|Posted by Amine Slimani on October 12, 2010 at 12:46 PM|
I had initially written the article below as part of headphone cable review, but I thought it could be more helpful by being presented as a separate article since it is in no way related to a specific brand of cables or equipment.
I have to say that I have always been puzzled to see on head-fi people spend thousands of dollars on headphone amps and tubes while dismissing the possibility that a “simple” piece of wire could make any sort of difference at all. And of course, according to that group of people, all those who think they hear a difference between cables are probably just experiencing the placebo effect. Indeed, there have been quite a few double blind tests where listeners could not distinguish very expensive cables from very cheap ones (though a high price in cables doesn’tnecessarily mean high quality).
So from that so-called “fact”, there is no possibility according to the objectivists that a “simple” wire could make any audible difference in the “real”world.
However, what double blind tests have also shown in the past is that itis impossible to tell a difference between different transports, DACs, amps, interconnects and speaker cables. So according to the cumulative results of those tests, we should not be able to distinguish any difference between a $100 audio chain from a $100,000 one. The only audible difference would come from the speakers ... and, even worse, if the speakers have similar frequency response, telling the difference should not be possible.
In my opinion, there are 3 flaws in that reasoning:
1. On the measurement side:
People focus only on the frequency domain and don’t know that the human ear is far more sensitive to time domain resolution. It is only until recently that some researcher started to test for the actual temporal resolution of the human ear, which is far higher than expected.
Here (http:www.physics.sc.edu/kunchur/Acoustics-papers.htm) you can read some interesting research papers about the temporal resolution ofhuman ears.
While conducting his research, Kunchur realized that no single CD player could generate properly the square waves he needed for his testing. He had to resort to an analog wave generator to test the limits of human temporal resolution. That means that people that were complaining so far about the poor quality of the digital reproduction of most CD players were dead spot on. Most measurements focused on the frequency domain (frequency response, THD...) but totally discarded the time domain performance.
So while those people might have been laughed at because there was supposedly no measurement to back their claim, you can draw your own conclusions after reading those research papers or by looking at the graphs here (http:www.mother-of-tone.com/conversion.htm).And for those who don’t know it, most CD players and DACs today use the sigmadelta chips that give those horrible high frequency sine waves.
That is to say that if people who complained about the sound of CD players were only affected by placebo, there wouldn’t be measurements to back their claims (jitter, square waves...). So those that were really affected by the placebo effect were those who were told that CD is as good as it gets ...and convinced themselves that is indeed as good as it gets.
The same rule, could be applied to cables, if we measured the impulse response, square waves or better signal music we would have more meaningful results.
By the way, just before finishing the review, I came across an ongoing project by Nordost. See here: http:www.nordost.com/downloads/New%20Approaches%20To%20Audio%20Measurement.pdf.
Instead of measuring static signals, they used real recorded music and their results show that even external factors (such as vibration support andpower cables) will affect the output of audio equipment
2. Most ABX are conducted improperly:
I have been baffled when I bought an AES research paper about jitter a few months ago. There has been a head-fier posting all around the place about the fact that jitter is inaudible below certain (high) levels because no double blind test showed otherwise.
Well to my surprise, I found that they used for the research some cheap $30 Sony headphones for the tests. Those same headphones would have been dismissed by any audiophile (I wouldn’t personally used them even in a portable rig).
Also, there was no mention in the paper about the headphone amp, interconnects, power filtration...
So unfortunately, that paper is used as a “reference” to say that the jitter of the source is not important as long as it is below 10000ps (audiophile equipment has jitter lower than 100ps). What it means is that when other “scientific” double blind tests are conducted on other areas (cables, amps...) they start from the supposition that the quality of the source doesn’t matter.
So when testing to see if there is a difference between cables or amps, they make some a few assumptions backed by previous tests. And the new (yet flawed) tests give them even more "solid data" to back their claims that every thing sounds the same.
So sadly, many of those double blind tests are meaningless, in my opinion, because the systems they use are often not nearly revealing enough.
What is even sadder is that the portion of people who can actually conduct scientific tests is the one that takes the least scientific approach to this problem.
Real audiophiles spend months or even years building their systems. They go through cycles of development, learning and growth. The progression is not linear. I can't feel but surprised when so-called scientists dismiss so quickly parameters that might affect the result of the test.
