Lii Song has been founded by a small group of people with love and passion for the beauty of full range speakers and we are honored to have been supported by many from the audiophile community, which encouraged us to go this far. Positive or negative, all constructive opinions from people help us to grow and improve and are much appreciated.
However, we found that there is one topic that often generate unnecessary debates and conflicts, that is related to the endless “holy war” between “hearing” and “measurements”.
Concretely speaking, we hold in belief that hearing is more important than measurements but not saying that measurements are not important, or else we won’t spend time and energy sending our speakers to professional lab with standard anechoic chamber and equipment to get the test report. But as always said, we develop speakers that is pleased to our ears and just let the lab test afterwards and publish the raw report to public regardless it’s beautiful or not.
However, even doing like this there come still doubts and one most typical criticism is that “I tested the SPL curve and that’s totally not like what you published, are you honest with your report?” Which we find tiring to explain again and again, thus in this article, we would like to discuss all the details in this aspect.
- How the specs are measured
We design drivers with our ears and our hearts. After we are satisfied the drivers will be supported by third party professional lab (objective) measurements for those who need this. The reason we do like this is that formally measuring is much more complicated than what many think.
To start with, the equipment of lab and the method of measurement should follow the norm ISO 3745:2012. This norm is a regulation that has 62 pages of content in English version. The table of contents looks like this:
The English version is paying so in case you are interested, here is the link to the Chinese version free to all:
By the way, in the norm, even the temperature and atmosphere are to consider and control for measuring the speakers.
Besides, the anechoic chamber and equipment of such a lab cost more than US$ 1,000,000 (equivalent value) and look like this:
Even the floor is treated with anechoic pads and everything suspended on a steel wire net.
If someone believe that one can hand make an airplane in the backyard better than Boeing 777, then he/she can believe that a measurement at home with a mic and software is more reliable than these.
- Is a smooth curve unreal and “beautified”?
One doubt we often hear, not only towards us, but towards almost all manufacturers, is that the published curve is “beautified”. The doubt itself might be OK but their reason is weird: because the published curve is smooth.
Many home measuring experts believe that the SPL curve should be a bunch of peaky waves like the following:
They believe that any smooth SPL curve is fake or edited to look “beautiful”.
But let’s take a step back and look at some basic physic rules. Firstly, a simple question:
Which of graph above, is showing a car’s speed through time in REAL WORLD?
For anyone with basic knowledge, there will not be a doubt to choose A, because in the real world, the physic rule exists. With inertia, the movement status can’t be changed instantly. Which, reflected to the graph, is that there can only be continuous curve but not a broken line. That’s also why SPL curve is called “curve”.
If we can have agreement on the car speed graph, then the same logic can be applied to SPL curve. Can anyone believe that the speaker cone can stop vibration at a frequency point but all of sudden at the next frequency point the vibration gets back? It is like believing that in actual world a car can instantly change its speed from 100km/h to 10. If this can be true, congratulations! The one who proves this has found the way to break inertia and deserve a Nobel Prize!
The SPL curve, if accurate, should and must be a smooth and completely continuous curve.
Besides, if one sees small waves in SPL curve it is consequence of room echo and interference of back wave. Even in the extreme anechoic chamber above these are not completely disabled. So in standard norm the final result should be actually a calculated one excluding all the interferences.
For avoiding debates we even publish now only the raw curve without computing…
- You might have reason on the smooth part but how can you explain that everyone finds the high end not reaching as high as you announce?
It’s a good question. As far as we have seen, in every home measuring result, none of our speakers can reach 20k as published on our page, almost all top-end result drops greatly like the following:
Before getting into further explanation, most people above 30-year-old can’t hear above 16kHz and above 50 can’t hear over 12kHz. Anyway, although one can’t hear one might still “feel”. So generally if the speaker can’t reach the top end it would be disappointing. So we can completely understand how someone will be disappointed seeing and believing this result.
However, we would like to remind a little reality.
The microphone has also its frequency response curve.
