Results 1 to 10 of 86

Thread: An important discussion / evaluation on recent blind shootout.

Hybrid View

Previous Post Previous Post   Next Post Next Post
  1. #1
    Join Date
    Aug 2003
    Location
    USA
    Posts
    5,538

    Default An important discussion / evaluation on recent blind shootout.

    Upon receiving many emails and private messages regarding the recent GTG at Craig Chase’s home, I decided to look a bit deeper at some of the various comments and results.

    Some of the comments I read regarding the performance of our 340 SE were very far off from the thousands of actual customers and professional reviewers who either own or who have auditioned the 340 SE. Upon further investigation, I realized that Craig had purchased a pair of our 340 SE “Recert” speakers. These are used speakers but have been recertified using brand new woofers and tweeters, the one caveat being that the tweeter is a non-magnetically shielded unit compared to our stock model.

    A warning sign went off in my head --- could it be that this tweeter was somehow not performing to spec? From a technical perspective, performance should be identical but had I missed something?

    After spending a week running an intense suite of measurements, I have concluded that there is no difference between these non-shielded tweeters and our stock shielded models. Performance is identical.

    I then looked a bit deeper into the results of the listening sessions and found a glaring discrepancy – something I found quite disturbing. In every single comparison test, the speaker with higher sensitivity scored lower. In fact, it seems that the larger the discrepancy in speaker sensitivity, the greater the difference in the result scoring.

    I would expect from a “blind” listening session such as this that perhaps 50% of the higher sensitivity loudspeakers to be preferred and 50% of the lower, or at least perhaps 40% - 60% -- but DEFINITELY NOT 100%. Once again, a few warning signs went off in my head… What exactly was taking place? I know Craig was level matching but just how was he doing this?

    Further research revealed that Craig was using a passive remote switchbox. This is a speaker selector box that is placed directly in-between the loudspeakers and the amplifier. This particular switch box is made up of high-power potentiometers, which are variable resistors. In order to level match, series resistance is used to pad down the power being received by one of the two speaker pairs. The higher sensitivity speaker must have series resistance applied to the signal in order to bring the output level down enough to match the lower sensitivity speaker (which would not have resistance applied).

    This is a problem and a “No-No” in the audio world. Adding series resistance to a loudspeaker completely changes the character of the loudspeaker’s crossover and complex impedance. It can be very dramatic as I will detail. A loudspeaker crossover is designed specifically to “see” a very specific impedance load, the load presented by the drivers. Adding series resistance changes that complex impedance load and the crossover no longer functions as it has been designed to. If we have an 8ohm rated loudspeaker and we add 4ohms of resistance, the loudspeaker now becomes a 12ohm loudspeaker yet the crossover components have been optimized for an 8ohm load.

    Generally speaking, and this depends on the design of the crossover, the more complex the crossover the worse the effects. As an example, if we have a 4 ohm loudspeaker, whereby 4 ohms of resistance have been added – in order to keep all things being equal, all inductor values would have to be doubled and capacitor values halved to keep things consistent with the addition of the series resistance. While this it is not as simple as my example here, from a mathematical standpoint, it is correct.

    Furthermore, information obtained from the company that supplied Craig with the remote speaker switcher confirmed my suspicions. The speaker selector does indeed use straight series resistance in order to level match speakers…

    I do not believe that any of this was done knowingly and I certainly would not blame anyone (especially Craig) however, this was most certainly not an “all things being equal” blind listening session. In every case, the higher sensitivity loudspeaker was at a definite disadvantage, it was being handicapped (unknowingly, but nevertheless - handicapped) and both the results and information we have obtained regarding how the speaker switch box is built confirm this.

    I have personally conducted many blind A/B tests (on a professional level and for myself personally). There is only one way to properly level match loudspeakers (and even this can be argued by some as still not infallible) – level matching must always be performed at the pre-amp level – never at the output of an amplifier (high-level). In the manner I mentioned, the crossover of the loudspeaker and the amplifier itself are seeing the correct impedance load. I truly hate to say this, but I feel from a technical standpoint I must – the scoring / results from that listening session are invalid (at least where A/B testing was performed).

