My REW Measurements - any advice

Discussion in 'Official REW (Room EQ Wizard) Support Forum' started by kalorew, Jan 27, 2019.

  1. kalorew

    kalorew New Member
    Thread Starter

    Joined:
    Jan 27, 2019
    Messages:
    4
    Likes Received:
    0
    Hi all,

    My first post here. I am excited about my first measurements with REW. I live in a small apartment and perhaps my measurements would not make sense but let me describe what I did.

    Firstly, I calibrated my speakers with the built in AVR sounds. Then, I fired up SPL meter and used pink noise to get where the SPL shows 75 dB. Now onto the measurements.

    I measured both my front speakers with my sub as well as did separate measurements with just the sub and the left speaker and the sub and the right speaker (see the first image attached - the three separate lines). Two things I saw was a bump in the bass and also the individual speaker measurements showed much flatter (although lower dB) than the combined speakers with the sub which produced a huge dip in the high frequency range.

    To solve the bass bump I simply reduced my subwoofer with 3 dB, so technically the SPL meter would measure 72 dB if I played it only by itself. Nevertheless, that reduced some for the bass bump (see the second image).

    Next, I dealt with the high frequencies using the AVR's built-in GEQ: -1.5 db @ 500, +5 dB @ 4k, +5 dB @ 8k. That managed to raise up the high frequency measurement (see the third image).

    Lastly, I took a capture of the Before & After to contrast the as-is and the improvement I made (see the fourth image). Now that I look at it, I probably should have applied some boost at the 16k band as well..

    I know you are not supposed to apply EQ beyond 200-300 Hz but I decided just to experiment. Did I do an improvement or am I better off to remove the GEQ on the high frequencies. Perhaps the chart showing an improvement is not an improvement? As far as listening -- I do notice more pronounced vocals and cymbals, etc. as one would expect. I just cannot say if it is more correct than before. Hoping for some feedback and guidance from the knowledgeable folks on here.
     
  2. Grayson Dere

    Grayson Dere Moderator
    Staff Member

    Joined:
    Dec 19, 2018
    Messages:
    134
    Likes Received:
    41
    Location:
    Bay Area, CA
    "Next, I dealt with the high frequencies using the AVR's built-in GEQ: -1.5 db @ 500, +5 dB @ 4k, +5 dB @ 8k. That managed to raise up the high frequency measurement (see the third image)."

    Perhaps some one with more experience in the EQ realm can chime in here, but from what I've gathered over time is that one generally shouldn't add dB to portions of the EQ, only subtract. I believe the thinking is that by adding dB to any frequency can cause other types of issues in the frequency response. Again, maybe EQ correction should be dealt with solely case by case and there are no hard and fast rules?
     
  3. John Mulcahy

    John Mulcahy REW Author

    Joined:
    Apr 3, 2017
    Messages:
    1,203
    Likes Received:
    201
    Measurements at the listening position generally have a downward slope, a consequence of the speaker having less HF energy off axis and the absorption of the room. We hear through that and mainly hear the direct signal from the speaker. If you try and "correct" the downward slope you end up creating an upward slope in the direct signal from the speaker moving it to a "thin" or "bright" sound. EQ is best reserved for the low end where it can tackle modal resonances and beyond that only applied as a tone control to deal with recordings where the tonal balance isn't to your liking.
     
    Grayson Dere likes this.
  4. kalorew

    kalorew New Member
    Thread Starter

    Joined:
    Jan 27, 2019
    Messages:
    4
    Likes Received:
    0
    I understand. Just for the heck of it, I decided to re-run Audissey today. My measurements before were all manual settings, levels, distances, etc. So today I wanted to see how Audissey would perform. I did not use Audissey for several months. Below is a graph which shows the new settings set by the AVR and Audissey. I measured the Audissey standard curve (red) vs. No Audissey (green) vs. Audissey FLAT curve, disabling things like Dynamic EQ and Dynamic Volume. Looking at the graph, Audissey actually does flatten my bass very well. The drawback is it also touches higher frequencies. So, Audissey FLAT measured the best.

    I also decided to do one more experiment - since Audissey FLAT measured the best - I moved the mic about 4-5 inches back from the original position and measured again. You will see in my 2nd graph that bass response increased as well as HF also increased.

    So, what should I do? Should I listen to Audissey FLAT (no dynamic eq, nor volume) moving forward for flatter bass or should I turn off Audissey and know that I am lacking bass since I cannot fix it using the AVR's limited graphical EQ?
     
  5. Matthew J Poes

    Matthew J Poes Staff Writer
    Staff Member

    Joined:
    Oct 18, 2017
    Messages:
    1,729
    Likes Received:
    349
    Your measurements are a little odd so my suggestion would be to first narrow down what’s going on. I would take a set of 3-6 measurements all around your listening position. For example where your head is, to the left, right, in front, slightly up, down, etc.

    Remember what John said. You aren’t measuring what you are hearing. These measurements at high frequencies don’t equal what you hear. The best way to get a sense of what the issues are is with these multitude of spatially diverse measurements.

    I would also say that, to answer your question, use your ears. An omnidirectional microphone taking measurements in room is no replacement for two ears and a brain. Those measurements are confounding a lot of room reflections your brain filters out. Trying to eq that screws up what is going on. It’s very hard to measure in a room and figure out what is going on with the actual speaker you hear. That is why eq at high frequencies is best based on anechoic measurements.

    If you watch my REW video you will see I touch upon this.
     
    kalorew likes this.
  6. kalorew

    kalorew New Member
    Thread Starter

    Joined:
    Jan 27, 2019
    Messages:
    4
    Likes Received:
    0
    Thank you, Matt. Here is what I did as my next set of experiments.

    In the first picture attached, I compared the response of how my systems sounds "raw" (orange line) and how it sounds with Audissey but in a couple of alternatives - with @100Hz crossover (red) and 80Hz crossover (blue). Furthermore, I used the Audissey Bypass L/R which forces Audissey only to work on the Subwoofer and skips the mains, limiting its effects to only 200 Hz. My conclusion in this experiment is that for music listening, I could have Aussey in Bypass L/R mode and get flatter bass response. For movies I can enable full Audissey so all speakers are balanced.

    Question 1: should I use @80 hz or 100Hz (100Hz is significantly flatter). My sub is between the center and the right main.

    In the second experiment I wanted to see what is the right level of subwoofer trim. By default, Audissey set it to -10.5 so I played with changing that (while sub was crossed over at 100Hz) and ambition to have as flat as possible curve. In my second image you can see how results differ when I lower the trim of the Audissey sub level.

    Question 2: Does lowering the trim to -12 make this flatter?

    Thirdly, I though if I want to have a house curve, can I achieve that by simply raising the subwoofer trim. This time I crossed over the sub @ 80Hz and did increments from the subwoofer base trim set by Audissey -10.5.

    Question 3: what is the right trim as you are looking at the third graph?


    Lastly, I did what Matt suggested, to take a number of measurements around where my head would be moving around with 5-6 inches in various directions. The results are in the 4th image I attached. I used the same setting for all measurements: subwoofer crossed over @80 Hz, Sub trim -7.0; Audissey Bypass L/R.

    Question 4: What can we tell from this graph? That my nasty bump around 500 Hz is permanently there?
     
    #6 kalorew, Jan 30, 2019
    Last edited: Jan 30, 2019
  7. kalorew

    kalorew New Member
    Thread Starter

    Joined:
    Jan 27, 2019
    Messages:
    4
    Likes Received:
    0
    Any input on my last measurements? If my setup sucks, just tell me, I can take it :)
     

Share This Page