SPL and Waterfall Interpretation

Discussion in 'Official REW (Room EQ Wizard) Support Forum' started by dwillis60, Jan 4, 2019.

  1. dwillis60

    dwillis60 Member
    Thread Starter

    Joined:
    Jan 2, 2019
    Messages:
    38
    Likes Received:
    0
    I have a fairly small (21x20) music studio that I use for both mixing and recording. I have a fair amount of the room treated with OC 703 bass traps and broadband absorbers. All four corners have 4" bass traps (floor to ceiling) and I have roughly three 2" panels on each of the four walls. I also have four 24x48x2" cloud panels mounted on the ceiling running down the center of the room. These are located at the mix position, in front of my mixing desk, and two behind the mix position. Finally, the room has multiple Producer's Choice sound blankets (mounted ceiling to floor) which cover some open shelving and countertops that are located in the room against the side walls and rear wall. I'm also in the process of adding eight additional 12x48x4" bass traps at the ceiling-wall corners.

    I've begun playing with REW and I was hoping to get some input on my initial measurements. Specifically, the SPL graph and a waterfall chart. For now, I'm just looking at the low end from 40 to 400Hz.

    Here's what I think I see: First, a fairly significant bump at around 52Hz. Luckily, most of my work is acoustic, with no mixing of EDM, House, Hip Hop, or even Pop/Rock. So I'm not terribly worried about addressing this bump. At some point, I might consider some tuned bass traps, but not now. Second, and more worrisome, there is a deep null at around 120Hz followed by a large bump in the neighborhood of 150Hz. Both of these should have an important impact and are likely to keep me from accurately assessing the low end when mixing. I'm hoping that they can be addressed to some extent by additional bass trapping but I have my doubts based on the size of both the null and the bump. I also worry that 120 Hz might be a little low for conventional DIY OC 703 traps.

    The Waterfall confirms these impressions, but also suggests less ringing than I might have expected. Most of the data are consistent across the low end and well within roughly 250ms. Again, there is a problem just above 50Kz. But as I mentioned before, this is not really an area of concern right now.

    Any additional thoughts would be appreciated. 40-400 SPL.jpg 40-400 Waterfall.jpg
     
  2. dwillis60

    dwillis60 Member
    Thread Starter

    Joined:
    Jan 2, 2019
    Messages:
    38
    Likes Received:
    0
    Here are a few pictures of the room just in case your curious.
     

    Attached Files:

  3. Wayne A. Pflughaupt

    Staff Member

    Joined:
    May 21, 2017
    Messages:
    209
    Likes Received:
    26
    Location:
    Corpus Christi, TX
    The 150 Hz bump is probably not severe enough to worry about. A high-rez graph tends to visibly exaggerate things once you get above the low bass frequencies. I’d suggest switching to 1/3-otave smoothing to get a graph that should better look like what you actually hear.

    The null, though – that’s probably going to have to be fixed via placement changes. I doubt you have room for the amount of traps it would require to fix something like that. Spreading the speakers further apart may make a difference.

    Regards,
    Wayne
     
  4. Matthew J Poes

    Matthew J Poes Staff Writer
    Staff Member

    Joined:
    Oct 18, 2017
    Messages:
    1,653
    Likes Received:
    307
    https://www.avnirvana.com/threads/a-beginners-guide-to-room-eq-wizard-rew.3723/

    You may find my video answers some of these questions.

    I would ignore the waterfall. Nothing there that isn’t in the steadystate. If the bass rings it would show up in the steadystate and that’s what we see.

    I agree that the biggest problem evident in that measurement is the dip. I suspect it’s a ceiling or floor reflection. Before doing anything I would take more measurements over a wider range of positions than just the one you showed. I would take maybe 5-6 different measurements all around where your head would be changing the lateral and vertical position. This will help us figure out the source of the dip and if it moves around.

    Now as for a fix, you can try moving the speakers. If it is a floor or ceiling dip then you won’t be able to fix it. However, thankfully, adding subwoofers will. Your tonal balance actually shows that the speakers have an overly flat response in room which is not what it should be. An in room measurement should actually be more tilted with more bass. Subs might help.
     
  5. dwillis60

    dwillis60 Member
    Thread Starter

    Joined:
    Jan 2, 2019
    Messages:
    38
    Likes Received:
    0
    I tried smoothing at 1/3 and 1/6 (See attachments). 1/3 is a little optimistic to the point of eliminating or at least masking the issue. It makes it appear that the frequency range falls within about 6db, which would be great for my purposes. 1/6 smoothing seems a little more appropriate.

    Rearranging the speaker and mix positions makes some sense and I'm in the process of doing that. My monitors sit at roughly 55" tweeter-to-tweeter. I've seen wider layouts, including Bobby Owsinski who suggests 67 inches. So that's an easy move and test.

    I've also been moving the monitors and the mix position forward towards the front wall. The current measurements are based on a mix position at 38 percent of the room length, which is a standard starting point. In my case, that's about 7'6". I've moved everything forward in two foot increments and also did a near flush mount measurement. Unfortunately, this seems to create more problems than it solves. The depth of the primary null is reduced by several db's but other anomalies pop up--especially in the 80-100Hz range.