    I have attached several illustrations of exactly how series resistance will affect the CMT-340 SE crossover. To be as precise as possible, the electrical graphs below were taken from a random 340 SE sample, in the middle of a production run. Wires were run from the tweeter and woofer to the outside of the cabinet through the port tube so they can be measured with reference standard equipment. I used several resistor values to illustrate the effect. It is also important to note that the resistors I used are good quality wire-wound models, which are used in the vast majority of quality crossovers. These resistors offer much lower inductance and distortion than any variable resistor I know of. In other words, signal degradation from the passive speaker selector will be worse. And finally, all aspects of crossover performance are affected by series resistance, frequency response, phase, filter slopes and Q. Imagine running 1 set of loudspeakers with speaker cable that presents a resistive load of 0.1 ohm (typical for decent cable) and another set of speakers off of cable that presents a resistive load of 4 ohms?

    I somewhat blame myself for not stepping in and preventing this. However, as many of you know – I have much going on in my personal life and while I knew that Craig was performing some form of blind listening session, I did not conduct my typical due diligence and I regret this. In the past, Craig had always called and asked our permission to use our products for such an event. It did not happen this time, which might have prompted me to do a bit more research, but I admit, I should have been more on the ball. I know there will be a TON of mudslinging from this but you can not blame Craig --- Craig hosted a wonderful event and truly put himself “out there” for the benefit of the industry.

    Test Results are precisely conclusive. Winner in Red, higher sensitivity speaker in Blue.

    1. NHT Classic 2 86dB vs. Swan D2.1SE: 85dB
    2. PSB Image B25 89dB vs. SVS SCS-01 87dB
    3. Ascend 340SE 91dB vs. Acculine A1 87dB
    4. AV123 xl-s 87dB vs. Ascend Sierra 86.5 dB
    5. Dana Model 630 86dB vs. B&W 805S 89dB
    6. Swan D2.1 SE 85dB vs. Ascend Sierra 86.5dB


    Measurements:

    To effectively match the sensitivity of the Acculine A1, I estimate that approximately 4 ohms of resistance was added inline to the 340 SE, resulting in a drop of about 3-4dB. As you can see from the electrical response graph, as little as 2 ohms of added resistance will result in rather dramatic changes.

    The results of this are clearly evident on the in-room frequency response graph of the 340 SE, comparing stock configuration to one with 4 ohms of inline resistance applied. Notice the dramatic difference in the 100Hz – 800Hz range. This range is now subdued by about 5dB compared to the 1 kHz – 3 kHz range (nearly 1/4 the output), whereas without inline resistance, the ranges are of equal amplitude. This is considerable and will have a dramatic affect on the balance of vocals and midrange instruments. In addition highs become subdued in comparison to the lower frequency range (< 100Hz) and subdued when compared to the 1 kHz – 3 kHz range. This will result in the speaker sounding a bit dull and lifeless, lacking detail. All closely matching the 340SE comments from the listeners at the get together…

    I must detail an important point, the facts I presented here hold true for any of the loudspeakers that had to be padded down to match a lower sensitivity speaker. I picked the 340 SE as an example simply because it is our product and easy for me to modify to accommodate this type of tests. While inline resistance will affect different speakers in different ways, the results will always be that the speaker is no longer performing to specification. I hope many of you will find the information I posted helpful and, perhaps, influence much needed changes for any future listening sessions…
    Attached Images Attached Images
    .
    .
    .
    Good Sound To You!

    David Fabrikant
    www.ascendacoustics.com

  2. #2
    Join Date
    May 2007
    Posts
    69

    Default

    Wow...teriffically informative post, Dave! That's very interesting and I don't know if anybody else made the connection between efficiency and performance, but it's clear to see and understand why it would "handicap" the more efficient speaker. Nice catch!