It is like if a physicit based his calculations on the speed of sound being 340m/s (because it was what his measurements gave him when he did). While it is true that the speed of sound is 340m/s when measured at 0m (0ft.) altitude, the speed of sound drops to 300m/s at 10,000m (32,000ft). So assuming the speed of sound is a constant to construct one's theory is a flawed approach.
The same can be applied to various A/B tests. It is not because under certain test conditions we found that the selected panel could not detect any significant different between 2 pieces of equipment or cable that we can assume that it is always true for everyone.
Conducting a proper A/B or DBT test on a particular piece of equipment or cable would require that all the audiophile "myths" and tweaks be applied because it can only give good results.
If we assume that those myths and tweaks are false and, incidently, don't have an effect on the result, then there is no harm in applying them on the testing.
If we assume that those myths and tweaks are true, but choose to discard them because it is deemed insignificant, then the results are skewed.
3. There is a strong psychological factor to the A/B tests:
First, there is the stress. When we are stressed and are “forced” to hear a difference our minds can play tricks. When we are listening casually, we are in a different state of mind where we are looking to forget about the equipmentand concentrate on the music instead.
Second, when we are doing a double blind test, we concentrate too muc hand our brain starts filling the gaps with missing data that isn’t really there.
Let’s take for example a phone conversation. If we concentrate enough,we can recognize a voice on a poor connection. So how come? Well, our brains interpolate the available data points and check it against our “database” of known voices.
The same is done when listening critically in a double blind test: our brains records the details with the better equipment and then fill in the gaps when listening to lesser equipment.
Unfortunately, this “imprinted memory” fades away after a while.
On the other hand, when we are listening casually, the differences jump at us and are not usually what we expected to hear. When listening very hard to hear a difference we try to simplify things and focus only on limited aspects of the music in order to highlight any difference. For example, when A/Bing 2components on a large orchestral piece of classical music, should the listener focus their attention on the placement of the performers on the soundstage, the size of the soundstage, the timber of each instrument (and which one?), the timing, the tonal balance, the overall transparency, the low level details, the dynamics, the transient speed... There are so many things one can focus on and the more stressed and constrained by time we are, the more restrictive that list become and that is why you find so many people limiting their A/Bcomparisons to bass/mids/highs.
Moreover, even in the "flawed" AES jitter paper I mentioned earlier, it was noticed that the more people listened, the better they became at discerning difference between jittery sources. When you replace hours of experience with a test set-up with months or even years, you start getting the picture. It far more easy to track differences on long listening sessions than on short, stressful A/B sessions where instead of focusing on the music, you have to listen to the sounds, which are 2 very different ways of listening.
To sump up, what good cables, bring to the mix, in my personal experience, is less hard work to“reconstruct” the recorded event. With better equipment in general, your brain doesn’t have to work as hard to visualize the event, which makes it a more pleasurable listening experience on long listening sessions.
This leads me to explain my testing methodology:
Instead of quick A/B testing on a single track, I prefer to listen to a component/cable through a various range of albums, and then take some notes. It is only after I start getting a grasp of whatis going on, that I do a few A/B switches with my reference gear to confirm on infirm my preliminary findings.This usually implies dozens or hundreds of listening hours.
I also like to listen to the equipment at low volume level and high volume levels. Sometimes when we listen exclusively at an average volume level, we can miss some characteristics. For example, a component that can be clear sounding at regular volume levels can turn into acid sounding at high volume levels.
Listener fatigue is also an important factor. Since this whole audiophile hobby is about listening to music in the first place, I automatically discard equipment that can be pleasing in short listening sessions and that end up being tiring on long listening sessions. I sometimes have my headphoneson for hours at a time so it is a very important factor for me.
The difference between an experienced listener and a less experienced one is how easily one can track miniscule changes. However, the additive and cumulative effects of interconnects, power cables, and various "tweaks" is easily noticeable and is what sets appart systems that can reproduce good hifi sound and systems that can reproduce high resolution enjoyable music.
To me a system can be truly detailed and lifeless at the same time. A system that is truly and genuinely detailed should be able to convey the emotions behind the music. If the emotions are missing, then it is only "half detailed".
Categories: Cables & Tweaks