Here are some results of various blind tests. I won't vouch that they were done correctly as I haven't scrutinised the methodology to that extent.
Results are organised into two groups - those that indicated a difference and those that didn't.
Please keep in mind two things. Firstly, a large sample size is needed for an accurate result and secondly, both individual and group results could be considered. Typical results show no more accuracy than random chance, but some individuals may be quite accurate.
Results showing no difference
David Clarke Amplifier comparison
Conducted by: David Clarke
Components tested: 5 amplifiers including a low cost Pioneer Receiver and Audio labs tube amp
Number of participants: 25 with a similar number of "believers" and "skeptics"
Level matched? Yes
Number of tests: Refer writeup
Results posted: view here > (PDF file)
Comments:
This test was conducted by respected Engineer David Clarke of DLC Design, who played a major role in the development of the ABX comparator. No audible difference was revealed in the results. This test was conducted in the 80s and apparently caused a stir at the time. In response, John Atkinson of Stereophile conducted a larger blind test set up in a similar way. This subsequent test did suggest differences.
Ozmillsy preamp shootout - Lightspeed vs ME24 pre
Conducted by: SNA member Ozmillsy
Components tested: Lightspeed attenuator & ME24 preamp
Number of participants: 4
Level matched? Yes
Number of tests: 8 rounds of 3 samples
Results posted: view here >
Comments:
Each round had one of each, then a third sanity check where either was played. Each marked a preference, with the sanity check to see if they could pick it again. Results on the sanity check were just below 50% despite comments that each felt sure they could hear a difference.
Budget vs high end system comparison
Conducted by: Matrix Hifi
Components tested: Two complete systems sharing the same speakers
Number of participants: 38
Level matched? Yes
Number of tests: Unknown
Results posted: view here >
Comments:
The two systems shared a pair of ATC speaker with manual switching. Participants chose which system they preferred. 14 chose system A, the same number chose B and the remaining 10 admitted they couldn't pick a difference.
Monster vs Opus cables
Conducted by: ChrisWiggles & Mike L
Components tested: Speaker cables
Number of participants: 4
Level matched? Yes - within 0.01V
Number of tests: 8
Results posted: view here >
Comments:
Chris did the write up on the AVS forum, and noted that during the first 3 of 8 trials, he was doing the test sighted and initially felt that one was audibly superior. However, in then doing the test blind, he came to the opposite conclusion - no difference could be picked.
AC power cords
Conducted by: Secrets of Home Theater and High Fidelity & Bay Area Audiophile Society (BAAS)
Components tested: AC power cords
Number of participants: 15 (9 in one session, 6 in the other)
Level matched? Yes
Number of tests: 10
Results posted: view here >
Comments:
Answers were correct 50% of the time. Questions were asked which enabled them to determine that there was no connection between the accuracy of their answers and their opinion of their own dedication to audio, or age. There was also no link between the confidence in their answers and their accuracy. Those who were sure they heard a difference scored no better than those who were uncertain.
Cowan Audio CD player comparison
Conducted by: William Cowan & Dane Fehlberg
Components tested: CD players, $300 Sony vs a $1800 player
Number of participants: 2
Level matched? Yes
Number of tests: Unknown
Results posted: view here >
Comments:
William had the cheaper player and did the test when comtemplating an upgrade. Both were expecting to hear a difference, but both guessed which CD player was in use 50% of the time.
CD digital transports
Conducted by: StereoNet members Dritz, Luckydog & others
Components tested: CEC & Marantz, Dynaudio speakers
Number of participants: 5
Level matched? Yes
Number of tests: 10
Results posted: view here >
Comments:
Transports tested were both in the same price range and the final result was 50% accuracy.
Results showing a difference
Stereophile John Atkinson Amplifier comparison
(In response to the David Clarke blind test)
Conducted by: John Atkinson (Stereophile)
Components tested: Adcom vs VTL amplifier
Number of participants: 500
Level matched? Yes
Number of tests: Refer writeup
Results posted: view here >
Comments:
This test was conducted by John Atkinson of Stereophile in response to the David Clarke test. The results overall show a slight bias towards an audible difference, but certain individual results show that certain individuals clearly could pick a difference. It highlights the importance of looking at individual results. If not for certain individuals, the results could be subject to debate.
SNA front end comparison
Conducted by: Andrew Ward (Aslan) & Terry Jones
Components tested: Two front end systems
Number of participants: 10
Level matched? Yes
Number of tests: Refer writeup
Results posted: view here >
Comments:
This test was a front end comparison. The initial concept was to test the idea that speakers and the room make up 95% of the sound and everything else makes up the rest. So for this test the same speakers were used (Aslan Minstrals) with two front end setups. A 20k front end was compared to a low cost front end, although the CD player was actually quite old and certainly not considered budget in it's time.
Most were able to tell the front ends apart and describe the differences. It was a surprise for many of them how close the sound was.