    J.

  3. #3
    Join Date
    Jan 2006
    Location
    Socal
    Posts
    56

    Default

    Here's a virtual punch in the neck for Craig.

    Maybe you should be a detective Dave.

  4. #4
    Join Date
    Aug 2003
    Posts
    913

    Default

    Interesting write-up Dave - I like the graphs, it's good to see how simply adding a series load can do more than just attenuate the levels.

  5. #5
    Join Date
    Aug 2003
    Posts
    913

    Default

    Quote Originally Posted by boludaso View Post
    Here's a virtual punch in the neck for Craig.
    I think we can do without this. I expect Craig didn't know that this approach would have this kind of impact - it's not like he's out to get all speakers with higher sensitivity specs.

    I imagine one of the few ways to do a good level-matched blind test would be to have a pre-amp with multiple outputs that allows you to set a fixed attenuation on each output. I know the Meridians will let you set different attenuations on inputs, but I don't know if they will do outputs.

  6. #6
    Join Date
    May 2007
    Posts
    69

    Default

    Quote Originally Posted by BradJudy View Post
    I think we can do without this.
    Definitely agreed...Craig works his butt of doing these tests and opens his home to folks from all over. If this indeed was the culprit (if there is a culprit), I don't think that it's something that would have easily been foreseen...heck, it took Dave several days to come across his findings.

    J.

  7. #7
    Join Date
    Mar 2004
    Location
    SouthWest of Cleveland
    Posts
    1,924

    Default

    Agreed. As a guest and participant at Craig's that day, I can assure everyone that there was no intentional bias towards any speaker over another. Craig put out a monumental outlay of time, effort and expense to put this together, and had the best intentions. Too bad Dave had more urgent, personal issues to attend to, or I'm sure this would have been addressed.

    Maybe Craig can correct for this next September!
    Ed

    * Sierra-2EX's W/V2 crossover upgrade
    * (2) Rythmik F12's
    * Parasound Halo P6
    * Audio by Van Alstine DVA-M225 Monoblock Amps
    * MiniDSP 2x4HD For Sub calibration
    *World's Best Cables Canare 4S11 speaker cables

  8. #8
    Join Date
    Aug 2003
    Location
    Manhattan Beach, California
    Posts
    7,032

    Default

    Quote Originally Posted by JasonColeman View Post
    Definitely agreed...Craig works his butt of doing these tests and opens his home to folks from all over. If this indeed was the culprit (if there is a culprit), I don't think that it's something that would have easily been foreseen...heck, it took Dave several days to come across his findings.

    J.
    I don't think anybody really understood what kind of device Craig was using to do the switching, and maybe we still don't know the full story. It is generally known, at least among speaker designers/engineers, that passive resistance between the amp and speaker causes changes to how the crossover behaves.
    -curtis

  9. #9
    Join Date
    Aug 2003
    Location
    USA
    Posts
    5,538

    Default

    Quote Originally Posted by BradJudy View Post
    I think we can do without this. I expect Craig didn't know that this approach would have this kind of impact - it's not like he's out to get all speakers with higher sensitivity specs.
    Agreed... there is no blame in any of this. Nevertheless, it is important to understand that no one should use these results as a basis to make a purchase. This is a learning process and I am sure things will be corrected next time.

    Quote Originally Posted by BradJudy View Post
    I imagine one of the few ways to do a good level-matched blind test would be to have a pre-amp with multiple outputs that allows you to set a fixed attenuation on each output. I know the Meridians will let you set different attenuations on inputs, but I don't know if they will do outputs.
    This is one way or someone could build up a low-level switching device. I use a low-level switch box, 2 pre-amplifiers and a multi-channel amp.
    .
    .
    .
    Good Sound To You!

    David Fabrikant
    www.ascendacoustics.com

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •