Hung0702
Head-Fier
- Joined
- Dec 19, 2008
- Posts
- 98
- Likes
- 26
Without a nearby professional audio distributor, few can listen before they buy. Many noobs (myself included) often rely on impressions and reviews rather than personal experience. Each reviewer, however, cannot be expected to define and explain each technical specifications in every review.
I will try to explain the numbers based on my sophomoric knowledge of circuits and helpful articles regarding the science. If you discover my findings inaccurate or misleading, PLEASE highlight them. The last thing I want is for newcomers to leave with the wrong overall impression or a misinterpretation of the significance (or insignificance) of the information. I hope you receive a context to help you decide the importance of manufacturer assessed technical data. The data may or may not affect your listening experience, I'm not trying to tell you it does or doesn't.
Although these specifications are facts, good sound quality is not about any single fact, or any big group of facts. It is about ALL the facts, and since we can never know EVERY SINGLE THING about anything, we cannot reduce sound quality to figures and graphs. Further, even the most accurate testing equipment and well-designed experiments can never replace a human ear. YOU are the judge and you can't selectively evaluate a lone aspect with your ear, so the readings from a microphone or meter tuned can be turned on their heads.
[size=small]Frequency Response:[/size] [size=x-small]Ranged and Charted[/size]
Wikipedia defines frequency response as "the magnitude of the system's response, measured in decibels (dB), and the phase, measured in radians, versus frequency."
Remember that earphones reproduce music. The most accurate earphones recreate audio as it was intended by producers. Many companies deliberately design their product to emphasize certain regions over others. This results in the louder bass, mids, or treble, a colored sound. Manufacture defects are ubiquitous, but should be understated. Undesirable distortion plays a much smaller role than unwanted color (both from defects). For a certain sound signature, distortion and color are required, and the former could play a larger role if the developers choose so.
HD800: 6–51,000 Hz (-10 dB); 14–44,100 Hz (-3 dB)
The frequency response threshold lists audible frequencies. We can hear from 20 Hz to 20,000 Hz, but the cutoff at either end is variable and dependent on the person. I personally have never heard audio below 35 Hz (may be a deficiency in my equipment) and can distinguish audio as high as 22,000 Hz. Sennheisers specs reveal that seemingly concrete FR is actually dynamic and dependent on volume. The many electrical properties of the many components within the HD800 respond differently to different testing conditions. The lower volume has a wider range, but this is no comment on how it sounds at different volumes. Again, when you see a figure, you must not treat it as a thumbnail for the whole image. It's merely the file-size, to use an analogy.
As a buyer: Don't worry about this too much. Most earphones can at least cover the entire audible range. Harmonics are an important aspect of music, but are handled different by the earphones themselves, but this isn't addressed by a FR graph.
Seeing this graph may raise some questions. Hills and valleys for the HD800 are frequent and pronounced. Many consider +/-3 dB an audible difference, but I HIGHLY recommend you download SineGen and explore how much 1, 3, and 5 dB increase the volume. You'll see that these peaks in amplitude are no small variations. But a flat response graph cannot be determined for every person. HD800 reviews typically stress neutrality. Our ears do not interpret sound linearly. In fact, in graphical terms, leftward frequencies need to be stronger to sound as loud as those on the right. People perceive higher frequency sounds as louder, even if they measure the volume. Further, the shape and travel of our ear canals were naturally selected for survival, not audiophilic enjoyment. We hear best in the range of vocal frequencies (mids). The most common type of FR graph is one derived on microphone + algorithm correction. A microphone intended to test audio equipment will yield a FR graph and each point will be shifted to account for resonance in ear canals. A human generated FR graph will be subjective, so each person's graph differs. Almost always, the former type will be provided.
As a buyer: you are probably a noob like me and should ignore this graph. Unless you know why it's important, don't treat it as a deal-breaker/maker.
More info you might not care about: the bode magnitude plot is most commonly used to represent FR. Head-Room uses reference audio equipment and attempts to correct for canal-derived differences. Educate yourself at Wikipedia or Carnegie Mellon's brief MatLab tutorial.
[size=small]Impedance:[/size] [size=x-small]A Measure In Ohms[/size]
The Ohm (Ω) used in alternating current represents impedance (Z). Impedance relates resistance (R), inductive reactance (X_L), and capacitive reactance (X_C). Due to my lowly high-school physics experience, I'm not too well-versed, so please look here and here for further information. Although, if you weren't classically educated in circuits, you might not have too much luck.