Stereo Mojo Integrated amps
Conducted by: Stereo MOJO
Components tested: 8 integrated amps ranging from ~ US $120 - 2200
Number of participants: 38
Level matched? Set at 75 - 76 db
Number of tests: Refer writeup
Results posted: view here >
Comments:
This test is different to many blind tests as it was set up based on the assumption that there are differences. The test sample is small and amps were tested in pairs. If 4 preferred one while 3 the other, then the one with one more vote progresses to the next round. Tested this way, there will virtually always be a winner. It is interesting to note, however, that in a test such as this, the most expensive item didn't win. The surprising result is that a small cheap amplifier with very little power was preferred.
Other comparisons on the Stereo Mojo website >
June 16, 2010
June 12, 2010
How to solder spades
This applies to spades for speaker cables. Assuming your spades are simply a tube with nothing that can be crimped and you are relying on solder only to make the connection.
1. Firstly clamp it. You want aluminium in your clamp to avoid damage.
2. Apply solder to the spade first. It will need more heat than the cable and you want to make sure the solder is happily connected to it. You might even attach something to the end to stop solder going too far and getting messy. Put the iron on the underside to apply heat, feed the solder in to the inside of the tube. Fill er up (not quite).
3. Now tin the end of the cable. Again put heat to the bottom, feed the solder on to the top. I tend to touch the tip with some solder so there is a bit sitting on the tip - it tends to give better heat transfer to the cable and the solder will melt quicker.
4. If you have done all that properly, this part is now easy rather than difficult. All you do is put the tip on the spade to soften the solder inside the spade tube. As soon as it softens, you push the cable in.
5. You might like to slide some heatshrink over, and if your work is less than perfect this can save the day as well as mark + and - but if you're doing this, make sure to slide on the heatshrink before everything else, and keep it away from the heat.
1. Firstly clamp it. You want aluminium in your clamp to avoid damage.
2. Apply solder to the spade first. It will need more heat than the cable and you want to make sure the solder is happily connected to it. You might even attach something to the end to stop solder going too far and getting messy. Put the iron on the underside to apply heat, feed the solder in to the inside of the tube. Fill er up (not quite).
3. Now tin the end of the cable. Again put heat to the bottom, feed the solder on to the top. I tend to touch the tip with some solder so there is a bit sitting on the tip - it tends to give better heat transfer to the cable and the solder will melt quicker.
4. If you have done all that properly, this part is now easy rather than difficult. All you do is put the tip on the spade to soften the solder inside the spade tube. As soon as it softens, you push the cable in.
5. You might like to slide some heatshrink over, and if your work is less than perfect this can save the day as well as mark + and - but if you're doing this, make sure to slide on the heatshrink before everything else, and keep it away from the heat.
June 11, 2010
Why is blind testing such a hot topic?
Blind testing has been the focus of a lot of flame wars in the audio community. Part of the reason for this is that it cuts to the heart of the biggest divide. Audio enthusiasts are inclined to hold to either an objectivist or subjectivist position. It's not black and white, but there are many shades of grey and there will always be some that sit on the fence. But the flame wars happen between those who sit at the extreme ends.
Are you an objectivist?
An objectivist takes a scientific view. They believe that a sound system is best set up following good scientific and engineering principles. Sonic differences can be measured and explained in scientific terms. This does not necessarily mean a denial that subjective differences exist, but when it comes to audio beliefs, the objectivist believes there is an answer that can be found, even if it hasn't yet been found.
The objectivist believes in blind testing as the ideal way to compare equipment. They see it as the sensible way to remove bias and make a comparison honest and valid.
Are you a subjectivist?
A subjectivist takes a view that is more like that of a critic. They prefer to evaluate the sound of components subjectively while knowing what they are listening to. They don't necessarily deny the value of science or engineering, but they believe that there is an art to setting up a system and that it's best done by ear. They tend to see measurements as having limited value and believe that the aspects of a system that most interest them can't be measured.
The subjectivist finds blind testing unnecessary and even detrimental. They see it as a sign that one hasn't learnt to trust their ears and will tend the feel that the process obscures the result. They believe that sighted listening tests are more valid.
Why can't we all just get along?
It happens like this. Often a discussion will start on a topic like cables, or in fact anything other than speakers or room acoustics. Someone will claim that all cables sound the same and demand that a blind test should be done to prove the claims that are made. An ordinary person might expect a response like "you can do it if you like but I'd rather not." But audiophiles aren't ordinary people, but a bunch of passionate nut cases. These discussions heat up and let's face it, audio enthuasiasts can't be trusted to just get along.
I'd like to appeal to both sides of the divide.
An appeal to objectivists
There is a tendency among some to be a little obtuse and blunt. It can be perceived as very arrogant and abrasive. Many opinions in this camp are formed without doing any kind of listening test.
My appeal is this - try to cut people some slack. When you see long descriptions of the sonic differences of cables, find another thread. If you feel the need to change other people's point of view, then do it in a helpful way. Offer to host a listening test designed to reveal differences if they do exist. If you feel that you are superior to these people, then you should be warned. Some of them may hear differences in a blind test that you cannot. It could end up being a humbling experience for you. Make sure your invitation is a friendly one and if you find that unpalatable, perhaps you should re-consider the circles you are mixing in. If you can't find many others that share your view online, then there is a problem!
