Results 1 to 12 of 12

Thread: Importance of Phase, Time Alignment, and other Musings

  1. #1
    Administrator Mr. Widget's Avatar
    Join Date
    Apr 2003
    Location
    San Francisco
    Posts
    11,213

    Importance of Phase, Time Alignment, and other Musings

    I had today off and the house was empty so I played in audio land.

    I wanted to compare a technically very nearly "perfect" loudspeaker and my Project Widgets which are not phase coherent, time aligned, low diffraction, or correct in any of the modern "critical" ways if one is to believe all they read.

    I brought home a pair of Meyer Sound Acheron Designer speakers. These are 2-way self powered speakers with built-in DSP phase, time, and frequency correction. They are designed to be used with one of several Meyer Sound subwoofers. These speakers were designed for use in high end cinema mixing and screening rooms, a modified version is used in the Meyer Sound Bluehorn system which is arguably one of the most accurate speakers currently available.

    The Acheron Designer speaker has a 4" aluminum domed compression driver with a 1.5" exit on and exceptionally well behaved 50° by 80° horn. Below it is a 4" voicecoiled 12" high power woofer.

    https://meyersound.com/download/ache...=58b9e6f0d9b9a

    I placed the Acheron Designers on a pair of small project boxes to elevate them off the floor. I didn't feel like lugging home the matching 18" subwoofers so I ran them full spectrum. They extend into lower 40s without a sub.

    Listening:

    The imaging was incredibly solid. I attribute this to the controlled off axis response as well as their being phase/time coherent. Even though there is no tweeter, the DSP is able to extend the 4" driver nicely and there is no hard sound that is often present when you push the output of a driver running well up into its breakup mode range. The sound was very natural and effortless. I did miss the bottom octave plus, but that is not fair since these are designed to always be paired with a proper sub. The image is a bit more forward than I am used to and holographic in a headphone sort of way... that may be due to placement though. I set the Meyers outboard of the resident Project Widgets and I didn't feel like rearranging the furniture.

    I was concerned that the built in electronics and DSP might suppress the transparency that I appreciate with my all analog system, but I was impressed with the inner detail and the stage depth. Years ago when using the second generation DEQX preamp with my Project Widgets, I was never able to get much stage depth out of them. Moving over to very high quality analog, the depth and additional inner detail came alive. The Meyer Acherons were very surprisingly good in this area.

    All that said, when I switched back to the Project Widgets, the slightly more diffuse image with a bit more inner detail, and slightly darker sound was really appreciated. Am I biased because I am familiar with their sound? Do they sound the way they do because I built them to my ear? Maybe. For me I really preferred the sound of The Project Widgets over the sound of the technically superior $35K a pair speakers. (List price including an X400C sub per side.)


    Widget
    .
    Attached Images Attached Images  

  2. #2
    Administrator Mr. Widget's Avatar
    Join Date
    Apr 2003
    Location
    San Francisco
    Posts
    11,213
    The benefits of DSP, phase correction, and less drivers is pretty evident in the frequency plots. These plots are both taken at 5' on axis of each speaker. The digitally corrected Acheron Designer is quite flat with most of the pronounced deviations from flat being caused by floor bounce.

    The driver interaction with obvious high frequency comb filtering around the high frequency crossover and phase issues around the 750Hz crossover points are visible in these 1/6th octave plots.


    Widget
    .
    Attached Images Attached Images   

  3. #3
    Administrator Mr. Widget's Avatar
    Join Date
    Apr 2003
    Location
    San Francisco
    Posts
    11,213
    I made these measurements at the listening position. In these plots I have taken 8 measurements within a 12"-15" cloud where the listener sits.

    There is a plot showing all 8 measurements for each speaker and then the average. I smoothed these measurement to 1/3 octave to better represent what we actually hear and also to make the differences in the multi position measurements easier to see.

    You can clearly see that the Acheron's phase corrected behavior is more stable throughout the space and retains a curve that is closer to the 5' on axis curve.

    Widget
    .
    Attached Images Attached Images     

  4. #4
    Administrator Mr. Widget's Avatar
    Join Date
    Apr 2003
    Location
    San Francisco
    Posts
    11,213
    My take away is that while I can certainly make a case for a simple two-way that is time and phase corrected and I can "prove" it is the better solution by showing plots and graphs, at the end of the day, I choose the speaker that looks a bit messier and doesn't measure as well. To me it sounds better.

    Next I would like to take a modern DEQX to the Project Widgets and see if it is possible to have the best of both worlds. This appears to be where GT has gone in his exploration.
    To be continued...


    Widget

  5. #5
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,986
    Hi Widget,

    Your posts are very well prepared and it must have taken some time to prepare.

