Announcement

Collapse
No announcement yet.

DSP - Automatic EQ, Crossover and Alignment of Speaker Systems

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • DSP - Automatic EQ, Crossover and Alignment of Speaker Systems

    * Ops - bad link
    Last edited by Sydney; 03-29-2017, 11:50 AM.

  • Squidspeak
    replied
    Most newer Pre/Pro units (such as the Marantz unit I may buy) has auto EQ functions. Also I use the Berh DCX 2496 for a 3-way active set-up, thinking of
    ​upgradind (if is an upgrade) to the DBX DriveRack unit that has auto EQ. Use is home theatre/music. My question is what you guys are discussing is for
    design or actual use or both?

    Leave a comment:


  • ernperkins
    replied
    I've played around with AutoEQ in SigmaStudio and think it has a lot of potential. It does have a few warts you have to work around, but in general I think it can be very similar to working in PCD. You can import a frequency or impulse response (impedance responses no longer needed) , set a target curve, manually add whatever EQ and crossover you want and see the predicted response. Sound familiar? Yes, you can let it do some or all of the work, but that's up to the user. I also think that passive crossover designers using this tool have a big advantage - they understand acoustical versus electrical roll off, off axis response, minimum phase responses, acoustic offsets and a host of other topics needed for successful crossover development, active or passive.

    Leave a comment:


  • dcibel
    replied
    Originally posted by neildavis View Post

    Thanks--looks interesting. I'll save it for a slow day

    ​I wanted to add an optimizer algorithm to Active Speaker Designer and I've got a folder with articles and links, but haven't gotten serious about it yet.
    It's way over my head, so I hope you can make some sense of it.

    Leave a comment:


  • neildavis
    replied
    Originally posted by dcibel View Post
    Neil,
    I thought this might help you. Chapter 13 of the SoundEasy manual details the algorithm used to optimize system response. See page 9.
    Thanks--looks interesting. I'll save it for a slow day

    ​I wanted to add an optimizer algorithm to Active Speaker Designer and I've got a folder with articles and links, but haven't gotten serious about it yet.

    Leave a comment:


  • dcibel
    replied
    Originally posted by neildavis View Post
    I wish they provided a few more details and some code .
    Neil,
    I thought this might help you. Chapter 13 of the SoundEasy manual details the algorithm used to optimize system response. See page 9.
    http://www.bodziosoftware.com.au/Chapter_13.zip

    Leave a comment:


  • neildavis
    replied
    Originally posted by jclin4 View Post
    Here's a DIY effort to accomplish the same thing I think, but perhaps the extent of optimization and automation may be different:
    ​That approach uses an FIR filter, which is straightforward to design from a measurement file. It's a good approach for a Raspberry Pi which has plenty of memory, but the number of resources required make it unsuitable for a DSP chip like the ADAU1701. The approach in the article uses a sequence of 2-pole FIR filters (biquads), which are easily implemented in the ADAU1701. The algorithm to find the "optimal" set of parameters for each filter is much more involved for IIR filters as it requires a search, employing a "cost" function. The article basically describes a type of curve-fitting algorithm in which the filter parameters are varied to find the optimal "fit" for a given number of filters. It's a good approach for the ADAU1701, but I wish they provided a few more details and some code .

    Leave a comment:


  • jclin4
    replied
    Here's a DIY effort to accomplish the same thing I think, but perhaps the extent of optimization and automation may be different: http://techtalk.parts-express.com/fo...-fir-crossover

    Leave a comment:


  • fpitas
    replied
    I'm guessing the idea is to accomplish the job faster with less skilled engineers.

    Leave a comment:


  • dcibel
    replied
    Putting the article back in this thread:
    http://www.almainternational.org/yah....106170930.pdf

    Similar results can be obtained without Sigmastudio, using a DSP and software such as REW or Omnimic that can simulate the PEQ bands. Simply take your acoustic measurements, adjust the response using the PEQ bands in the software, then enter the resulting data into the DSP. I think REW will even spit out the biquads for a minidsp so all you have to do is copy and paste. IMO the "automatic" process is irrelevant when using a full-featured DSP, since adjusting the response is a fairly quick and painless task and really the whole fun of the device. That's like buying a BMW M5 and then getting someone else to drive it for you.

    Leave a comment:


  • neildavis
    replied
    Originally posted by charlielaub View Post
    DSP cannot fix certain problems that are inherent in a loudspeaker's design, most notably the variation in the radiation pattern and power response and any non-linearity in the system (e.g. IMD of the drivers, etc.).
    This is true, but that's not what the article is about

    Automating the EQ/voicing process does not change anything.
    ​The article describes an "optimizer" algorithm that has been integrated into SigmaStudio. If you are trying to figure out which filters you need to achieve a certain amplitude response, it is a very convenient and useful tool. I don't know how well this particular tool works, but TI had a similar tool around 10 years ago for the TAS3004 chip--it was called Automatic Loudspeaker Equalization (ALE). I played around with that program and was impressed with it, but it sometimes had some problems with converging on a solution.

    Leave a comment:


  • fpitas
    replied
    Maybe intended for manufacturers of low-cost high-volume "box store" speakers, where the development time is far more expensive than the actual components.

    Leave a comment:


  • charlielaub
    replied
    DSP cannot fix certain problems that are inherent in a loudspeaker's design, most notably the variation in the radiation pattern and power response and any non-linearity in the system (e.g. IMD of the drivers, etc.). The expression that "you can put lipstick on a pig, but it's still a pig" applies here. Automating the EQ/voicing process does not change anything.


    Leave a comment:

Working...
X