An appeal to subjectivists
I notice a trend with some to become a little elitist. You feel that your ears are superior because you can hear what the blind test brigage can't. You may actually have better listening skills, but that's besides the point. It is the attitude that "I'm a cut above you" which others find abrasive. "If you can't hear cable differences, you should choose another hobby!" Or you might refer to some as unbelievers with disdain. This attitude only feeds the flame wars.
Time for a truce?
The flame wars end when each side drop the attitude and accept that the other side has something to contribute to the community. It doesn't mean we have to agree or compromise views on which we are passionate. It just means we have to treat other enthusiasts with respect and conduct ourselves as we would at a social event.
Have you been demanding people take a blind test as proof? Instead, repeat after me "wooooo - saaaaaaa" and take a deep breath. For a full description of the technique, see the movie Anger Management.
In case you have wondered, I lean towards the objectivist camp but I don't take an extreme position. I believe science gets us most of the way there, but a little salt to the taste isn't a bad thing.
Are you an objectivist?
An objectivist takes a scientific view. They believe that a sound system is best set up following good scientific and engineering principles. Sonic differences can be measured and explained in scientific terms. This does not necessarily mean a denial that subjective differences exist, but when it comes to audio beliefs, the objectivist believes there is an answer that can be found, even if it hasn't yet been found.
The objectivist believes in blind testing as the ideal way to compare equipment. They see it as the sensible way to remove bias and make a comparison honest and valid.
Are you a subjectivist?
A subjectivist takes a view that is more like that of a critic. They prefer to evaluate the sound of components subjectively while knowing what they are listening to. They don't necessarily deny the value of science or engineering, but they believe that there is an art to setting up a system and that it's best done by ear. They tend to see measurements as having limited value and believe that the aspects of a system that most interest them can't be measured.
The subjectivist finds blind testing unnecessary and even detrimental. They see it as a sign that one hasn't learnt to trust their ears and will tend the feel that the process obscures the result. They believe that sighted listening tests are more valid.
Why can't we all just get along?
It happens like this. Often a discussion will start on a topic like cables, or in fact anything other than speakers or room acoustics. Someone will claim that all cables sound the same and demand that a blind test should be done to prove the claims that are made. An ordinary person might expect a response like "you can do it if you like but I'd rather not." But audiophiles aren't ordinary people, but a bunch of passionate nut cases. These discussions heat up and let's face it, audio enthuasiasts can't be trusted to just get along.
I'd like to appeal to both sides of the divide.
An appeal to objectivists
There is a tendency among some to be a little obtuse and blunt. It can be perceived as very arrogant and abrasive. Many opinions in this camp are formed without doing any kind of listening test.
My appeal is this - try to cut people some slack. When you see long descriptions of the sonic differences of cables, find another thread. If you feel the need to change other people's point of view, then do it in a helpful way. Offer to host a listening test designed to reveal differences if they do exist. If you feel that you are superior to these people, then you should be warned. Some of them may hear differences in a blind test that you cannot. It could end up being a humbling experience for you. Make sure your invitation is a friendly one and if you find that unpalatable, perhaps you should re-consider the circles you are mixing in. If you can't find many others that share your view online, then there is a problem!
An appeal to subjectivists
I notice a trend with some to become a little elitist. You feel that your ears are superior because you can hear what the blind test brigage can't. You may actually have better listening skills, but that's besides the point. It is the attitude that "I'm a cut above you" which others find abrasive. "If you can't hear cable differences, you should choose another hobby!" Or you might refer to some as unbelievers with disdain. This attitude only feeds the flame wars.
Time for a truce?
The flame wars end when each side drop the attitude and accept that the other side has something to contribute to the community. It doesn't mean we have to agree or compromise views on which we are passionate. It just means we have to treat other enthusiasts with respect and conduct ourselves as we would at a social event.
Have you been demanding people take a blind test as proof? Instead, repeat after me "wooooo - saaaaaaa" and take a deep breath. For a full description of the technique, see the movie Anger Management.
In case you have wondered, I lean towards the objectivist camp but I don't take an extreme position. I believe science gets us most of the way there, but a little salt to the taste isn't a bad thing.
Exploding a bass myth
Are big drivers evil?
First, a story. Once I walked into a local store looking for a demo. I heard a speaker with some 6.5" drivers and a number of subs. Not one of the subs sounded nearly as good as the bass from the smaller drivers. In fact, all of their subs were just plain awful. It's tempting at this point to conclude that big drivers are the problem. A small driver is tight, fast and accurate while a big heavy driver is muddly and sloppy. Kind of like comparing a truck to a luxury car. This particular myth is seductive because it's one that intuition and experience would often suggest.
Many audiophiles quickly form a belief about things like this. Human nature then steps in as once an opinion is formed, the brain tends to order any extra information according to the belief. Truth is often suppressed in order to propagate an established belief. This is often applied to religion, but in reality this works across the board. It's a way for our minds to establish stability and avoid changing opinions about everything all the time. We don't worry about floating off into space because we instinctively trust the law of gravity.
Why does it appear that smaller drivers are more accurate?