As far as I know, V = IR applied to AC yields V = IZ. Thus, a high impedance requires higher voltage given the same current. At the same volume from a source, a higher impedance driver will not travel as far. In regard to sound quality, if the current demand is not met, then bass and impact should suffer significantly. The driver will not travel as far for higher frequency, so highs will not sound as degraded as lows. Sensitivity is more relevant to volume and will be discussed later.
As a buyers: If it's <30 Ω, it's probably suitable for portable audio. 16 Ω is very common and a practical solution. >30 Ω may be driven loud enough, but sound will suffer from an underpowered driver. Remember that this is a spectrum, and the difference is very gradual. I'm driving 60 Ω IEMs with a PMP, and it still sounds pretty damn good.
[size=small]Sensitivity:[/size] [size=x-small]The Loud Standard[/size]
The actual volume of earphones is a direct measurement rather than a derived figure. Volume or SPL is measured in dB and in perfectly manufactured equipment will gradually and logarithmically increase. So it requires an increasing amount of increase to achieve the same perceived level of increase (anyone got a better way of stating this?). For SPL to actually benefit customers, a standard has been defined. 1,000 Hz sound and 1 V_rms. Louder is not better. Quieter is not better. 100 dB typically preferred, but only because of convenience and not sound quality.
As a buyer: Sensitivity has little to do with quality, but is a convenience for customers. If you're using an PMP, you'll look for higher sensitivity models with lower impedance. You might also opt for a lower sensitivity and low impedance if your PMP is particulary weak. High sensitivity/impedance headphones require amping in order to efficiently drive the headphones while maintaining healthy volume levels.
[size=x-small]Final Note[/size]:
More experienced audiophiles and technicians will find some information lacking. While I encourage those users to post what kind of information would help them (i.e. how much x units of THD affect sound and the context of THD), I will not add these sections to the original post. This thread is simply a wading pool for newer users and not the end-all, tell-all of technical guides. My goal is to enlighten spec-freaks. You can't compare a car to a train. Both can go the same speed, but clearly have different roles and unique strengths. The same applies to earphones. If all of them were the same design, then specs would matter a great deal.
But they are not.
If anyone's got the time (and experience) to talk about BA, Dynamic, Electrostatic, and Planar Magnetic, it'd be a great help. Not which is better, but objective strengths and weaknesses so readers can decide "is this a well made BA?" rather than "is this BA better than this Dynamic?"
I will try to explain the numbers based on my sophomoric knowledge of circuits and helpful articles regarding the science. If you discover my findings inaccurate or misleading, PLEASE highlight them. The last thing I want is for newcomers to leave with the wrong overall impression or a misinterpretation of the significance (or insignificance) of the information. I hope you receive a context to help you decide the importance of manufacturer assessed technical data. The data may or may not affect your listening experience, I'm not trying to tell you it does or doesn't.
Although these specifications are facts, good sound quality is not about any single fact, or any big group of facts. It is about ALL the facts, and since we can never know EVERY SINGLE THING about anything, we cannot reduce sound quality to figures and graphs. Further, even the most accurate testing equipment and well-designed experiments can never replace a human ear. YOU are the judge and you can't selectively evaluate a lone aspect with your ear, so the readings from a microphone or meter tuned can be turned on their heads.
[size=small]Frequency Response:[/size] [size=x-small]Ranged and Charted[/size]
Wikipedia defines frequency response as "the magnitude of the system's response, measured in decibels (dB), and the phase, measured in radians, versus frequency."
Remember that earphones reproduce music. The most accurate earphones recreate audio as it was intended by producers. Many companies deliberately design their product to emphasize certain regions over others. This results in the louder bass, mids, or treble, a colored sound. Manufacture defects are ubiquitous, but should be understated. Undesirable distortion plays a much smaller role than unwanted color (both from defects). For a certain sound signature, distortion and color are required, and the former could play a larger role if the developers choose so.
HD800: 6–51,000 Hz (-10 dB); 14–44,100 Hz (-3 dB)
The frequency response threshold lists audible frequencies. We can hear from 20 Hz to 20,000 Hz, but the cutoff at either end is variable and dependent on the person. I personally have never heard audio below 35 Hz (may be a deficiency in my equipment) and can distinguish audio as high as 22,000 Hz. Sennheisers specs reveal that seemingly concrete FR is actually dynamic and dependent on volume. The many electrical properties of the many components within the HD800 respond differently to different testing conditions. The lower volume has a wider range, but this is no comment on how it sounds at different volumes. Again, when you see a figure, you must not treat it as a thumbnail for the whole image. It's merely the file-size, to use an analogy.
