How to evaluate changes in a playback system.
First know that it’s your system and that playback is art. There is no such thing as "absolute sound" or "perfect fidelity" and not everyone values the same things in music or listening.
Some facts to consider:
- Playback systems function as a system and any change within it will very likely affect how you hear it. A change anywhere within a system can and should impact the system’s sound.
- We can never assume that an assembled array of playback equipment, or any single device, is perfect.
- The perfect recording does not exist.
- Things change over time.
- People are easily fooled.
- Each person listens and hears differently.
- All systems are different and a device under test will perform differently in them.
- Systems will sound different at different levels and most people tune for a particular listening level, music, or coloration emphasis.
- Unless you are the producer, recording engineer, tracker, mix guy and mastering man (rarely even the same team let alone the same person) responsible for the recorded material, with a vast understanding and experience with your recording and playback systems, you can never know the recording’s nature and accuracy. Besides, isn’t music about fidelity of emotion? (If you roll in at a Zu demo, chattering about how you know what a particular recording is supposed to sound like… well, let’s just say it’s hard to take your comments seriously and it’s possible that we’ll call you out if you start stinking up the room.)
- General listening-specific assumptions can be made after a large enough sampling of systems and observers has been taken. But these are only relevant to the state-of-the-art of playback systems within their sampled generation. Loudspeakers that are known to sound good in today’s systems will generally sound bad when used within a system from the ‘40s or ‘50s. That is not to say a loudspeaker from the ‘40s will sound bad in today’s systems—it’s all about the matchup and knowing where you the listener places value.
So, acknowledging all of these things, how do we approach the evaluation process? It’s easy to present situations that cause a desired result, both in oneself and in those you wish to influence. For the most part, it’s quite easy to recognize performance changes relative to primary system attributes. Shades and gradients of primary and secondary attributes are much harder to wrap your ears around, but it is usually within the combined secondary attributes that we find the magic we are looking for.
Zu makes cables, so let’s take a look at cable evaluation. If observers are really trying hard to hear a difference between cables because they believe the difference will be subtle, this will compound the potential for outside or psychological influence. If the observer has a predisposition towards a model or brand, there will likely be subconscious events that will shape the outcome.
Also, if a demonstrator wants to influence the observer, there are several tricks he or she may use to produce the desired results:
- The second time listening to the same material- the “second set”- results in increased awareness by the observer, particularly after a brief calm or distraction. Therefore the second device tested is usually psychologically perceived as sounding better.
- Observers usually relate a small increase in amplitude (1 – 3dB) to increased fidelity, especially if the cut is played at a level that is lower than you would experience it in real life (low-level listening is distortion). The closer the intensity of sound is to what we expect, the better we think it sounds.
- A demonstrator may use playback of recordings that are known to work with or against the bias.
- A demonstrator may use playback with equipment that works with or against the bias.
- Observers are influenced by the demonstrator. Combine this with the “second set” phenomenon and the demonstrator can easily skew data.
- Rumor, reputation, price, looks of the device and personal investment effect our values.
What a Fair Test Looks Like
The only way to know the performance of a device is to follow some semblance of scientific observation. A double blind test, within a large enough sampling of gear and people, is best- as long as the observers keep their notes to themselves until the completion of the study. But even then, the whole system thing is pretty massive.
A simple cable test should consist of two or more observers and a controller, all of whom have at least a basic understanding of musical acoustics and can accurately communicate using musical or scientific terms. The controller must not communicate anything to the observers during the test except the test number. The controller must also account for any change in amplitude between the two cables prior to running the test. (Electrical characteristics of a cable influence power transfer between transmitter and receiver.)
Then there is the very large topic about recorded material and how it factors in, but a song or two should be selected and the set played twice without change to the system. This is so the observers can listen to the music the first round and then take notes as she or he listens to the set the second time.
There should be two cables to be observed; three usually increases the level of complexity and duration to unusable limits in terms of information and listener fatigue. Observers are not allowed to know which cable is under test or know the device cycle.
The controller randomly changes between the two cables, sometimes leaving one cable in for multiple cycles and so on. The evaluation session should not last longer than an hour.
All of these factors combined should make for a fair test.
But the bottom line is that you must search for your sense of sound and trust yourself-you know what you are looking for. If you don’t, you won’t find it in the forums.