There are a few reasons for this. In particular, when the bigger driver is a subwoofer, and the smaller driver is in a fullrange speaker, we have an apples to oranges comparison. Firstly, with a subwoofer, there is more that can go wrong.
A fullrange speaker is plug and play and already has the levels balanced - it's passive. A sub on the other hand is just too easy to set the level wrong. If done by ear a person will probably set the level much higher than the mains. That tends to amplify any bass problems in the room. You will now become much more aware of room modes, since the bass is turned up too much, some of them will boom. A fullrange speaker draws less attention to this as you are hearing a flat response. The sub could easily end up 10 - 15 db louder, and then if you have a 15 db room mode peak which is very common, it really booms. Just by turning up my subs too loud, I can get them to sound like boom boxes. Yet done right I have not heard better bass at any price.
Another issue is cost. Subs use big expensive drivers. If you compare a good woofer to a cheapie sub, what do you expect? Good woofers cost ~$80 - 300 but decent sub drivers cost generally $300 - 1000. Comparing at the same price is like comparing a sedan to a bus at the same price point for ride comfort.
Also there is a difference between woofers and subwoofers. Woofers have lower excursion and have an easier task. They are designed generally for wider bandwidth and usually need less power handling. They don't do the same job and they have different compromises. Sub drivers need generally an extra octave extension, but less top end. Therefore they need to move 4x as much air as well as more power handling. This will mean a bigger VC diameter which in turn has higher inductance, but we can live with that in a sub. Good design also helps keep inductance reasonable. We can move that extra air by making the driver bigger, or increasing excursion.
Increasing SD (driver piston area), will increase efficiency. That means we need less power, less excursion and can live with a smaller voice coil. Those factors are the ones that we struggle with here, and which cause distortion to hike up. The mass goes up in proportion to the size, and the motor strength may also go up, although this will depend on if the driver is designed to have a bigger box or not. The alternative, if we buy into the driver size myth, is to increase excursion. That is a change that will cause problems. Let's suppose we believe a 6.5" driver is "faster" than an 18" driver. We want smaller drivers to do the same job as an 18" driver with 18mm p-p excursion.
The 6.5" driver has a piston area of 145 sq cm, but the 18" has 1220. That means one 6.5" driver needs 150mm excursion! It is compensating for 1/8 piston area. We could compensate by using 8x as many drivers but even there 9mm is quite high excursion for a 6.5" driver. It can be done, but we probably want to use more like 12 drivers with about 6mm xmax instead. Now that we can move as much air, we have a cost issue. A decent midbass costs around $100 - 300 for something high end. The 18" driver for similar quality would cost around $400 - 1000. So if we pick a good 18" driver even at the top end, we have to use lesser quality midbass drivers as there are so many of them. If we choose high end midbass drivers, the cost is nearly 4x as much! So if we consider this on the basis of price, we are at a big disadvantage with the smaller drivers.
What happens in reality is that the small midbass drivers are placed in a slim box that will give accurate bass at a moderate level. They are affordable and practical, but are limited in dynamics. This choice is based on domestic acceptance rather than performance, but audiophile myths and subs not properly set up have confused the issue.
If you want the ultimate in bass performance, you can't go past bigger drivers. Small drivers for bass is always a compromise. Achieving accurate bass is far more complex than most are ready to get their minds around, but in general it is true to say that size matters. Depending on how loud and how deep you want to go, bass drivers should idealy be 10 - 24" in size. You can get good quality bass from a 6.5" driver with enough extension and output for moderate music enjoyment, but you should realise that this is always a compromise and certainly not as good as it gets. All else being equal (that happens only in theory), bigger is better up to a point.
One 6.5" driver
First, a story. Once I walked into a local store looking for a demo. I heard a speaker with some 6.5" drivers and a number of subs. Not one of the subs sounded nearly as good as the bass from the smaller drivers. In fact, all of their subs were just plain awful. It's tempting at this point to conclude that big drivers are the problem. A small driver is tight, fast and accurate while a big heavy driver is muddly and sloppy. Kind of like comparing a truck to a luxury car. This particular myth is seductive because it's one that intuition and experience would often suggest.
Many audiophiles quickly form a belief about things like this. Human nature then steps in as once an opinion is formed, the brain tends to order any extra information according to the belief. Truth is often suppressed in order to propagate an established belief. This is often applied to religion, but in reality this works across the board. It's a way for our minds to establish stability and avoid changing opinions about everything all the time. We don't worry about floating off into space because we instinctively trust the law of gravity.
Why does it appear that smaller drivers are more accurate?
There are a few reasons for this. In particular, when the bigger driver is a subwoofer, and the smaller driver is in a fullrange speaker, we have an apples to oranges comparison. Firstly, with a subwoofer, there is more that can go wrong.
A fullrange speaker is plug and play and already has the levels balanced - it's passive. A sub on the other hand is just too easy to set the level wrong. If done by ear a person will probably set the level much higher than the mains. That tends to amplify any bass problems in the room. You will now become much more aware of room modes, since the bass is turned up too much, some of them will boom. A fullrange speaker draws less attention to this as you are hearing a flat response. The sub could easily end up 10 - 15 db louder, and then if you have a 15 db room mode peak which is very common, it really booms. Just by turning up my subs too loud, I can get them to sound like boom boxes. Yet done right I have not heard better bass at any price.