    I appreciate your subjective assessments. The human ear is a far more powerful tool that can discriminate a huge amount of data not yet visible to measurements we are currently able to perform.

    My only thoughts are:

    1. The driver locations of The Widget and the Meyer system are different relative to the x,y position with the floor and the ceiling. My view is this accounts for the deviation on these measurements from your first group of measurements.

    2. Based on my response above l would like to understand what your meaning is of phase corrected response of the Meyer? Neither loudspeaker have been corrected for room interactions. I would understand your comments on phase correction if this was the case.

    My own assessment of phase correction is in the case of vertical polar plots studies of crossover region where there is no non linear disruption in the crossover region.

    I enjoyed really your posts and look to your response.

  6. #6
    Senior Member spkrman57's Avatar
    Join Date
    Feb 2005
    Location
    Columbus, Ohio
    Posts
    2,005

    I've always been a 2-way fan

    Quote Originally Posted by Mr. Widget View Post
    My take away is that while I can certainly make a case for a simple two-way that is time and phase corrected and I can "prove" it is the better solution by showing plots and graphs, at the end of the day, I choose the speaker that looks a bit messier and doesn't measure as well. To me it sounds better.

    Next I would like to take a modern DEQX to the Project Widgets and see if it is possible to have the best of both worlds. This appears to be where GT has gone in his exploration.
    To be continued...


    Widget
    Less things to get complicated.

    Thanks for sharing Widget!

    Ron sends...
    JBL Pro for home use!

  7. #7
    Administrator Robh3606's Avatar
    Join Date
    Apr 2003
    Location
    Rocinante
    Posts
    8,843
    Hey Widget!

    Thanks for posting. Looking at you in room reminded me of the Harmon curve:

    http://www.audioheritage.org/vbullet...n-Target-Curve

    I have never had the opportunity to experience a time aligned speaker except maybe my old 811C center. It was uncannily good on vocals but panning left or rignt to systems that were not time aligned nothing fell apart.

    I don't have a set 2 way/3 way/ 4 way preference it's all in the implementation although a 2 way is so much easier with just a single crossover point. Tuff finding the right driver set.

    All the technology improvements available seem to be secondary to our preferences. Most of what's going on with phase and time are icing on the cake and subtle. I think if the changes were more pronounced it may swing us the other way.

    I tried Dirac live and it sucked the life out of my HT. I should revisit it with a new prepro and updated Dirac. Looking forward on what you find using DEXQ

    "All that said, when I switched back to the Project Widgets, the slightly more diffuse image with a bit more inner detail, and slightly darker sound was really appreciated. Am I biased because I am familiar with their sound? Do they sound the way they do because I built them to my ear? Maybe."

    I find that interesting I had a similar experience changing from 2344 to the PTH1010. The 2344 imaging was pin point vs the PTH1010 which was less precise but sounded more natural. I preferred the PTH1010 and still have them in my system.

    Have Fun!

    Rob
    "I could be arguing in my spare time"

  8. #8
    Administrator Mr. Widget's Avatar
    Join Date
    Apr 2003
    Location
    San Francisco
    Posts
    11,213
    Quote Originally Posted by Ian Mackenzie View Post
    Hi Widget,

    Your posts are very well prepared and it must have taken some time to prepare.

    I appreciate your subjective assessments. The human ear is a far more powerful tool that can discriminate a huge amount of data not yet visible to measurements we are currently able to perform.

    My only thoughts are:

    1. The driver locations of The Widget and the Meyer system are different relative to the x,y position with the floor and the ceiling. My view is this accounts for the deviation on these measurements from your first group of measurements.

    2. Based on my response above l would like to understand what your meaning is of phase corrected response of the Meyer? Neither loudspeaker have been corrected for room interactions. I would understand your comments on phase correction if this was the case.

    My own assessment of phase correction is in the case of vertical polar plots studies of crossover region where there is no non linear disruption in the crossover region.

    I enjoyed really your posts and look to your response.
    The point of the measurements is not to show an idealized quasi anechoic measurement of the speakers. Simply raising the Meyer Acheron Designer speaker off the floor to about 5' (in this 12' high room) would render a more accurate measure of the speaker, but that wasn't the point. I was showing how they measure at the listening position in my room. Of course there are floor and room interactions in both cases and even some degree of interaction between the two speakers being right next to each other, but this was the most convenient way to audition them.

    What I am saying is that the response in these plots below about 700Hz is showing these interactions and yes they will be different due the relative distances between the LF drivers and the floor and other boundaries. What I wanted to show was the differences between a well behaved phase corrected speaker and a multiway speaker with rather severe time incoherence.