    I haven't tried moving the mix position towards the center of the room because I only have a limited amount of space in that direction. I'm also leery of mixing near the center of a nearly square room. And I don't think that moving 2-3 feet would shift the primary null close enough to the frequency limits of the monitors (45Hz).

    I did get some interesting reductions in the null when I moved the mic position forward (about 12") and left the desk and speakers stationary. This seemed problematic to me, though, because I'm not clear if there are ramifications to mixing inside of the standard equilateral triangle. This is a rule that most people seem to swear by, and violating it might reduce my ability to properly hear the stereo image. Again, though, I'm not clear. Some rules are really not iron clad and some are meant to be violated. This could also be a tradeoff situation, where I'm gaining one benefit and loosing another. There's a lot of that going around in the home studio world.

    Thanks again for your suggestions
     

    Attached Files:

  6. Matthew J Poes

    Matthew J Poes Staff Writer
    Staff Member

    Joined:
    Oct 18, 2017
    Messages:
    1,653
    Likes Received:
    307
    7'6" is exactly the length of 150hz, sounds like that is your null. That is likely SBIR then. It is possible that it is so big because it is also a multiple of other reflections such as the full path distance between the monitor, back wall, and back again. Ceiling to floor, etc. How tall are your ceilings (normally ceiling and floor bounce effects hover at 150 and 250hz or so. My SBIR article get's into this as well I believe.
     
  7. dwillis60

    dwillis60 Member
    Thread Starter

    Joined:
    Jan 2, 2019
    Messages:
    38
    Likes Received:
    0
    I'll give that a try and see if there is any change in the null.

    Adding a subwoofer would probably involve an upgrade to new monitors. The ones pictured above are ancient Event PS6's and I'm not sure that a matching sub would be available. I'm also not clear if a matching sub is necessary. Frankly, I don't know that much about subs. I've never really given them much thought, largely because of the type of music I play. I was planning an upgrade to better monitors but I wanted to check my room carefully before spending the money. If my space is fatally flawed, it makes more sense to split mixing tasks between decent headphones and monitors.
     
  8. dwillis60

    dwillis60 Member
    Thread Starter

    Joined:
    Jan 2, 2019
    Messages:
    38
    Likes Received:
    0
    The ceilings are standard 8 foot and made of drywall.
     
  9. Matthew J Poes

    Matthew J Poes Staff Writer
    Staff Member

    Joined:
    Oct 18, 2017
    Messages:
    1,653
    Likes Received:
    307
    They don't need to match. At LF's the sub doesn't really have a sound, the room does and what you hear at LF's is the room.

    Because the null is so high, you would actually need to use a pair of subs operated in stereo to avoid detection. You wouldn't want to sum to mono the bass that high.

    Try moving the speakers back toward the front wall and see what you get. The null should move, and sometimes! it goes away.
     
  10. dwillis60

    dwillis60 Member
    Thread Starter

    Joined:
    Jan 2, 2019
    Messages:
    38
    Likes Received:
    0
    I'm guessing that, with an 8ft ceiling, adding more OC703 to the ceiling cloud won't have much of an effect on a null of that frequency and size.
     
  11. Matthew J Poes

    Matthew J Poes Staff Writer
    Staff Member

    Joined:
    Oct 18, 2017
    Messages:
    1,653
    Likes Received:
    307
    I doubt it, but maybe. How much of the ceiling is covered.

    I ran some calculations on your room based on what information you provided and I actually can't seem to quite match that 150hz null.

    When you measure, are you measuring only one speaker at a time? Left or right? Not both at the same time? Just want to make sure there isn't something else going on.

    Also, would you be able to save out the MDAT file and post it? The double dip in that null makes me think there are compounding issues going on.

    Also, you mentioned that the null goes away when you move the speakers but new issues arrise. That moving the mic also changed things? If that is true, then it is both SBIR and room modes.

    What is the dimensions of the room itself?

    When I have stubborn nulls or peaks in a response that I can trace down, I often pull the entire room apart, take everything out, and just put back the speakers (no treatment) and take measurements. If there is no null, that tells you something. If there is a null, that too tells you something, but with no mixing desk or anything else in the way, at least you have the freedom to easily move the speakers around to see what might be going on.
     
  12. dwillis60

    dwillis60 Member
    Thread Starter

    Joined:
    Jan 2, 2019
    Messages:
    38
    Likes Received:
    0
    My tests show the null moving slightly, perhaps closer to 130hz and decreasing by a few db. This is with the monitors and the desk nearly flush to the wall, However, this seemed to create other issues--particularly in the 80-100hz region. At that point, I assumed that I was not driving the null far enough to be effective trapped by adding more standard bass traps. And now I had multiple problems: The null and some issues in the 80-100Hz area.
     
  13. Matthew J Poes

    Matthew J Poes Staff Writer
    Staff Member

    Joined:
    Oct 18, 2017
    Messages:
    1,653
    Likes Received:
    307
    Also, in terms of mixing rules, I'm not a studio guy beyond occassionally doing acoustic work for them. I know nothing about mixing. However, @DanDan has been around lately and may be able to chime in with some advice.