Another issue is cost. Subs use big expensive drivers. If you compare a good woofer to a cheapie sub, what do you expect? Good woofers cost ~$80 - 300 but decent sub drivers cost generally $300 - 1000. Comparing at the same price is like comparing a sedan to a bus at the same price point for ride comfort.
Also there is a difference between woofers and subwoofers. Woofers have lower excursion and have an easier task. They are designed generally for wider bandwidth and usually need less power handling. They don't do the same job and they have different compromises. Sub drivers need generally an extra octave extension, but less top end. Therefore they need to move 4x as much air as well as more power handling. This will mean a bigger VC diameter which in turn has higher inductance, but we can live with that in a sub. Good design also helps keep inductance reasonable. We can move that extra air by making the driver bigger, or increasing excursion.
Increasing SD (driver piston area), will increase efficiency. That means we need less power, less excursion and can live with a smaller voice coil. Those factors are the ones that we struggle with here, and which cause distortion to hike up. The mass goes up in proportion to the size, and the motor strength may also go up, although this will depend on if the driver is designed to have a bigger box or not. The alternative, if we buy into the driver size myth, is to increase excursion. That is a change that will cause problems. Let's suppose we believe a 6.5" driver is "faster" than an 18" driver. We want smaller drivers to do the same job as an 18" driver with 18mm p-p excursion.
The 6.5" driver has a piston area of 145 sq cm, but the 18" has 1220. That means one 6.5" driver needs 150mm excursion! It is compensating for 1/8 piston area. We could compensate by using 8x as many drivers but even there 9mm is quite high excursion for a 6.5" driver. It can be done, but we probably want to use more like 12 drivers with about 6mm xmax instead. Now that we can move as much air, we have a cost issue. A decent midbass costs around $100 - 300 for something high end. The 18" driver for similar quality would cost around $400 - 1000. So if we pick a good 18" driver even at the top end, we have to use lesser quality midbass drivers as there are so many of them. If we choose high end midbass drivers, the cost is nearly 4x as much! So if we consider this on the basis of price, we are at a big disadvantage with the smaller drivers.
What happens in reality is that the small midbass drivers are placed in a slim box that will give accurate bass at a moderate level. They are affordable and practical, but are limited in dynamics. This choice is based on domestic acceptance rather than performance, but audiophile myths and subs not properly set up have confused the issue.
If you want the ultimate in bass performance, you can't go past bigger drivers. Small drivers for bass is always a compromise. Achieving accurate bass is far more complex than most are ready to get their minds around, but in general it is true to say that size matters. Depending on how loud and how deep you want to go, bass drivers should idealy be 10 - 24" in size. You can get good quality bass from a 6.5" driver with enough extension and output for moderate music enjoyment, but you should realise that this is always a compromise and certainly not as good as it gets. All else being equal (that happens only in theory), bigger is better up to a point.
One 6.5" driver
Does fast bass exist?
The topic of "fast bass" tends to polarise opinions. There are two main camps here. In one, "fast bass" is a myth that doesn't exist. They reject the use of the term completely. This is an objectivist opinion. In the other, the term is accepted as a subjective one meaning "tight and accurate bass." It would be helpful if the audio community could agree on this so that we can stop arguing over the meaning of a term, or if it even exists!
Taking it literally ...
Out of interest, let's first explore a literal interpretation of the term. Let's see how fast bass really is, as measured by the speed of cone travel. What is the fastest bass we can get, as measured by the speed of cone travel. First we'll take a 6.5" midbass driver and play it at a level that audiophiles will use in a room to get "fast bass." We'll play it at a moderate level of 70db at a distance of 4m and we'll see how fast the cone moves at 60 Hz. We'll use the classic Scan Speak 6.5" midbass - 18w8545, which is a fairly expensive midbass used in many high end designs. We need 0.4w to achieve this level and the cone excursion is a mere 1mm from peak to peak. What is the speed involved? The average speed considers distance and time.
Distance is 1/1000m and time is 1/60 sec. So that is 0.001 metres in 0.016 seconds (16ms). So we have 0.06 m/s or 0.2 km/h. The average person walks 25 x faster! If we want the cone to move faster, we could turn up the volume. If we use 4x as much power, we get twice the excursion and twice the speed, but we all know that everything goes backwards from there.
What happens if we use a subwoofer? If we use the Exodus Tempest-X, which is a 15" driver and compare at the same spot with identical output, the cone moves one third as much due to it's size. It is even slower. But we want to get some speed here so we feed it 2kw of power and the excursion is now 28mm - so the speed is 28x as much. Now we've sped up to the speed of a person walking at last, but we can hardly call this fast! Let's rule out velocity as being relevant to "fast bass."
What should we do with this term "fast bass?"
