Everything after the link in Post #17 was a quote from the Stereophile article. I'm not sure if you were mixing up what I was saying vs. what John Atkinson was saying.
With care, most of the things they measure in a Stereophile speaker review can be measured by an individual these days. To do it accurately requires understanding the methods and their limitations though. It's very easy to make bad acoustical measurements.
If I understand your question, this is likely due to resolution limitations in the quasi-anechoic method typically used by anyone without an anechoic chamber. To collect accurate quasi-anechoic data in a normal reverberant room, you have to window out (exclude with a gate) the reflected speaker output that is bouncing off walls, equipment, etc. You only want the sound that travelled directly from the speaker to the microphone in this scenario. The shorter that time window is, the less resolution there is in the frequency response. This inherently smooths small peaks/dips in the frequency response plot. In an anechoic chamber there are more options to collect accurate high-resolution data.
It's pretty common these days to apply additional smoothing to the data on spec sheets as well. Most acoustical measurement software includes smoothing based on fractional octaves (1/3, 1/2, etc.).
Also, in Stereophile reviews they spatially average the upper portion of the frequency response (above 1 kHz) to smooth position-dependent anomalies that are very common and can be misleading.
Last edited:
i was referring to this passage Probably are Stereophile words It seems that in general magazine cannot stay completely scientific and so objective
The mix words like coloration poor presentation ot timbres with FR and distortion that are science
Some mags like TAS completely avoid any kind of measurements in favor of listening impressions
Sounds science is physics not literature I much prefer the lab approach I am willing to select a speaker just on the basis of a set of measurements without even listening to it
The biggest problem i see is that amps are tested on pure resistive loads while speakers are a more complex load
The solution could be to look for 2 ohm stable amplifiers They should drive most speakers around
Stereophile does use a speaker simulating load for some tests.
There are also things like the PowerCube from AudioGraph that tests amplifiers into a range of reactive loads. That data isn't published very often, but my guess is tests similar to it happen a lot in the background.
There are also things like the PowerCube from AudioGraph that tests amplifiers into a range of reactive loads. That data isn't published very often, but my guess is tests similar to it happen a lot in the background.
now i see thanks what i find missing in Mr Atkinson set of measurements are distortion measurements Imho they can be very telling
I really do not why they never check the dynamic ability of a speaker i.e. the SPL generated at a certain distance and frequency without exceeding a certain level of distortion Maybe i am naive but i think that distortion is a bad thing Why not check ?
yes and thank you very much for your kind and valuable explanations
But the very amusing part is when a reviewer provide a in room FR that looks very badly and then explain that is mostly due to the listening room not adequate This is comic Better doing nothing at all Measurements are a serious thing Much more than listening with old dirty ears
John Atkinson instead is much more scientific But he skipped distortion measurements that i rate very important At least to avoid to break drivers
i see i guess that after all this FR in room are just descriptive And clearly depends on the environment were the measurements are taken
Just to explain what i mean about distortion measurements
https://www.audiosciencereview.com/...ds/revel-m106-bookshelf-speaker-review.14363/
this graph shows the range of usability of the speaker in terms of distortion and spls
this speaker has no chance at all to reproduce correctly frequency below 100Hz and 96dB
to get the full range a sub is needed There is no need to listen after seeing this A limited range speaker
great from about 200 Hz and above by the way
Hi, this could have been that the other didn't have as good of a channel balance for example. I could also speculate particular harmonic distortion profile could enhance the original harmonics, while soume could compress, and also the noise could contribute. If this happened, the distance where perception changes (auditory system pays attention) could have changed, from front of you to behind you. If this happened, you would have needed to move bit closer to speakers with the other amp to make auditory system lock in and provide similar perception. This would also change D/R balance, so sound of the room would change as well if you move closer.
Yes, very much so. Both by either helping you auditory system to pay attention, also how the envelopment and sound stage is affected in many ways so sound would change at least when listening close enough, when auditory system pays attention. Listen too far in the room and small changes with speaker position don't make too big of a difference, because sound of the room dominates perception auditory system doesn't pay attention, so the sound is quite undefined and mushy in typical domestic room no matter what direction your speakers point at.
For exanple, if you have active speakers woth DSP go listen otherside of the room and start playing with tweeter delay. I notices the sound changes hardly at all until the delay is significan't enough so that power response starts to dip. Conversely, go listen close enough to speakers so that brain pays attention and smaller changes seem to change the perception.