As a buyer: Don't worry about this too much. Most earphones can at least cover the entire audible range. Harmonics are an important aspect of music, but are handled different by the earphones themselves, but this isn't addressed by a FR graph.
Seeing this graph may raise some questions. Hills and valleys for the HD800 are frequent and pronounced. Many consider +/-3 dB an audible difference, but I HIGHLY recommend you download SineGen and explore how much 1, 3, and 5 dB increase the volume. You'll see that these peaks in amplitude are no small variations. But a flat response graph cannot be determined for every person. HD800 reviews typically stress neutrality. Our ears do not interpret sound linearly. In fact, in graphical terms, leftward frequencies need to be stronger to sound as loud as those on the right. People perceive higher frequency sounds as louder, even if they measure the volume. Further, the shape and travel of our ear canals were naturally selected for survival, not audiophilic enjoyment. We hear best in the range of vocal frequencies (mids). The most common type of FR graph is one derived on microphone + algorithm correction. A microphone intended to test audio equipment will yield a FR graph and each point will be shifted to account for resonance in ear canals. A human generated FR graph will be subjective, so each person's graph differs. Almost always, the former type will be provided.
As a buyer: you are probably a noob like me and should ignore this graph. Unless you know why it's important, don't treat it as a deal-breaker/maker.
More info you might not care about: the bode magnitude plot is most commonly used to represent FR. Head-Room uses reference audio equipment and attempts to correct for canal-derived differences. Educate yourself at Wikipedia or Carnegie Mellon's brief MatLab tutorial.
[size=small]Impedance:[/size] [size=x-small]A Measure In Ohms[/size]
The Ohm (Ω) used in alternating current represents impedance (Z). Impedance relates resistance (R), inductive reactance (X_L), and capacitive reactance (X_C). Due to my lowly high-school physics experience, I'm not too well-versed, so please look here and here for further information. Although, if you weren't classically educated in circuits, you might not have too much luck.
As far as I know, V = IR applied to AC yields V = IZ. Thus, a high impedance requires higher voltage given the same current. At the same volume from a source, a higher impedance driver will not travel as far. In regard to sound quality, if the current demand is not met, then bass and impact should suffer significantly. The driver will not travel as far for higher frequency, so highs will not sound as degraded as lows. Sensitivity is more relevant to volume and will be discussed later.
As a buyers: If it's <30 Ω, it's probably suitable for portable audio. 16 Ω is very common and a practical solution. >30 Ω may be driven loud enough, but sound will suffer from an underpowered driver. Remember that this is a spectrum, and the difference is very gradual. I'm driving 60 Ω IEMs with a PMP, and it still sounds pretty damn good.
[size=small]Sensitivity:[/size] [size=x-small]The Loud Standard[/size]
The actual volume of earphones is a direct measurement rather than a derived figure. Volume or SPL is measured in dB and in perfectly manufactured equipment will gradually and logarithmically increase. So it requires an increasing amount of increase to achieve the same perceived level of increase (anyone got a better way of stating this?). For SPL to actually benefit customers, a standard has been defined. 1,000 Hz sound and 1 V_rms. Louder is not better. Quieter is not better. 100 dB typically preferred, but only because of convenience and not sound quality.
As a buyer: Sensitivity has little to do with quality, but is a convenience for customers. If you're using an PMP, you'll look for higher sensitivity models with lower impedance. You might also opt for a lower sensitivity and low impedance if your PMP is particulary weak. High sensitivity/impedance headphones require amping in order to efficiently drive the headphones while maintaining healthy volume levels.
[size=x-small]Final Note[/size]:
More experienced audiophiles and technicians will find some information lacking. While I encourage those users to post what kind of information would help them (i.e. how much x units of THD affect sound and the context of THD), I will not add these sections to the original post. This thread is simply a wading pool for newer users and not the end-all, tell-all of technical guides. My goal is to enlighten spec-freaks. You can't compare a car to a train. Both can go the same speed, but clearly have different roles and unique strengths. The same applies to earphones. If all of them were the same design, then specs would matter a great deal.
But they are not.
If anyone's got the time (and experience) to talk about BA, Dynamic, Electrostatic, and Planar Magnetic, it'd be a great help. Not which is better, but objective strengths and weaknesses so readers can decide "is this a well made BA?" rather than "is this BA better than this Dynamic?"