We have a dictionary to define what words mean, so that we can all communicate without debating what words mean - we can get on with life. In technical fields certain terms are defined and generally accepted due to practical need. There is a need to agree on commonly used terms so that we don't get caught up in arguing over words, and get on with the discussion. I suggest the best solution is to simply accept the term "fast bass" in the way it is usually intended - as a subjective term that simply means "accurate bass." I personally don't use the term, as it becomes something of a red herring and often derails discussion. It's much easier to say "accurate." But if someone uses the term fast bass, I know what they mean. We can debate what actually constitutes fast bass, and how to achieve it, but I wish we could at least agree on the term itself!
Taking it literally ...
Out of interest, let's first explore a literal interpretation of the term. Let's see how fast bass really is, as measured by the speed of cone travel. What is the fastest bass we can get, as measured by the speed of cone travel. First we'll take a 6.5" midbass driver and play it at a level that audiophiles will use in a room to get "fast bass." We'll play it at a moderate level of 70db at a distance of 4m and we'll see how fast the cone moves at 60 Hz. We'll use the classic Scan Speak 6.5" midbass - 18w8545, which is a fairly expensive midbass used in many high end designs. We need 0.4w to achieve this level and the cone excursion is a mere 1mm from peak to peak. What is the speed involved? The average speed considers distance and time.
Distance is 1/1000m and time is 1/60 sec. So that is 0.001 metres in 0.016 seconds (16ms). So we have 0.06 m/s or 0.2 km/h. The average person walks 25 x faster! If we want the cone to move faster, we could turn up the volume. If we use 4x as much power, we get twice the excursion and twice the speed, but we all know that everything goes backwards from there.
What happens if we use a subwoofer? If we use the Exodus Tempest-X, which is a 15" driver and compare at the same spot with identical output, the cone moves one third as much due to it's size. It is even slower. But we want to get some speed here so we feed it 2kw of power and the excursion is now 28mm - so the speed is 28x as much. Now we've sped up to the speed of a person walking at last, but we can hardly call this fast! Let's rule out velocity as being relevant to "fast bass."
What should we do with this term "fast bass?"
We have a dictionary to define what words mean, so that we can all communicate without debating what words mean - we can get on with life. In technical fields certain terms are defined and generally accepted due to practical need. There is a need to agree on commonly used terms so that we don't get caught up in arguing over words, and get on with the discussion. I suggest the best solution is to simply accept the term "fast bass" in the way it is usually intended - as a subjective term that simply means "accurate bass." I personally don't use the term, as it becomes something of a red herring and often derails discussion. It's much easier to say "accurate." But if someone uses the term fast bass, I know what they mean. We can debate what actually constitutes fast bass, and how to achieve it, but I wish we could at least agree on the term itself!
Subwoofer bracing primer
How should you design bracing for a DIY subwoofer? It's very common for DIY enthusiasts to get it wrong. There are many different ways to brace a subwoofer enclosure, but the one shown here is one that will yield the best performance possible without going crazy.
What are we trying to achieve?
With a subwoofer box, the critical aspect is stiffness. In a fullrange enclosure, the challenge is different as we are also interested in damping qualities. Accelerometer measurements of fullrange enclosures show benefit from mass damping with lead, but for a sub box this merely adds weight. With a sub we simply want to make the box stiff. We can achieve this with curved walls, exotic materials, very thick walls but the cost effective way to do it is with bracing applied to reasonably thick walls.
So if we start with 18mm MDF, we might first double the thickness for the baffle, then brace internally. Each brace should touch 4 walls, and 3 intersecting sheets should be used. We then make cut outs for the driver and amp (if used in the box), and openings so that we don't have enclosed air cells.
Here is an example:
This bracing system could be improved slightly but making the cut-outs circular, although the differencs isn't critical.
Another example can be seen here of a 10" subwoofer with a downfiring driver.
A similar bracing design could be scaled up for 12" or 15" drivers. As the enclosure gets larger you may choose more bracing sheets, but as a general guide aim to break up the box with a grid no greater than 200mm.
What are we trying to achieve?
With a subwoofer box, the critical aspect is stiffness. In a fullrange enclosure, the challenge is different as we are also interested in damping qualities. Accelerometer measurements of fullrange enclosures show benefit from mass damping with lead, but for a sub box this merely adds weight. With a sub we simply want to make the box stiff. We can achieve this with curved walls, exotic materials, very thick walls but the cost effective way to do it is with bracing applied to reasonably thick walls.
So if we start with 18mm MDF, we might first double the thickness for the baffle, then brace internally. Each brace should touch 4 walls, and 3 intersecting sheets should be used. We then make cut outs for the driver and amp (if used in the box), and openings so that we don't have enclosed air cells.
Here is an example:
This bracing system could be improved slightly but making the cut-outs circular, although the differencs isn't critical.
Another example can be seen here of a 10" subwoofer with a downfiring driver.
A similar bracing design could be scaled up for 12" or 15" drivers. As the enclosure gets larger you may choose more bracing sheets, but as a general guide aim to break up the box with a grid no greater than 200mm.
June 2, 2010
What do the numbers tell you about a driver?
So you want to use a particular driver and you are wondering if it's suitable for your application. If you look at the data sheet, you will see "Thiele Small" parameters. What do the numbers mean? What can you judge about a driver based on these numbers? In this article the focus is on bass drivers.
