, smooth, and I genuinely need to keep recording, it might be workable, but only if continuity survives.<\/p>\n
My default is still to wait. Waiting is clean. Waiting preserves tone. Waiting keeps the edit boring.<\/p>\n
Denoise feels like a magic wand. The trouble is that narration is full of low-level detail that makes it sound human:<\/p>\n
Aggressive noise reduction does not only remove noise. It can remove some of that detail, or replace it with modulation artefacts. That might not be obvious over one sentence. Over a whole book, it can become tiring.<\/p>\n
So denoise is a tool, but in this context it is usually a last resort, not a default setting I use just because the waveform looks untidy.<\/p>\n
This is a demo that makes the problem audible. Take a short narration paragraph and put a train bed under it. Use a real section from a recording if you have one. Then make two edits.<\/p>\n
Version one: leave the bed. If it is smooth and consistent, listen to how quickly your brain can stop caring.<\/p>\n
Version two: cut around the loudest train moments, so some parts are clean and some parts have train. Now listen again.<\/p>\n
The train itself may be less present overall, but the bed change becomes the star of the show. The listener’s brain keeps detecting the switch.<\/p>\n
That is the paradox. Removing noise can create a more noticeable pattern than leaving it alone.<\/p>\n
By this point, the decision rule is simple: intermittent, attention-grabbing noise gets retakes, or I wait it out. Steady, truly low, non-modulating noise usually gets left alone.<\/p>\n
And to avoid breaking flow, I use a simple marking method. I can drop a marker in the DAW, clap once, or say a quick low “mark” and keep going.<\/p>\n
The important thing is that the mark is not an invitation to start editing while I am still recording. It is a breadcrumb for later.<\/p>\n
The exact method is less important than the behaviour: do not stop to solve the problem mid-performance. Flag it and move. Because speed does not come from pushing through every noise. Speed comes from not shattering focus every time reality makes a sound.<\/p>\n
Noise is mainly a focus problem and a continuity problem. The wrong noise pulls the listener out. The wrong repair creates a new thing to notice.<\/p>\n
So the aim is not perfect silence. The aim is stable, boring audio that lets the story stay in front.<\/p>\n
Next time, I will deal with the other small problem that becomes enormous over hours: clicks, breaths, and plosives, and why prevention beats repair.<\/p>\n","protected":false},"excerpt":{"rendered":"
The noise that steals attention Quick experiment to start. Imagine you are listening to a narrator in a quiet room, and underneath the voice there is a faint something. Not loud. Not obviously bad audio. Just a little texture. Now imagine that texture changes every twenty seconds. That little moment of checking is the problem […]<\/p>\n","protected":false},"author":34,"featured_media":36405,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_feature_clip_id":0,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_post_was_ever_published":false},"categories":[147],"tags":[7881,7882,8309,8311,8310,6686,8307,8308,3674,7894],"class_list":["post-36174","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-blog-posts","tag-audiobook-audio-v3","tag-audiobook-production","tag-background-noise","tag-bed-changes","tag-bird-noise","tag-denoise","tag-recording-noise","tag-room-tone","tag-trains","tag-voice-recording"],"jetpack_sharing_enabled":true,"jetpack_featured_media_url":"https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/07\/07-noise-you-notice-hero.png","_links":{"self":[{"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/posts\/36174","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/users\/34"}],"replies":[{"embeddable":true,"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/comments?post=36174"}],"version-history":[{"count":2,"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/posts\/36174\/revisions"}],"predecessor-version":[{"id":36404,"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/posts\/36174\/revisions\/36404"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/media\/36405"}],"wp:attachment":[{"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/media?parent=36174"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/categories?post=36174"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/tags?post=36174"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}