    Quote Originally Posted by Ian Mackenzie View Post
    2. Based on my response above l would like to understand what your meaning is of phase corrected response of the Meyer? Neither loudspeaker have been corrected for room interactions. I would understand your comments on phase correction if this was the case.
    Since the Meyer Acheron Designer speakers have integrated amplifiers and DSP, they are factory corrected to a response of 38Hz-17KHz +/- 4dB. (It appears that even in my far from idealized real world measurements these specs hold.) But more importantly, these speakers have a +/- 30° phase response between 230Hz and 17KHz. That would not be possible without time/phase correction in the DSP.

    You are right, no attempt to correct for room interactions was made, but it is my supposition that since the Meyer Sound Acheron Designer is so idealized, it is less impacted by the room interaction and is easier to deploy. We use these behind the screen in our better home theaters. (And yes, in those applications we use multiple subwoofers and external room correction.) Room correction is typically from Trinnov, Datasat, Lyngdorf, or Meyer's own Galileo processors. These particular speakers will be going into a room with 4 subs and 14 surround and ceiling speakers. That will be 21 channels of room correction... but I am getting off topic.

    My over arching point is that while my measurements appear to confirm that a time/phase corrected speaker may measure better, at least to this listener, the speaker measuring more poorly sounds better. FWIW: Both systems do not suck... If I had to live with either system, I could be quite happy.


    Widget

  9. #9
    Administrator Mr. Widget's Avatar
    Join Date
    Apr 2003
    Location
    San Francisco
    Posts
    11,213

    Love of 2-way

    Quote Originally Posted by spkrman57 View Post
    Less things to get complicated.
    Intellectually a two-way is definitely my first choice. Unfortunately I have yet to perfect one or find one I like better than these more complicated systems.


    Widget

  10. #10
    Administrator Mr. Widget's Avatar
    Join Date
    Apr 2003
    Location
    San Francisco
    Posts
    11,213
    Quote Originally Posted by Robh3606 View Post
    Looking at you in room reminded me of the Harman curve:

    http://www.audioheritage.org/vbullet...n-Target-Curve
    Ultimately my speakers were voiced by ear. Over hundreds of hours of listening to a wide range of music. This is the balance that sounds "right" for the widest range of recordings... definitely some recordings end up sounding bass heavy, and then others sound a little thin, so I have to conclude it is about right.

    Quote Originally Posted by Robh3606 View Post
    I find that interesting I had a similar experience changing from 2344 to the PTH1010. The 2344 imaging was pin point vs the PTH1010 which was less precise but sounded more natural. I preferred the PTH1010 and still have them in my system.
    Exactly. Listening to the Meyer Sound Acheron Designers, the imaging is locked in and spooky cool, but it doesn't sound natural. With the Project Widgets a hi-hat will be firmly planted in space as will a kick drum, or a sax solo... a singer will typically be centered in the image, but there is more space around them.

    We've talked and I know you listen to a lot of live music. Live music does not "image" like a studio album. To me having a little more air sounds more realistic and more pleasing.


    Widget

  11. #11
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,986
    Quote Originally Posted by Mr. Widget View Post
    The point of the measurements is not to show an idealized quasi anechoic measurement of the speakers. Simply raising the Meyer Acheron Designer speaker off the floor to about 5' (in this 12' high room) would render a more accurate measure of the speaker, but that wasn't the point. I was showing how they measure at the listening position in my room. Of course there are floor and room interactions in both cases and even some degree of interaction between the two speakers being right next to each other, but this was the most convenient way to audition them.

    What I am saying is that the response in these plots below about 700Hz is showing these interactions and yes they will be different due the relative distances between the LF drivers and the floor and other boundaries. What I wanted to show was the differences between a well behaved phase corrected speaker and a multiway speaker with rather severe time incoherence.

    Since the Meyer Acheron Designer speakers have integrated amplifiers and DSP, they are factory corrected to a response of 38Hz-17KHz +/- 4dB. (It appears that even in my far from idealized real world measurements these specs hold.) But more importantly, these speakers have a +/- 30° phase response between 230Hz and 17KHz. That would not be possible without time/phase correction in the DSP.

    My over arching point is that while my measurements appear to confirm that a time/phase corrected speaker may measure better, at least to this listener, the speaker measuring more poorly sounds better. FWIW: Both systems do not suck... If I had to live with either system, I could be quite happy.


    Widget
    Thank you for your reply.

    I don’t believe your point is technically conclusive on the evidence you have provided. I accept your subjective position. Each to their own. There is no absolute right or wrong. It’s about what a particular listener prefers.