What you can't tell ...
The parameters don't tell the whole story. They won't show how well it is built or anything about the quality or distortion performance. They also won't show if the driver has a nice flat response or if it has nasty break-up or resonances. What the parameters will do is show you if it's suitable for the box you have in mind and what kind of output and extension you can expect. This is a good starting point.
The main parameters to watch
fs
free air resonance
This indicates the low frequency limit of the driver. In general this becomes the low frequency limit in a vented box, although you can run a driver lower than fs if the box is large. Driver excursion becomes critical as this creates an extra demand. In a sealed box, the bass limit will be one octave above fs. 12db of EQ is required to get extension down to fs however, this will quadruple driver excursion and require more than 10x as much power!
Qts
total Quality factor
This has nothing to do with quality. It is a measure of the damping of a driver. A driver with a low Qts value (<0.3)>0.5) has less damping, often resulting in a peak around fs. Mid Q drivers are a popular choice for subwoofers as they are generally flexible and will suit vented and sealed designs.
High Q drivers are often used in a sealed box use as they tend to need a very big vented box. Some of the new Exodus audio drivers have a relatively high Q. While a fairly large vented box is generally required, it will be more efficient than a lower Q driver in a smaller box.
Low Q drivers allow small boxes, but their efficiency is generally lower so that high power is required. The result is often a sub that is expensive for the performance that is offered. Many commercial subs use low Q drivers to allow small size, but due to the inefficiency, very high powered amplifiers are used. Early versions of the Peerless XLS driver were very low Q and worked in a small box but were not practical for a vented design. A workable vent would not fit in the box so passive radiators were required. The cost jumps up as a result.
VAS
equivalent volume compliance - the volume of air which has the same compliance as the driver's suspension
VAS in conjunction with Qts gives an idea of the size of the box required. Generally, driver with a large VAS need a large box. If the driver has a high Qts, it needs to be even bigger.
SD
effective piston area
This is the total cone surface area that produces output. Generally part of the surround is included as part of the piston area. Drivers with very large surrounds have a lower SD.
xmax
maximum linear excursion (one way)
xmax is typically defined as the point where the force of the motor (BL) drops to 70% and is considered to be the maximum excursion which is useful. This is critical for determining the output capability of the driver, when combined with power handling and SD.
xmech
mechanical excursion limit
This is the limit of cone travel before damage, although some drivers avoid damage when pushed beyond xmech. Ideally you should design a subwoofer so that there is some kind of limiting so that there will never be enough power to push a driver to this point. This can be done by a combination of a few strategies:
Electrical power handling
This number should always be taken with a pinch of salt. Often an inflated number is given and even when accurately quoted, may not be significant. Car audio subs often have a high power handling and low xmax. This works well in their application, where they will often be placed in a small sealed box in the small space of a car which provides massive acoustic bass boost. In a home theatre subwoofer application, xmax is more critical. Where a home theatre subwoofer is used for music, thermal power handling becomes the limiting factor.
Le
Voice coil inductance
Ideally inductance should be as low as possible. In a crossover, inductors act as a low pass filter. High inductance in a driver has the same effect. If it is very high, it can limit the upper useable bandwidth. It is also said to have a deterimental impact on transient response, although this aspect is the subject of debate.
mms
Driver moving mass
Avoid the temptation to choose a driver with low mms thinking it will sound faster. Mms is chosen to give a desired fs and when considered in isolation does not show how "fast" a driver is.
BL
Force factor
This is an indicator of motor strength. Avoid the temptation to choose a driver because of high BL. Considered in isolation, this means nothing.
Sensitivity
This number is misleading. The design of the box has such a dramatic impact on the actual efficiency, that this number tells us very little. A high sensitivity driver may end up less efficient than a much lower sensitivity driver.
Parameters to ignore
Most of the other parameters don't warrant consideration, at least not in a quick analysis of a driver.
Example 1 - Peerless XLS (830500)
fs: 18 Hz
Qts: 0.2
VAS: 139 L
xmax: 12 mm
xmech: 22 mm
This driver works in a very small box. It suits a sealed box or passive radiator box. A 40L box can get decent extension, but there is no way a practical vent will work. Low tuning in a small box doesn't work. You might make it as big as 85L, which is bigger than required but at this size, you can make a workable vented box with a 4" vent. Xmax is moderate at 12mm so this driver will only need 240 - 350w. xmech is quite high at 22mm due to the large rubber surround so there is a fair safety margin. SD is quite low for a 12" driver.
Example 2 - Exodus Tempest-X2
fs: 20 Hz
Qts: 0.42
VAS: 255 L
xmax: 26 mm
PE: 1kw
The parameters show this is a very different driver. It needs a much bigger box as seen in the higher Qts and VAS. It also has more than double the excursion and very high thermal power handling. For home theatre use, a large vented box with a high power amp will yield more extension and output than the modest XLS driver.
The next step
After taking a quick look at the parameters, it's time to simulate the driver. A great deal can be learnt this way. After trying different settings, you will find that you can force fit a driver into different applications.
What you can't tell ...
