Inside the Studio · Round Table Recording Company Blog
Every engineer has had this conversation. The new mix is better by every measure anyone can name, and the artist keeps asking why it doesn't sound like the voice memo on their phone.
The industry calls it demoitis. The same thing happens a stage later with rough mixes: an artist lives with a rough for two weeks while the song finishes, and the final mix arrives sounding wrong. So is this a real effect or studio folklore? The mechanism behind it holds up well. The specific diagnosis is less settled. And sometimes the artist is simply right.
What Are Demoitis and The Rough Mix Blues?
Both describe the same thing at different stages: an artist becomes attached to an early version of a song because they've heard it more times than any other version.
Demoitis usually refers to the original demo, often a phone recording or a quick home session captured in the excitement of writing. Rough mix blues shows up later in the process, when an artist lives with a rough for a few weeks, and that rough quietly becomes the reference point for everything that follows.
Tom Fox described the pattern in Production Expert in September 2026. The artist plays the demo so often that it becomes the definitive version in their head, so a clearer and more detailed studio take arrives sounding unfamiliar even when it's better by every technical measure.
What Does the Research Say For It?
The underlying mechanism has solid support. Repeated listening makes people like music more, and familiarity predicts liking better than the music itself does.
Guy Madison and Gunilla Schiölde tested this directly in Frontiers in Neuroscience in 2017. Fifteen participants rated 40 music examples across 28 separate presentations over roughly four weeks. Liking rose steadily with repeated listening at every level of musical complexity.
Their second finding is the one that matters here. Familiarity with the material predicted liking better than complexity, leading the authors to describe familiarity as the single most important variable for liking.
Apply that to a songwriter who has played their demo 200 times while the mix engineer hears the new version for the first time. Those two versions aren't being judged on equal footing. One has a month of familiarity behind it, and familiarity does much of the work in how much anyone likes what they hear.
Where the Belief Breaks Down
In three places, and anyone who uses demoitis as a blanket explanation will eventually get burned by at least one of them.
First, no study has tested demoitis itself. The exposure research examines how people come to like music in general, not how artists weigh demos against finished mixes. The psychology underneath it is solid. The tidy name the industry put on it came from us, not from a lab.
Second, the shape of the effect is genuinely disputed. A long-standing position in psychology predicts an inverted curve, where liking climbs with exposure, peaks, then falls as listeners tire of the material. Madison and Schiölde set out to test exactly that and found no such decline in their data. When researchers can't agree on whether people eventually get tired of a track, nobody in a control room should feel confident predicting when an artist will come around.
It's also worth saying plainly that their study was small. Fifteen participants, all music-engaged adults, across a limited range of musical complexity. The authors say so themselves. It's good evidence for the direction of the effect and thin evidence for anything more precise.
Third, and most useful in practice: sometimes the demo really is better. Fox is direct about this in the same article that describes the phenomenon. The demo may have been recorded on a night when the artist felt inspired and unguarded, while the studio session happens weeks later when they're tired, self-conscious, and on take nineteen.
The record business has the receipts on that one. The Carpenters kept the scratch vocal on “Superstar” because the performance was judged good enough that re-recording it wasn't necessary. Rupert Holmes found that a re-recorded vocal for “Escape (The Piña Colada Song)” lost the energy and spontaneity of the original take, so the original stayed.
How Do You Tell Unfamiliarity From a Real Problem?
You make the artist be specific.
“It doesn't feel right” isn't actionable, and it usually isn't accurate either. Fox's recommendation is to match playback levels between the two versions, then ask the artist to name the exact moments they prefer in the demo. A timing choice. A particular phrase. The way a word lands at the end of a line.
That question sorts the problem quickly. If the artist can point to specific moments, there's a real performance difference worth chasing. If they can't point to anything and simply prefer the overall feel, you're usually looking at familiarity rather than a defect in the work.
The level-matching half of that advice deserves more than a passing mention, because it's where most of these comparisons go wrong.
How Do You Objectively Test Which Version Is Better?
You run the comparison the way a listening test is supposed to be run: level-matched, blind, and repeated. Anything looser than that measures familiarity and hands you the answer you already had.
Match the loudness before anything else
Demo bounces are usually pushed loud. Max Rayden at Major Mixing points out that demos often arrive heavily limited and overloaded, while a professional mix shows up unmastered, with its dynamics intact. Put those two side by side, and the demo wins before anyone has evaluated a single musical decision, because listeners reliably hear louder as better.
Match them by measured loudness rather than by ear, since your ear is the thing being fooled. Most metering tools report integrated loudness, and several comparison plug-ins automatically trim one source to match the other. A comparison that skips this step was decided before it started.
Make the listening blind
Production Expert traces proper comparison methodology back to Bell Labs research, which established that both the listener and whoever is switching between sources should be unaware of which version is playing. Knowing which file is the new mix is enough to tilt the result.