    The reason is you have not disclosed if your Widget system was designed for a flat response in the first place or not. My guess is you are using the native response of the Tad 4003 on the TH 4003 cloned horn without any passive EQ.

    l could design a passive or analogue active response with your drivers so the response is indistinguishable from the Meyer loudspeaker on axis. Phase and delay can be designed with multi order Bessel Allpass filters in the analogue active domain ref Linkwitz. I am happy to discuss this off line.

    If both systems were then measured at the same x y z position then the amplitude response measurements would be more meaningful.

    What you could do is a “minimum phase transformation” of your Widget” system and compare it with the Meyer system. I am sceptical of marketing claims that take tech jargon out of context. Acoustics as you know is about solving specific problems for a client. Not making ambiguous claims about one technology or another.

    Scientistic research in psycho acoustics does not support the audibility of phase shift in a loudspeaker drivers pass band. The exception being deliberate intervention to augment the LF response of a low frequency system which can cause audible “group” delay. This is where FIR digital dsp has the advantage. But that advantage has specific applications such as the perceived audibility of phase tracking in the crossover region. As you know some of the most coveted and most expensive loudspeakers are full passive systems without any dsp intervention at all. So please do not be offended by this post.

    Another exception is the sound arrival time of two different transducers with long path length difference such as a Klipschorn which was well researched by Bell Labs about 100 years ago. That is where dsp has an advantage but l see no evidence of Klipschorn adopting dsp correction in the design.

  12. #12
    Administrator Mr. Widget's Avatar
    Join Date
    Apr 2003
    Location
    San Francisco
    Posts
    11,213
    Quote Originally Posted by Ian Mackenzie View Post
    The reason is you have not disclosed if your Widget system was designed for a flat response in the first place or not. My guess is you are using the native response of the Tad 4003 on the TH 4003 cloned horn without any passive EQ.
    There is no EQ in the system other than an Ashly DSP unit in front of the amp for the Sub1500. The Ashly is after the Pass Labs analog crossover.

    The TAD TH-4003/TD-4003 is extremely flat (+/- <2dB) when measured quasi anechoically between 800Hz and 13KHz, but in this application I have a -6dB pad on the 4003 combo and a passive 18db/octave crossover at 12KHz. This makes a real mess of the response, but to my ears it sounds better than deleting the tweeter or any of the other crossover frequencies or slopes I have tried.

    Quote Originally Posted by Ian Mackenzie View Post
    If both systems were then measured at the same x y z position then the amplitude response measurements would be more meaningful.
    I was really not attempting to compare the frequency responses of the two systems. I was more interested in showing how a digitally corrected system has a more consistent pattern in the measured cloud than the time/phase incoherent speaker. If my goal was to take frequency measurements of the two speakers for comparison purposes, I would need to perform ground plane measurements in a larger space. FWIW: I don't think our industry standard loudspeaker measurements tell us much about how a speaker sounds. I use a variety of measurements to design speakers or fix problems, but not so much to evaluate the sonic quality of a system.

    Quote Originally Posted by Ian Mackenzie View Post
    What you could do is a “minimum phase transformation” of your Widget” system and compare it with the Meyer system.
    I probably won't be comparing it to another system, but when I have the opportunity to borrow a DEQX or other high end DSP, I will likely explore this as I am curious if I can preserve the bits I like about these speakers and "correct" the bits I know to be wrong.

    Quote Originally Posted by Ian Mackenzie View Post
    I am skeptical of marketing claims that take tech jargon out of context. Acoustics as you know is about solving specific problems for a client. Not making ambiguous claims about one technology or another.
    I have talked at length with the engineers at Meyer Sound and know that they perform real science on a level parallel to the work done at Harman. Their statements should not be confused with the claims we regularly see in the high end world.

    Quote Originally Posted by Ian Mackenzie View Post
    So please do not be offended by this post.
    Not at all... this is all about exploration and trying to sort out why we like or don't like the way a system sounds.


    Widget

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Similar Threads

  1. Time alignment dimension
    By jbl4ever in forum Lansing Product Technical Help
    Replies: 11
    Last Post: 07-12-2018, 09:53 PM
  2. Time-alignment
    By boputnam in forum General Audio Discussion
    Replies: 15
    Last Post: 08-09-2013, 02:18 PM
  3. Need help with phase alignment
    By djnagle in forum General Audio Discussion
    Replies: 9
    Last Post: 03-22-2010, 03:25 PM
  4. Tips on time alignment?
    By hpyle in forum General Audio Discussion
    Replies: 5
    Last Post: 06-21-2007, 01:36 PM
  5. Time Alignment
    By alskinner in forum Lansing Product General Information
    Replies: 76
    Last Post: 03-15-2005, 06:01 AM

Posting Permissions

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