It’s the time of year for saving money!
In our everyday lives, being accurate about most things is important. For example, when we visit the doctor, we want an accurate diagnosis if we feel bad. If the washing machine is on the fritz, we want the repair guy to be accurate in going about fixing things back to right.
How about in music? In the vast array of sonic presentations available from a huge variety of systems and technologies, how achievable is accuracy as a goal?
I have had people in my audio room with one request – “crank it up!” Somehow, for some, playing music abusively loud is the key to enjoying a song. I’ve had others request a string quartet, turn the volume down lower than audiophile normal, close their eyes and relish, become enveloped even by what they are hearing. There seems to be no correct or incorrect method of listening to a song.
For most of us, from an audiophile perspective anyway, the hallmark of good sound, and enjoyable listening, may be determined by several factors: dynamics, soundstage, imaging, clarity and yes, our good friend accuracy. Remove any one of these attributes and music sounds, essentially, less captivating.
Let’s jump forward a few giant steps and mention digital signal processing (DSP). This is a technology whose time has come. Ten years ago, it was almost exclusively a stand alone component. Now, many components as well as speakers have DSP built right in. There are scores of audiophiles who will swear by the achievements DSP provides.
DSP has, as a utopian goal, the communization of sonics. It alters the digital signal to an algorithmically derived waveform. It has the ability to correct problems with a song based on the room in which the song is being played. Many find this a highly valued achievement.
As audiophiles, we all know there are good recordings and there are less than good recordings. Which format does not matter one whit. Any recording – streamed, CD, LP even reel to reel can be better or worse than another recording. When we play a song and it sounds far worse than something previously played, we sort of roll our eyes and think to ourselves how bad it sounds. Maybe we might make a mental note to not play that song again.
A natural question then becomes, which is more important – accuracy in playback or using alternative means to homogenize sonics?
For some, and I fall into this group, I want the song playing on my system to be as accurate a representation as possible of how it was recorded. I’ll take a poorly recorded song any day over one that is altered or modified by artificial means to achieve a predetermined result – even if that result improves the sonic portrayal. Then again, my preferences apply to only one person, me. And just because I may happen to like something one way does not make it universally accepted. Because in actual practice, my guess is more people prefer playback improvements however they happen – through artificial means or not.
All that said, we have not yet answered the question of the importance of accuracy.
Determining an answer to this question is not quite so easy. DSP alters the sound we hear, that much is true. But does it always and consistently do so at the expense of accuracy? I see that as a very difficult question to answer – difficult because there are so many variables yielding an army of answers.
If we can agree that any change to the recording is an alteration of the accuracy of that recording, we open up the doors to a wide interpretation of sonic changes. If we increase the bass and decrease the treble by the use of tone controls, is that not changing the recording? Isn’t that doing fundamentally the same thing as DSP, only not to the same degree? And if any change in the signal produced by the recording after it leaves the initial source component, however small, is an alteration of that recording, how can we say accuracy is not effectually compromised – however small?
I realize I am being decidedly literal in this analogy. Most people hardly give any of this a passing thought. Tone controls can be adjusted, DSP may or may not be used or even necessary and we can still enjoy listening to a song. That part is unassailable.
But it is worth thinking about. Our hobby seeks to obtain the closest representation of perfection of musical accuracy. We want our systems to sound like live music despite knowing it is not possible. Yet how many of us have said something along the lines of “it sounded like she was singing right in front of me!” We want perfection. We chase that dragon endlessly in its pursuit.
For me personally, I place a high value on accuracy. I want a piano to sound like a real, live piano. I want instrumentals to sound somehow live. I see accuracy equally necessary as any other of our sonic attributes and an essential requirement for the total enjoyment of a song. It follows then, compromising the musical signal, however minimally that alteration may be, degrades accuracy.
My one enduring question, and one, I suppose, without a universal answer, is where should the line be drawn? When have I gone too far in my efforts to improve sonics? If DSP goes too far, how can tone controls not be viewed the same way? If DSP is a welcomed and enjoyable resource, is there no limit on how far accuracy may be maligned to fulfill our own definitions of improving what we hear?
In the end, I suppose it is a question we all must answer for ourselves. Because just like the wide variety of system choices available to play music, there are equally as many ways to interpret how that music should sound. Basically, its listener’s choice.
What kind of DSP are you talking about? If it’s room correction DSP, I would argue those are at least attempts to improve accuracy. As I understand their function, they use a microphone to compare the in room frequency response to the “ideal” of the recorded frequency. Then they boost or attenuate the frequencies the room alters back to the levels in the recording. Granted not all DSP is room correction and some are intentional colorations away from accuracy. However, it doesn’t seem you’ve addressed room correction DSP and its potential impact on accuracy. Or do I not understand room correction DSP?
You are absolutely correct in your perception of DSP.
If you will notice, I termed DSP as an “algorithmically derived waveform.” Many would argue that DSP is an attempt to improve accuracy, as you state in your comment. However, the means by which DSP improves the signal is an alteration of the original recording (done by math) and the one supplied by the source, be it a CD player, streaming device or whatever. If DSP is, in fact, changing the source to create an “ideal of the recorded frequency,” is that not ultimately accomplished by altering the accuracy of that source?
When you state: “Then they boost or attenuate the frequencies the room alters…” is that not changing the original source, and what the artist and recording engineers created? Granted, the source recording cannot help what the room is doing to the overall sonic presentation. But when we change the source, even if those changes are for the better, are we not altering the original recording? And that speaks to the crux of this article — how willing are we to change the source and therefore the accuracy of that source?
Now obviously I doubt there are very many people who would not like to somehow improve sonics. If room anomalies are a problem and DSP can correct them, well, of course it is a welcomed technology. Shoot, it is very commonplace today. I used DSP at one time myself. These days, personally, I prefer to fix the room and not mess with the signal through mathematical processing but that’s me.
This is not an article meant to criticize DSP. Not in the least. It’s not really about DSP at all except that it does alter the original signal. How much or little this matters to you or anyone else is a personal decision. Let’s be honest, I feel sure it would be difficult to find very many people who do not like the changes DSP can provide. My question is this, if accuracy is a stated audiophile goal, is changing or modifying the signal not an elimination of that goal?
Thanks for your comment!
I guess I never understood “accuracy” within the audiophile hobby to apply to fidelity to the recorded signal. I understood “accuracy” to mean fidelity to what the artists and engineers intended the listener to hear. Relative to your point of measurement – which seems to be fidelity to the recorded signal – then yes, you are correct that any form of DSP or tone control/equalization is a reduction of accuracy. It seems to me though that the more meaningful point of measurement would be at the end listening position. At the listening position there will be room interactions (assuming speakers are used and not headphones – which is a whole ‘nother set of variables) that reduce the perceived signal’s fidelity to what the artists/engineers intended. Even the best treated room is going to have a nonzero impact on the perceived frequency response, resulting in an end listening experience that is not “accurate” to the intended listening experience.. So at that point of measurement, then DSP is about “accuracy.” We agree on that last point but were not lining up on where in the sound reproduction path we were thinking about the meaning of “accuracy.”
Thanks for your thoughts.
I believe your comments are an extension of the article, that is, what is a compromise of accuracy and are we willing to accept it?
And I also believe that is a decision only the individual listener can make.
Thanks for the well thought out comments!
I want my audio system to be able to produce an accurate reproduction of what’s on the recording as a baseline. That means that the frequency response is what’s on the recording, there is no added distortion, and similar things, as heard or measured from the listening seat. From there, one can alter the sound to taste, whether that means using DSP to correct bass problems caused by the room, or using the treble control to fix a bad recording.
Systems that cannot produce an accurate reproduction as a baseline tend to make recordings sound similar to one another, and I don’t care for that. I have noticed that the better my system gets, in terms of accuracy and a quiet environment, the more distinctive each recording tends to sound.
I’d like to think of tone controls as user manipulated DSP. Degree of accuracy stands until the tone controls are changed from default neutral characteristics by the user. I think it’s immaterial once those knobs are turned.
Accuracy is the most important thing in Audio. You may like ketchup on your steak or prefer the Mona Lisa as a Blonde but for me that is not what HE Audio is supposed to do. Audio should not be about fixing it should be about reproducing the Artists vision. Many recordings are less that stellar but many are really excellent. I want my gear to show me whats there not interpret it for me. There are so many false assumptions based on many things however a quality operating system requires many things to work and transport me like a time machine to these venues and places around the world. Proper set up, room orientation and treatment, dimensions, choices of gear are all important to the end result, the result of getting an accurate reproduction of the event, some suspension of disbelief, goose bumps. If you don’t have it you are doing it wrong.
The #1 determinant of playback accuracy starts with room design, followed by speaker performance and suitability to the room and listener location, followed by signal path electronics. The first two follow well-known and measurable “accuracy metrics.” The electronics are harder to qualify. The best examples of delivery accuracy will be found in the world’s finest mastering and post rooms, designed and fine-tuned by acousticians.
The advent of object-coded audio (e.g., Atmos, etc.) has the potential to bring us closer to a true concert hall experience, with a 360 degree hemisphere of reality. Recording engineers are just starting to work with immersive experiences, and the early work is stunning.