In practice, that means having someone else load both versions under neutral labels, or using a comparison tool that relabels them as X and Y. Then pick repeatedly before anyone reveals which is which. One pass tells you almost nothing. Five passes where you choose the same file tells you something real.
Judge both against commercial releases, not against each other
This is Rayden's most useful suggestion and the one most artists skip. Stop comparing the demo to the new mix and compare both of them to recent releases in your genre. Your demo isn't the standard. A record that already works on real systems is the standard, and measuring against it tells you which version is closer to finished.
Borrow ears with no history
Someone hearing both versions for the first time has no exposure gap to overcome, so their preference is the cleanest data in the room. Rayden suggests paying for unbiased engineer feedback, and a trusted collaborator who hasn't lived with the demo for a month works just as well.
Let time and different speakers do their part
Set it down for two or three days before deciding, then listen in the car, on a phone, and on headphones. A preference that holds up across all of them is telling you about the mix. A preference that evaporated on the second system told you about your memory.
If your preference survives loudness matching, a blind listen, a few days away, and three playback systems, it isn't demoitis. Take it back to your engineer with specifics, because at that point you've found something real and the job is figuring out what it is.
How Do You Protect Yourself From It?
Stop playing the demo once tracking starts. That sounds glib, but it's most of the intervention. The gap in exposure between the two versions is the whole problem, so the fix is to stop widening it.
Ask for rough mixes that sound like records rather than work tapes. Fox recommends building roughs with some compression and reverb so they land closer to a finished record. A bare, unflattering rough practically invites the comparison you're trying to avoid.
Don't make the call on the day you first hear a new mix. Run the comparison properly instead, using the steps above. And keep what's worth keeping. If one line from the demo genuinely beats the new take, use the line. That's a normal part of making records, not an admission that the session failed.
How We Handle This at Round Table Recording Company
We bring it up before it happens, because it's much easier to name in advance than to argue about later.
When an artist arrives with a demo they love, we listen to it with them and ask what specifically they want to carry forward. Sometimes it's a vocal inflection. Sometimes it's a tempo that would get tightened out of existence. Sometimes it's that the demo was cut the night they wrote the song, and nothing reproduces that except catching the same feeling again.
We build rough mixes that sound finished enough to live with, and we tell artists up front that the studio version will sound different rather than automatically better. Different is the honest word, and it keeps the conversation productive instead of defensive.
When the demo wins, we say so. Our job is the best version of the song, not a defense of whatever we worked on most recently.
If you're sitting on a demo you can't stop listening to, bring it in and we'll work out together what's actually in it. Sometimes we rebuild it properly. Sometimes we keep a piece of it and build around that.
Start the conversation by contacting us today!
Related Links
The Power of Reference Tracks: The Most Underused Tool in Music Production
Before You Hit Record: Why Preproduction Is the Most Important Phase
Frequently Asked Questions
What is demoitis?
It's the industry name for becoming attached to a demo or early version of a song because you've heard it far more often than any later version. The finished recording then sounds wrong to you, not because it's worse, but because it's unfamiliar.
Is demoitis a real psychological effect or just studio folklore?
Both, in a sense. The mechanism is well supported: research on repeated listening shows that liking rises with exposure and that familiarity predicts liking better than the music's complexity. No study has tested demoitis specifically, so the named condition itself is studio shorthand resting on real psychology.
Why does my rough mix sound better than the final mix?
Usually for two reasons at once: you've heard the rough many more times, and rough bounces are often pushed loud while a finished mix arrives unmastered. Match the loudness of both versions, then try to name the specific moments you prefer. If you can't name any, familiarity is the likely explanation.
How do I avoid getting attached to a demo?
Stop playing it once studio work begins, and ask your engineer for rough mixes that sound closer to a finished record rather than bare work tapes. The attachment comes from an exposure gap between versions, so the practical fix is to stop widening that gap.
Can I use parts of the demo in the final recording?
Often yes, and there's real precedent for it. The Carpenters kept the scratch vocal on “Superstar” because the performance didn't need re-recording. If a line or a moment from the demo genuinely beats the new take, keeping it is normal practice.
How long should I live with a new mix before deciding?
Give it at least a few days and listen on more than one playback system. Deciding in the room on the day you first hear it tends to measure how familiar it is rather than how good it is.
Sources: Tom Fox, “Demo-itis: How To Fix It Without Losing Artist Trust,” Production Expert (September 2026); Max Rayden, “What Is Demoitis,” Major Mixing (updated September 2026); Guy Madison and Gunilla Schiölde, “Repeated Listening Increases the Liking for Music Regardless of Its Complexity,” Frontiers in Neuroscience, Vol. 11 (March 2017); Production Expert on level-matched and blind comparison methodology (March 2023). Scratch vocal examples are drawn from widely documented accounts of the “Superstar” and “Escape (The Piña Colada Song)” sessions.
Round Table Recording Company · Indianapolis, Indiana · www.thertrc.com