Blog Categories

Audyssey XT32 vs. Parametric EQ: A Comparison of Manual vs Automagic Calibrations

Setting the Stage for Battle!

Earlier this year Acoustic Frontiers had the opportunity to improve a two channel and home theater system in Tucson, Arizona.  Our client owned a Polk / Onkyo system which incorporated the latest Audyssey XT32 ‘room correction’ algorithms. However he was unhappy with overall sound quality and wanted to know how it could be improved…

Enter Acoustic Frontiers. It started like most of our remote consulting engagements, with the client purchasing an XTZ Room Analyzer II Pro and our fixed price acoustic diagnosis package. This allowed him to easily take acoustic measurements and send them to us for analysis.

This particular room had many issues. The major one was very poor bass response with large peaks and dips and significant ringing at modal frequencies, a consequence of the brick walls in the room. This had led the client to Audyssey XT32, a well reviewed ‘room correction’ algorithm embedded in a few top of the range AV receivers and processors. Like many who use Audyssey, however, our client complained that despite the improvements XT32 made to the bass it also negatively affected the higher frequencies.

There are a few good and interesting case studies that came from this room. One we are going to share with you today is related to how Audyssey XT32 handled bass vs. hand dialed parametric EQ* in the subwoofer range. We’ll look at this from two perspective – frequency response and time decay.

* A type of equalizer that allows continuous control over three parameters: frequency, bandwidth, and amount of boost or cut. We used the miniDSP to implement the parametric EQ fllters for this client. The miniDSP processors are low cost and OK for use in a subwoofer chain, like in this instance. We do not recommend their use in the main speaker feeds. For this you need a higher quality unit such as a Xilica, as used by Procella and Legacy Audio or a DEQX.

Frequency Response

  Frequency response before Audyssey XT32 or hand dialed parametric EQ

Before, frequency response 1/12th octave  

Frequency response after Audyssey XT32

After Audyssey XT32 frequency response, 1/12th octave smoothing

Frequency response after manual parametric EQ

After manual parametric EQ frequency response, 1/12th octave smoothing

 

Frequency response clearly looks like a wash doesn’t it! Audyssey is maybe a little flatter, but causes a suckout just above the sub range at 85Hz.

Time Decay (Spectrogram)

Spectrogram before Audyssey XT32 or manual parametric EQ

Before spectrogram

Spectrogram after Audyssey XT32

After Audyssey XT32 spectrogram

Spectogram after hand dialed parametric EQ

After manual parametric EQ spectrogram

A spectrogram is a way of looking at how sound decays in the bass frequencies. It shows SPL over time at different frequencies. It is very easy to see the influence of room modes in these graphs by looking for extended ‘tails’ which stretch out in the time dimension. One has to try and not be misled by how differences in measurement levels can affect the charts – the scale is sized to fit the highest peak. If that peak is significantly higher than the average bass level then the room modes will stand alone, and be easy to spot. Such a graph can be seen in the ‘before’ example. In the after examples you can see that Audyssey does not do as good a job at reducing the long time decay of the modes around 20Hz and 43Hz as hand dialed parametric EQ. You can also see that in general the right hand side of the graph is significantly higher in level than for the hand dialed parametric EQ graph.

Conclusions

Clearly Audyssey XT32 does not deal with modal ringing as well as hand dialed parametric EQ from Acoustic Frontiers. Please remember that we did all of this calibration work remotely, using our XTZ Room Analyzer II Pro. So, if you have a bass problem, and the room correction algorithms just ain’t cutting it, be sure to call the experts at Acoustic Frontiers!

A postscript message from Nyal Mellor, our founder…I am really looking forward to getting our demo room up and running so we can do more of these tests in a controlled environment! I’d personally like to look at Anthem’s ARC, do more investigation on Audyssey XT32 and also get to grips with Trinnov.

21 thoughts on “Audyssey XT32 vs. Parametric EQ: A Comparison of Manual vs Automagic Calibrations”

  1. Hi Nyal thank you i will download the calculator.

    If it helps, i have played around with the XTZ suggested filters and specially the Q.
    Basically by halving the Q e.g. XTZ suggested Q10.00 i loaded Q 5.00 and this produced the same response as the re-measurement for the Velodyne DD and SMS-1.
    I kept the same suggested gain and frequency.
    From then on i entered manual filters.

    I have more experimenting to do.

    Why would halving the Q do this?

    Peter

  2. Hi Nyal thank you i will download the calculator.

    If it helps, i have played around with the XTZ suggested filters and specially the Q.
    Basically by halving the Q e.g. XTZ suggested Q10.00 i loaded Q 5.00 and this produced the same response as the re-measurement for the Velodyne DD and SMS-1.
    I kept the same suggested gain and frequency.
    From then on i entered manual filters.

    I have more experimenting to do.

    Why would halving the Q do this?

    Peter

  3. Hi Nyal thank you i will download the calculator.

    If it helps, i have played around with the XTZ suggested filters and specially the Q.
    Basically by halving the Q e.g. XTZ suggested Q10.00 i loaded Q 5.00 and this produced the same response as the re-measurement for the Velodyne DD and SMS-1.
    I kept the same suggested gain and frequency.
    From then on i entered manual filters.

    I have more experimenting to do.

    Why would halving the Q do this?

    Peter

  4. Hi Nyal,

    I also would love to see a face-off between those three vendor's algorithms. However, I would like to throw one more into the mix; TACT, which is found in the forthcoming XMC-1 pre/pro from Emotiva.

  5. Hi Robert,

    I will TACT to my list. I think I read somewhere that another company was also licensing the TACT algorithms. Any idea who that might be?

    Also DIRAC might make it's way into more reasonably priced components. At the moment it's only in the Theta and Datasat processors, both of which are $$$.

    Some of the Japanese companies also have their own proprietary processes, including YPAO (Yahama) and Pioneer's.

    Nyal

  6. Hi Nyal,

    No, I have not heard of another company licensing TACT. I also have not heard of DIRAC, likely because I only focus on products with prices that are more down to earth. In that sphere, ARC and Audyssey XT32 have the most proven track records, with the consensus giving ARC the lead. Trinnov made it into a Sherwood Newcastle receiver and appeared promising with its speaker re-positioning algorithms, but it did not get much review time as it was buggy and so retired rather quickly. Also, Outlaw had planned to release a pre/pro with Trinnov, but they were not able to follow through. Emotiva's implementation of TACT should add an exciting competitor to the mix. Of the Japanese options I believe the Pioneer implementation in the Elite receivers is well received, but not of the same caliber as the Anthem and Audyssey offerings.

  7. Hi Nyal

    Did you load the filters recommended by the XTZ?

    I have Velodyne DD series and have loaded the filters suggested by XTZ but the filters or Q don't seem to match very well.

    Peter

  8. Hi Peter

    In this particular case I used the values suggested by XTZ as a starting point, and then from there manually iterated (the 'hand dialed' part) the values of frequency, Q and gain to flatten the response.

    With XTZ one trick is to right click in the frequency response window and drag your mouse. This allows you to add filters above those suggested by XTZ. You can also manually edit the filter values by double clicking each row in the EQ table. Finally be sure to use the 'Stimulus EQ' mode (i.e. parametric EQ simulation) to get everything nice and flat before transferring to an outboard EQ.

    Note that not all products that incorporate parametric EQ do so in the same way or accurately. For example I was using a Rives Audio PARC the other day during a calibration and the front panel frequency values were 5-8Hz off reality. There are also examples where EQs have different filter shapes. So the only real way to find out what is going on is to remeasure with the XTZ to confirm.

    Thanks,

    Nyal

  9. Hi Nyal thank you, hope you had a great Christmas.

    With your expereince have you noticed that the filters that are manually (right click response) and suggested by XTZ are accurate, but is the Q value need tweaking?

    I wonder if there is some sort of formula / info (calculation) for the the Q values and gain that can get myself / other users closer for equalizers like the Velodyne SMS-1 (which is also in the Digital Drive)?

    When you use to adjust the filters for EQ which smoothing do you use?

    Thanks again

  10. Hi Robert

    Yes I saw that one – but thanks for the link, many might have not. One of my goals for this year is to assemble a bunch of room correction products, including the Dirac, and do some testing on them to see how well they work!

    Nyal

  11. Hi Peter

    Yes, the Q often needs tweaking a little.

    Q defines the bandwidth of the filter. This calculator is pretty useful.

    For EQing I use 1/12th or higher resolution if it's available.

    Thanks

    Nyal

  12. Hi Peter

    It's a known fact that different parametric EQs apply the core frequency, gain and Q controls differently. You would think that they would all give the same results as all of the variables are mathematically defined!

    I don't know why the Q values suggested by XTZ and Velodyne were so different. For what it is worth I can take the Q values from XTZ and put them into Xilica or MiniDSP EQs with pretty much the desired effect.

    Thanks!

    Nyal

  13. Hi Nyal, Great article!
    I've done a lot of Audyssey setups and I've always wanted to verify the results, but
    I don't know how to use my measurment system (REW) to do it.
    How did you make a measurment with Audyssey turned on?
    As far as I know Audyssey works only on HDMI inputs (on analog audio is disabled) and
    XTZ has only analog audio output.
    How do you connect your XTZ system?

    Thomas

  14. Hi Thomas

    If you hook up your measurement rig to the front panel Aux left / right inputs then those will typically go route through the AVRs digital signal processing (DSP) and room correction. The pack panel analog inputs, especially the multichannel analog inputs, in my experience typically bypass DSP.

    Nyal

    1. Hi Nyal

      I am in the process of pre eqing sub frequencies with Xilica xp4080. Can you tell me if I run PEQ before Audyssey xt32 setup?

      1. Hi Dwight

        Good question. Generally I would pre-EQ with the Xilica before running Audyssey on top. Hopefully with the Xilica you will be able to do just as good a job of EQing your system in the bass as Audyssey XT32 (if not better, as this blog article demonstrates you can). Then you can switch XT32 off (I am not an Audyssey fan!).

        Nyal

  15. Pingback: ForoDVD
  16. Pingback: Anonymous

Leave a Reply to Nyal at Acoustic Frontiers Cancel reply

Your email address will not be published. Required fields are marked *