Keith_W
Major Contributor
Sorry for the clickbaity title but I thought it was time for another ASR bun-fight 
For some time now, I have noticed that both groups of audiophiles - subjectivists and objectivists - share the same belief: that the signal should be kept uncorrupted. Purity is all important. The only distinguishing feature between the two groups is the approach. Subjectivists believe their cables, power conditioners, cable elevators, DAC's, and analog systems keep the signal clean. The whole purpose of an analog system is to prevent the signal getting corrupted by digital. The intention of power tweaks is to keep power "clean" because they think they can hear "dirty power". Objectivists believe that the lower the distortion/SINAD, the better. Speakers are selected for low measurable distortion. The whole point of this engineering exercise is to keep distortion and nonlinearities as low as possible, preferably pushed into inaudible limits. There is certainly a large contingent of very vocal ASR members who say that the system is there to reproduce what the artist intended, and nothing more. So you can see, both are the same yet both are different.
However, there are two major aspects to sound reproduction that everyone agrees has a major impact on the sound: (1) the recording, and (2) the speaker/room interface.
The recording: this is the result of a bunch of subjective decisions made by the artist and the mixing engineer on their monitoring system, which is different to your home audio system - i.e. Toole's "circle of confusion". Take a look at the VST plugins offered by only one company, and realize there are thousands more offered by other companies. All these VST plugins manipulate the signal - compressors, expanders, EQ, harmonic distortion generators, emulators for tube sound, vinyl sound, tape sound, reverb VST's, and so on. Some recordings are completely artificial constructs that never existed in reality, e.g. all electronic music. Some recordings of less talented singers have been "autocorrected" to make them sing in tune. Even recordings of live acoustic instruments have been manipulated. I listen almost exclusively to classical music, and I can tell you that recordings sound crazy different from label to label, and even with the same label. Just compare Deutsche Grammofon from the 80's and 90's and current recordings, even though the recordings are digital. They have changed something in their workflow that makes the recordings sound different.
The speaker/room interface: despite the best intentions of speaker designers, the moment the speaker is placed in the room, the response is changed. I think that nearly all of us here on ASR agree that some type of room correction is necessary, preferably with DSP. And if you correct with DSP, you will need a target curve. Here is an older thread on ASR about target curves:
I have tried all these target curves, and they all sound different. All are purportedly supported by science. Some of these curves were arrived at by hypothesizing what the speaker's room response would look like if a speaker that measured flat under anechoic conditions were placed in a room. Some others are based on preference studies, e.g. the Harman curve. Even the Harman curve has three bands, "more bass", "less bass", and "Harman curve lovers". These bands are split into different demographics - younger men and older women respectively for the first two groups, and the largest group being "Harman curve lovers".
I suspect that whether we admit it or not, many of us pick the target curve that we like based on what sounds best to us. So here is the rub: the moment we pick a target curve, we are manipulating the signal to our preference. Even if we choose a speaker that has a certain target curve with no digital signal manipulation, we have still chosen a target curve that we prefer.
Another point of data: valve amps and turntables distort the signal and measure objectively worse. Yet there are many people, myself included, who think that they sound pleasant. Why? My hypothesis was that people just like some additional distortion, so I performed an experiment on distortion - thread here. I deliberately added distortion to the signal, and asked people if they prefer "A or B", without revealing what the intervention was. Nearly all the people I tested preferred some distortion. As I emphasized in that thread repeatedly, this was not a scientific test given that it was performed under uncontrolled conditions with only one system, but the trend seemed significant.
And yet another data point: in another thread, @ppataki recommended I try Pultec EQ to improve subjective bass slam. I downloaded a VST and tried it. It works by boosting bass followed by a little undershoot, i.e. it effectively modifies the target curve. It is still early days in my testing, but I like it very much. I need to throw more music at it to see if it still holds up after a few weeks of listening, but here is another example where signal manipulation has increased my enjoyment of the system.
Conclusion. My system is an unholy hodgepodge where I have manipulated the signal to an extent that would horrify both objectivists and subjectivists. It started by realizing that recordings all sound different, and Toole's circle of confusion meant that my system wasn't reproducing what the artist intended anyway. I began correcting the system with the intention of making the circle of confusion smaller, before I realized that this intervention by itself was manipulating the signal to my preference. So why not go further. I ended up installing all sorts of VST's and rejecting most of them on grounds of taste, so I have only kept a few. The important feature of my system is that I can turn everything off with a few clicks and everything is restored to unmolested signal, with the exception of the target curve that I can not avoid because it is baked into the system.
If you conclude that it is OK to manipulate the signal to your preference, the very uncomfortable corollary is this: there are no standards in audio, nor can there ever be. All the preference scores in spinorama.org are moot. We are back to the Wild West where anything goes. I myself am uncomfortable with this because as a scientist at heart and by profession, I can not accept a universe of disorder and chaos.
So my questions for ASR are: do you think signal integrity is important? Do you avoid all manipulation to the signal? If you did manipulate your signal, how did you choose your target curve? Do you think there is a role for preference when it comes to signal manipulation?
For some time now, I have noticed that both groups of audiophiles - subjectivists and objectivists - share the same belief: that the signal should be kept uncorrupted. Purity is all important. The only distinguishing feature between the two groups is the approach. Subjectivists believe their cables, power conditioners, cable elevators, DAC's, and analog systems keep the signal clean. The whole purpose of an analog system is to prevent the signal getting corrupted by digital. The intention of power tweaks is to keep power "clean" because they think they can hear "dirty power". Objectivists believe that the lower the distortion/SINAD, the better. Speakers are selected for low measurable distortion. The whole point of this engineering exercise is to keep distortion and nonlinearities as low as possible, preferably pushed into inaudible limits. There is certainly a large contingent of very vocal ASR members who say that the system is there to reproduce what the artist intended, and nothing more. So you can see, both are the same yet both are different.
However, there are two major aspects to sound reproduction that everyone agrees has a major impact on the sound: (1) the recording, and (2) the speaker/room interface.
The recording: this is the result of a bunch of subjective decisions made by the artist and the mixing engineer on their monitoring system, which is different to your home audio system - i.e. Toole's "circle of confusion". Take a look at the VST plugins offered by only one company, and realize there are thousands more offered by other companies. All these VST plugins manipulate the signal - compressors, expanders, EQ, harmonic distortion generators, emulators for tube sound, vinyl sound, tape sound, reverb VST's, and so on. Some recordings are completely artificial constructs that never existed in reality, e.g. all electronic music. Some recordings of less talented singers have been "autocorrected" to make them sing in tune. Even recordings of live acoustic instruments have been manipulated. I listen almost exclusively to classical music, and I can tell you that recordings sound crazy different from label to label, and even with the same label. Just compare Deutsche Grammofon from the 80's and 90's and current recordings, even though the recordings are digital. They have changed something in their workflow that makes the recordings sound different.
The speaker/room interface: despite the best intentions of speaker designers, the moment the speaker is placed in the room, the response is changed. I think that nearly all of us here on ASR agree that some type of room correction is necessary, preferably with DSP. And if you correct with DSP, you will need a target curve. Here is an older thread on ASR about target curves:
I have tried all these target curves, and they all sound different. All are purportedly supported by science. Some of these curves were arrived at by hypothesizing what the speaker's room response would look like if a speaker that measured flat under anechoic conditions were placed in a room. Some others are based on preference studies, e.g. the Harman curve. Even the Harman curve has three bands, "more bass", "less bass", and "Harman curve lovers". These bands are split into different demographics - younger men and older women respectively for the first two groups, and the largest group being "Harman curve lovers".
I suspect that whether we admit it or not, many of us pick the target curve that we like based on what sounds best to us. So here is the rub: the moment we pick a target curve, we are manipulating the signal to our preference. Even if we choose a speaker that has a certain target curve with no digital signal manipulation, we have still chosen a target curve that we prefer.
Another point of data: valve amps and turntables distort the signal and measure objectively worse. Yet there are many people, myself included, who think that they sound pleasant. Why? My hypothesis was that people just like some additional distortion, so I performed an experiment on distortion - thread here. I deliberately added distortion to the signal, and asked people if they prefer "A or B", without revealing what the intervention was. Nearly all the people I tested preferred some distortion. As I emphasized in that thread repeatedly, this was not a scientific test given that it was performed under uncontrolled conditions with only one system, but the trend seemed significant.
And yet another data point: in another thread, @ppataki recommended I try Pultec EQ to improve subjective bass slam. I downloaded a VST and tried it. It works by boosting bass followed by a little undershoot, i.e. it effectively modifies the target curve. It is still early days in my testing, but I like it very much. I need to throw more music at it to see if it still holds up after a few weeks of listening, but here is another example where signal manipulation has increased my enjoyment of the system.
Conclusion. My system is an unholy hodgepodge where I have manipulated the signal to an extent that would horrify both objectivists and subjectivists. It started by realizing that recordings all sound different, and Toole's circle of confusion meant that my system wasn't reproducing what the artist intended anyway. I began correcting the system with the intention of making the circle of confusion smaller, before I realized that this intervention by itself was manipulating the signal to my preference. So why not go further. I ended up installing all sorts of VST's and rejecting most of them on grounds of taste, so I have only kept a few. The important feature of my system is that I can turn everything off with a few clicks and everything is restored to unmolested signal, with the exception of the target curve that I can not avoid because it is baked into the system.
If you conclude that it is OK to manipulate the signal to your preference, the very uncomfortable corollary is this: there are no standards in audio, nor can there ever be. All the preference scores in spinorama.org are moot. We are back to the Wild West where anything goes. I myself am uncomfortable with this because as a scientist at heart and by profession, I can not accept a universe of disorder and chaos.
So my questions for ASR are: do you think signal integrity is important? Do you avoid all manipulation to the signal? If you did manipulate your signal, how did you choose your target curve? Do you think there is a role for preference when it comes to signal manipulation?