This is the fr curve of Shure MV7 of $250:
And this is the fr curve of Neumann U87 AI of $3200:
Most common microphones, cheap or expensive, are not designed to collect ultra-high frequency above 10kHz and this is a typical fr curve:
This might partially explain why it’s common believed by home measuring experts that manufacturers are “lying” on how high their products can reach.
- Why you believe ears are more important than measurements?
So why are people getting into HiFi? Is it for enjoyment or for realizing any technical achievement? Speaker is merely magnet+coil+cone+frame and other parts so why are developers like us spending so much time, money and energy to develop one after one and why are audiophiles paying a great more than the raw material costs to buy them? In terms of Hi-Tech a speaker is far inferior to a MacBook so why do people think it’s worth buying it much expensive than a MacBook?
Among our team we have someone working in speaker tycoon since 18 years old developing monitor speakers for audio studios. He knows that monitor speakers is the best matching of evaluation by specs: flat curve, minimum distortion, correct phase compensation, etc. However, we all agree that we are seeking for something more instinctive, and love to do something that can bring good feeling and enjoyment.
There is always difference between technical sound and musical sound. A simple example, piano does not flatly present the sound of the string inside, on the contrary, the sound of the string is reflected, composed, resonated, and finally presented into the ear to bring enjoyment to listener. Same case with violin, guitar, and almost all instruments… Also, the structure of theatre is designed to reflect and focus the sound to the audience. If someone does not enjoy merely striking a steel string or watching an opera in the wild field, he/she might understand why having musical signature is important.
Of course, this does not mean that reality can be ignored. No one would enjoy if piano sounds not like piano and voice sounds not like living human.
They key however, is to get a balance that one can feel the beauty and enjoyment. This is the belief and practice of our brand and we are only able and seek to serve people who share the same value.
- Are you saying that measurements are not useful?
If anyone claims such and tells you to completely ignore technical data and follow only your instinct, it is likely to be a cult leader wishing to collect what we call “IQ tax” from you. Specifications are important, in a way, but the overall judging system is not superficial, and one needs to understand what each spec is and how it will affect the performance in real life.
Giving this case as an example, the following SPL curve taken at home that looks very disappointing for our flagship Silver-10:
Ignoring the content that we have explained above, we can see that in the SPL curve, there is a sudden fall between 900 and 1k Hz and the bottom is at 950 Hz of 15 dB less (and greater 20 dB fall at 6k Hz.
The unit of decibel is a unit describing sound pressure values on a logarithmic scale. Which means, for an additional 10 dB, the sound pressure would become 3.162 times than before, so for a 10 dB less, it would be 1/3.162 of before. For 15dB less, the sound pressure will become 1/5.622 of before, that is less than 20% and you can consider it almost muted.
So if this curve is accurate, it will mean that, if we play the speaker on a normal sound volume, the 950Hz note would almost disappear.
In the A440 pitch standard, 440Hz is the standard pitch of la (6, a1) and 950 Hz is almost corresponding to High si (7 with a point hat, b2).
This is a note of pitch within the range of human voice (everyone can easily sing it) and can be easily find in a great deal of songs.
So, in another word, if physically such case can be true, that our Silver-10 flagship can sound other pitch notes while only muted on 950Hz, anyone can easily distinguish that in a song a note is disappeared. Imagine the famous song “do-re-mi” be heard like “Do a (whisper), a (whisper)male (whisper)”, one does not need a machine and can easily find out that something is greatly wrong.
It was a sincere criticism without any bad intention, but still very offending in point of view of professional who understand what it means. The problem is that many are too confident on machines but haven’t yet realized that there one needs lot of knowledge on “how to” and “what is”. Even the most experienced member of our team, worked in the industry for decades, dares not to tell that he can make judgement just with the specs.
That is why we recommend people to hear and judge with what they feel and offer a 45 day trial period. We think one doesn’t need to have learnt in a music academy to enjoy a concert, also we don’t want to turn the enjoyment of music into a PhD thesis defense with data and sheets.
We totally understand and agree that our products might not reach the expectation or fit the taste of everyone, and welcome and appreciate all opinions, positive or negative. However, we wish that if someone would give negative feedback, it will be like “I don’t like the sound of the speaker” instead of “My machine says that the speaker is bad so I believe it”.