luckily, I think it's just about listening skill, which in my view is howbwell you understand the perception. One day this stuff goes by as unnoticed or insignificant, but if you start experimenting with listening distance and how perception changes, you might finb out there is a difference, maybe a critical one. Luckily, most people seem to have speaker setups so that sound is fine anywhere, listening without brain paying attention, most music sounds just fine and is relaxing. But, in my opinion the magic sound is only when brain pays attention, the music goes really deep into heart (brain) then, also the sound has more nyance to it, much more overall experience making it feel like magic, the magic I know. Huge part of auditory system is unconscious processing of the input signal from ear, but also other senses and feelings and memories affect. What you perceive is not the soubd that enters ear canal, but what your unconscious self provide into existence, into your conscious perception 😉
It's something you cannot bypass with your conscious mind, but you can indirectly affect it of you have some experience how to do it. Most significant and easy method I know is move myself closer to speakers, paying attention how perception changes. The speakers and the room are exactly same, but your processor the brain adapts, and it's now easy to move yourself to a position which provides perception you most prefer at that moment. It's really powerful thing
Last edited:
very interesting Of course speakers parameters vary a lot I remember the famous Apogee Scintilla called like that because some amps caught fire ?
scintilla in italian means spark 😵
seriously i know that a speaker has a varying impedance and inductance and capacitance In a word a mess
i understand that even with small speakers to have a robust amp can be a good think expecially during peaks
i think i saw something similar in The Audio Critic magazine But as i said i am always looking for low THD speakers
Small speakers are limited range by their nature They need a sub to playback all instruments at realistic levels.
In-room response is still a useful measurement, especially if the reviewer has measurements from multiple speakers to compare. There are things simple on-axis anechoic response does not show. In-room response helps bridge the gap in information if a full power response/Spinorama dataset is not available, which is often the case with older measurement methods. It can also help the user position speakers more effectively.
The Klippel system used at ASR and Erin's Audio Corner is about a US$100,000 piece of equipment that hasn't been around very long. Most of Stereophile's methods were developed years ago, and I think they value having a consistent test set to compare speaker data over time.
With previous methods, these tend to be a little difficult. If you want to measure down to 0.1% distortion, you have to be able to measure 60 dB below the reference level. For 0.01%, you need 80 dB below the reference level. A typical quiet home has about 30 dB of background noise, and most measurement microphones run out of linear range around 120 dB. Averaging measurements helps, but takes more time. Nearfiled measurement helps, but complicates things with multi-way speakers, and again can run up against microphone limitations. However you want to do the measurement, high resolution distortion data is not easy to gather in a typical home or industrial setting. I think the newer Klippel software does more analysis of background noise levels to try to improve accuracy, but again this is a relatively new software package.
Good morning i see your point what i find weird is that a person who reviews for profession does not take care of treating his listening room and graphs show I see this as extremely unprofessional like a surgeon with a kitchen knife
at the beginning I was passionate about reading reviews
now I can't stand them anymore I want to see the measurements i have enough of chit chat
The distortion graph that I pasted for me is extremely telling
One just has to establish what frequency range he wants to reproduce and at what spl As I was saying the small two-way in the test doesn't have the slightest chance of reproducing some instruments with deep bass It just doesn't get there And this reflects my direct experience
The interesting thing is that you don't even need to listen You can predict from the measurements
I'm convinced that you can choose a speaker based only on measurements The coloration the timbre the transparency ... all things that the right measurement highlights
yes but is it really needed ? i think that a much more basic rig could be enough to provide an idea of the performance of a speaker
FR depends heavily on the room Not so distortion measurements
The real problem is that reviewers use fantasy names The speak of compression while they should speak of distortion
I guess the reason could be that most speakers have poor performance and high distortion And mags live on commercial speakers
Not so in the pro sector Both PA and studio applications They mention often the distortion performance I like this approach much more
Not my case I want to see what kind of response i can get without going above 5% of THD and IMD for instance at lets say 95dB/1 meter
if the result is that the woofer/speaker stays below that threshold only down to 200Hz this means that some music will be played back badly
like a Toccata e Fuga on a pipe organ or some bass lines or some drums Along with FR and maybe CSD would be more than enough for me
thanks a lot again It is a fact that in the past great speakers have been designed and built without Klippel Today just a calibrated mic with a sound card a pc and a sw if used correctly should be enough to design a speaker Also a DATS unit or similar could be very handy
I am sincerely impressed by this little DATS unit i have a woofer here with a very big magnet but a small diameter coil gap A stupid woofer imho
anyway i have lost the dust cup and there is dust in the gap I can see that the impedance curve is ragged I guess due to the dust in the gap
I love this little device a lot It provides automatically TS parameters of ANY driver Unbelievable
one of my main duties at work was to read operational instructions for workers
someone had noticed that if I understood them, 95% of the other workers understood them too
the remaining 5% didn't want to understand them, in short
Yes.. mostly. However once you have a speaker that performs in your room, you still need to voice it. Let's be clear though, if you ask people, voicing is often done to smooth away room problems but that's not what it should really mean. Voicing would be better described as making it sound natural, which is not always flat.
If you design your own you can make this less dependent. Your listening position will measure similarly to the anechoic version.
you mean to design the listening room ? that is just a dream
I have a little anecdote told to me by a technician of audio installations in cinemas and small theaters
One day it happened in a small theater with particularly well-finished acoustics They put a portable radio on the stage on a chair
The sound filled the theater in a very impressive way
a dedicated room remains my dream At that point everything could become less difficult to achieve