Acoustic testing finds HomePod's adaptive audio tech delivers highly consistent sound
In its advertising push for HomePod, Apple has touted its smart speaker as capable of delivering consistent sound to listeners positioned anywhere in a given room. New testing backs up those claims, and offers insight into how the company accomplished such a feat.
As part of an ongoing HomePod evaluation, Fast Company partnered with professional acoustic analysis hardware and software maker NTi Audio to measure the speaker's performance in reporter Mark Sullivan's living room.
Setting up HomePod on a table near a wall, a specialized microphone was used to capture white noise test sound from four different locations. Results were compared to generate a sound profile for the speaker, which displayed an average variance of 0.95 decibels across all audible frequency bands.
As NTi notes, humans are typically unable to detect changes below a decibel, meaning this particular test backs up Apple's claims that HomePod will deliver consistent sound to listeners positioned just about anywhere in a room.
"The developers have done an excellent job of having the HomePod adjust to the room; (it has) Impressive consistency in overall level and frequency response," said Brian MacMillan, associate general manager at NTi. "The HomePod automates spatial compensation that previously required a real audiophile's expertise, tools and time."
To achieve such consistency, HomePod employs an array of six exterior microphones and a digital signal processor to analyze its surrounding environment. Another microphone located inside HomePod's chassis detects nearby objects like walls and other large obstacles that might interfere with generated sound waves.
An onboard A8 chip applies collected microphone data to advanced algorithms to dynamically modify sound output from each of the speaker's seven tweeters and its single woofer. These adjustments are made when HomePod is set up, moved and during music playback.
The result is highly consistent, adaptive and immersive audio. Most importantly, the entire process is performed in the background without manual intervention from the user.
The Fast Company analysis validates AppleInsider's report detailing the inner workings of HomePod's audio magic and follows a similar evaluation that found Apple's speaker provides a nearly flat representation of sound, a metric often used to quantify audio quality.
As part of an ongoing HomePod evaluation, Fast Company partnered with professional acoustic analysis hardware and software maker NTi Audio to measure the speaker's performance in reporter Mark Sullivan's living room.
Setting up HomePod on a table near a wall, a specialized microphone was used to capture white noise test sound from four different locations. Results were compared to generate a sound profile for the speaker, which displayed an average variance of 0.95 decibels across all audible frequency bands.
As NTi notes, humans are typically unable to detect changes below a decibel, meaning this particular test backs up Apple's claims that HomePod will deliver consistent sound to listeners positioned just about anywhere in a room.
"The developers have done an excellent job of having the HomePod adjust to the room; (it has) Impressive consistency in overall level and frequency response," said Brian MacMillan, associate general manager at NTi. "The HomePod automates spatial compensation that previously required a real audiophile's expertise, tools and time."
To achieve such consistency, HomePod employs an array of six exterior microphones and a digital signal processor to analyze its surrounding environment. Another microphone located inside HomePod's chassis detects nearby objects like walls and other large obstacles that might interfere with generated sound waves.
An onboard A8 chip applies collected microphone data to advanced algorithms to dynamically modify sound output from each of the speaker's seven tweeters and its single woofer. These adjustments are made when HomePod is set up, moved and during music playback.
The result is highly consistent, adaptive and immersive audio. Most importantly, the entire process is performed in the background without manual intervention from the user.
The Fast Company analysis validates AppleInsider's report detailing the inner workings of HomePod's audio magic and follows a similar evaluation that found Apple's speaker provides a nearly flat representation of sound, a metric often used to quantify audio quality.
Comments
I just find it funny that people reference positive reports of HomePods performance that aren't any more valid than CR's report. An individuals ears have more to do with a speaker sounding ""good"" than the speaker itself. Combine the ears variable performance with the fact that installations vary widely and you can't really say if HomePod is good enough until you actually try it out.
https://finance.yahoo.com/news/head-head-apple-homepod-really-sound-best-160346138.html
Perception of what sounds good to an individual can be coloured by the speakers they regularly listen to.
This is similar to the first time someone views a properly ISF calibrated display panel after watching LCD TV's in "torch mode" for years - a properly calibrated display will leave them looking to crank up the brightness and the contrast.
This is a test by a lab using microphones to judge sound quality variance around a room; it is as objective as any other scientific test, as it does not rely on human ears in any way whatsoever. This is not, like some other articles, a subjective human-based test of which one "sounds better." No comment is made about the sound quality at all -- only that the HomePod's claim that the sound is consistent to below a human's ability to detect variance is supported by scientific testing.
So no, you are wrong. This report is more useful and objective than subjective listening reviews -- just as Jwdav, ihatescreennames, and Rob53 mentioned. it doesn't, and wasn't intended to, make up for doing your own first-hand listening test, but it was intended to, and does, verify one of Apple's claims on the product.
https://www.reddit.com/r/audiophile/comments/7wwtqy/apple_homepod_the_audiophile_perspective/
Small portion of the critique:
Not that it matters though, because I’m willing to bet if someone else do this kind of test the result will be different because that’s what all blind test get. Different results from different groups.
I stated this after hearing AppleInsider’s comparison. The HomePod does sound less open, transparent and the G-Max revealed detail that isn’t audible on the HomePod.
Just listen to it. It’s definitive. (80% volume test).
As far as consistency of sound - What’s better:
- Room Treatment to improve reflection of all sound?
- Algorithms that reduce frequencies on a source to compensate?
https://9to5google.com/2018/02/13/home-max-homepod-listening-test/
Should tell you all about reliability of this kind of testing really.
EDIT: Ah, you meant a former CEO. Found it now.
No, his measurements are still ok. The idea a speaker that adapts its output to a room should be tested in an anechoic chamber is pure horseshit. People don’t live in anechoic chambers - they live in real buildings that have all sorts of varying acoustical properties.
Edited: One thing for sure. Before the HomePod nobody gave a rats ass about doing listening tests or instrumented measurements. Nobody knew what imaging or soundstage are. Now everyone’s talking about it (even though most still don’t understand it). And as usual with any Apple product opinions on the HomePod are strongly divided. And as usual it will be highly successful.
Sonos and Bose should be worried. Apple will sell millions of these, and each sale represents one less potential sale for them. Google doesn’t have to worry about the Max because it wasn’t going to sell anyway (like all Google “premium” hardware).
Not even the guy who did the tests now says that. You plainly did not read the update, which the author himself promotes as more accurate than his original claims which he now says should be taken with a "huge brick of salt".
There is no “huge”, just a brick of salt.
And yes because it’s not done in anechoic chamber. his quote:
”these measurements are largely inconclusive, since the measurements were not done in an anechoic chamber.”
The discussion makes him question his method. But the measurement is still equivalent to the real world usage, so the consumer still get what he measures. His questioning is simply, how are these numbers if it’s tested in a chamber. It’s academic. But his test is realistic.
HomePod is designed to be used in household rooms, not sound absorbed listening room like other audiophile grade speakers.
Actually, This is more important since it is not subjective, it is quantitative It tells you the speaker will reproduce the sound exactly as it should and it will not sound any different in any place in the room.
The only subjective or opinion part is the individual persons ability to determine if something sounds good to them or not. Their evaluation has no baring on what someone elses experience will be. Knowing a speaker will reproduce sound exactly and will not distort it due its only design limitation is the most important thing to know before you introduce someone personal, and most likely unprofessional non trained subject knowledge opinion. If CR did not do a control test first, and they are rendering an opinion it is not worth much.