The parameters don't tell the whole story. They won't show how well it is built or anything about the quality or distortion performance. They also won't show if the driver has a nice flat response or if it has nasty break-up or resonances. What the parameters will do is show you if it's suitable for the box you have in mind and what kind of output and extension you can expect. This is a good starting point.
The main parameters to watch
fs
free air resonance
This indicates the low frequency limit of the driver. In general this becomes the low frequency limit in a vented box, although you can run a driver lower than fs if the box is large. Driver excursion becomes critical as this creates an extra demand. In a sealed box, the bass limit will be one octave above fs. 12db of EQ is required to get extension down to fs however, this will quadruple driver excursion and require more than 10x as much power!
Qts
total Quality factor
This has nothing to do with quality. It is a measure of the damping of a driver. A driver with a low Qts value (<0.3)>0.5) has less damping, often resulting in a peak around fs. Mid Q drivers are a popular choice for subwoofers as they are generally flexible and will suit vented and sealed designs.
High Q drivers are often used in a sealed box use as they tend to need a very big vented box. Some of the new Exodus audio drivers have a relatively high Q. While a fairly large vented box is generally required, it will be more efficient than a lower Q driver in a smaller box.
Low Q drivers allow small boxes, but their efficiency is generally lower so that high power is required. The result is often a sub that is expensive for the performance that is offered. Many commercial subs use low Q drivers to allow small size, but due to the inefficiency, very high powered amplifiers are used. Early versions of the Peerless XLS driver were very low Q and worked in a small box but were not practical for a vented design. A workable vent would not fit in the box so passive radiators were required. The cost jumps up as a result.
VAS
equivalent volume compliance - the volume of air which has the same compliance as the driver's suspension
VAS in conjunction with Qts gives an idea of the size of the box required. Generally, driver with a large VAS need a large box. If the driver has a high Qts, it needs to be even bigger.
SD
effective piston area
This is the total cone surface area that produces output. Generally part of the surround is included as part of the piston area. Drivers with very large surrounds have a lower SD.
xmax
maximum linear excursion (one way)
xmax is typically defined as the point where the force of the motor (BL) drops to 70% and is considered to be the maximum excursion which is useful. This is critical for determining the output capability of the driver, when combined with power handling and SD.
xmech
mechanical excursion limit
This is the limit of cone travel before damage, although some drivers avoid damage when pushed beyond xmech. Ideally you should design a subwoofer so that there is some kind of limiting so that there will never be enough power to push a driver to this point. This can be done by a combination of a few strategies:
- limiting amplifier power
- sealed box where the selected volume will limit excursion to safe levels
- rumble filter in the crossover
- DSP dynamic filtering
Electrical power handling
This number should always be taken with a pinch of salt. Often an inflated number is given and even when accurately quoted, may not be significant. Car audio subs often have a high power handling and low xmax. This works well in their application, where they will often be placed in a small sealed box in the small space of a car which provides massive acoustic bass boost. In a home theatre subwoofer application, xmax is more critical. Where a home theatre subwoofer is used for music, thermal power handling becomes the limiting factor.
Le
Voice coil inductance
Ideally inductance should be as low as possible. In a crossover, inductors act as a low pass filter. High inductance in a driver has the same effect. If it is very high, it can limit the upper useable bandwidth. It is also said to have a deterimental impact on transient response, although this aspect is the subject of debate.
mms
Driver moving mass
Avoid the temptation to choose a driver with low mms thinking it will sound faster. Mms is chosen to give a desired fs and when considered in isolation does not show how "fast" a driver is.
BL
Force factor
This is an indicator of motor strength. Avoid the temptation to choose a driver because of high BL. Considered in isolation, this means nothing.
Sensitivity
This number is misleading. The design of the box has such a dramatic impact on the actual efficiency, that this number tells us very little. A high sensitivity driver may end up less efficient than a much lower sensitivity driver.
Parameters to ignore
Most of the other parameters don't warrant consideration, at least not in a quick analysis of a driver.
Example 1 - Peerless XLS (830500)
fs: 18 Hz
Qts: 0.2
VAS: 139 L
xmax: 12 mm
xmech: 22 mm
This driver works in a very small box. It suits a sealed box or passive radiator box. A 40L box can get decent extension, but there is no way a practical vent will work. Low tuning in a small box doesn't work. You might make it as big as 85L, which is bigger than required but at this size, you can make a workable vented box with a 4" vent. Xmax is moderate at 12mm so this driver will only need 240 - 350w. xmech is quite high at 22mm due to the large rubber surround so there is a fair safety margin. SD is quite low for a 12" driver.
Example 2 - Exodus Tempest-X2
fs: 20 Hz
Qts: 0.42
VAS: 255 L
xmax: 26 mm
PE: 1kw
The parameters show this is a very different driver. It needs a much bigger box as seen in the higher Qts and VAS. It also has more than double the excursion and very high thermal power handling. For home theatre use, a large vented box with a high power amp will yield more extension and output than the modest XLS driver.
The next step
After taking a quick look at the parameters, it's time to simulate the driver. A great deal can be learnt this way. After trying different settings, you will find that you can force fit a driver into different applications.
Subscribe to:
Posts (Atom)