    Just as an audio enthusiest who likes to think he has a good handle on stereo imaging, I would think that creating a perfect equalateral triangle would not be a necessity for mixing. At least based on the work of Toole, he seemed to suggest that a mix engineer and a home listener should ideally be listening under similar conditions. If that is true, then when it comes to stereo imaging, Everyone and their mother has their own opinion. I think lots of the different "rules" create a great image, no one method seems right. In my theater, my mains sit closer together than the distance to me by about a foot. I currently have a set of speakers under review setup in the theater but spaced into more of an equalateral triangle. Both sound good and both have a rock solid image. The wider distance simply widens the stage a bit.
     
  14. dwillis60

    dwillis60 Member
    Thread Starter

    Joined:
    Jan 2, 2019
    Messages:
    38
    Likes Received:
    0
    As shown in the photos above, there are 4 panels on the ceiling, All panels are 24x48x2" and they run from in front of the desk to just past the center of the room. There is an air gap of about 4" between the ceiling and the panel. So additional insulation is possible.
     
  15. dwillis60

    dwillis60 Member
    Thread Starter

    Joined:
    Jan 2, 2019
    Messages:
    38
    Likes Received:
    0
    I measured each speaker and the results showed little appreciable difference. Then I moved on and started using both speakers in all subsequent measurements. I should also mention that I'm measuring with the mic (a Dbx RTA-M) pointed at the ceiling.

    Two other issues. I have no calibration file for the mic. I did use a generic .cal file for the Dbx-RTA-M, but it showed no differences compared to a run with no .cal file. So I haven't been using any .cal file. At some point, if it really makes a serious difference, I might send the mic out for calibration.

    The other issue is my interface. I'm using an Apogee Duet and my measurements are not based on a calibration for the interface. Apogee claims it's flat and I've seen measurements to that effect. I've also read Ethan Weiner who has asserted that he can't imagine any high quality modern interface that isn't close to perfectly flat. I'm not sure if any of these assumptions could cause the problem, but it's worth mentioning.
     
  16. dwillis60

    dwillis60 Member
    Thread Starter

    Joined:
    Jan 2, 2019
    Messages:
    38
    Likes Received:
    0
    I've looked at the null for days and it looks to me that it's closer to 120HZ with the bump at around 150hz.

    The room is 21x20x8 but because of some weird construction choices, it could also be about 20x20x8. I ran those numbers through the AMROC (?) room calculator and came up with the following axial room modes. Not sure how they match up but there are modes at 118hz, 140hz, 148, etc.

    1
    28.13 Hz
    A0
    1-0-0
    ax

    2
    29.61 Hz
    A0#
    0-1-0
    ax

    4
    56.27 Hz
    A1
    2-0-0
    ax

    5
    59.23 Hz
    A1#
    0-2-0
    ax

    8
    70.33 Hz
    C2#
    0-0-1
    ax


    13
    84.4 Hz
    E2
    3-0-0
    ax

    14
    88.84 Hz
    F2
    0-3-0
    ax

    25
    112.53 Hz
    A2
    4-0-0
    ax

    30
    118.46 Hz
    A2#
    0-4-0
    ax


    41
    140.67 Hz
    C3#
    0-0-2
    ax

    42
    140.67 Hz
    C3#
    5-0-0
    ax

    51
    148.07 Hz
    D3
    0-5-0
    ax

    74
    168.8 Hz
    E3
    6-0-0
    ax


    80
    177.68 Hz
    F3
    0-6-0
    ax

    104
    196.93 Hz
    G3
    7-0-0
    ax
     
  17. dwillis60

    dwillis60 Member
    Thread Starter

    Joined:
    Jan 2, 2019
    Messages:
    38
    Likes Received:
    0
    127hz and 149Hz for the null and the following peak--at least according to the info tab combined with the crosshairs.
     
  18. dwillis60

    dwillis60 Member
    Thread Starter

    Joined:
    Jan 2, 2019
    Messages:
    38
    Likes Received:
    0
    Try that when you hit 60 and you're a grandparent! I think you need a note from your cardiologist.
     
  19. Matthew J Poes

    Matthew J Poes Staff Writer
    Staff Member

    Joined:
    Oct 18, 2017
    Messages:
    1,653
    Likes Received:
    307
    I hear ya. Well hopefully we can figure out what this is without drastic measures. It's an odd null.
     
  20. Matthew J Poes

    Matthew J Poes Staff Writer
    Staff Member

    Joined:
    Oct 18, 2017
    Messages:
    1,653
    Likes Received:
    307
    Well I modeled your room in REW, did an SBIR calculation, and gave this a bit more thought. So here is what I am coming up with. it seems that it may be a perfect combination of many different kinds of nulls combining near by. Your room is basically square. Square rooms cause two identical modes for the width and length and this caused them to combine. They reinforce each other.

    Specifically they will reinforce at the midpoint of the room because the medial modes are compounded by this effect.

    So if you have it setup so your speakers are 1/3 into the room and you sit roughly in the middle of the room, it's exactly the worse possible place to put your head. It is where the peaks and dips from this will be their worst. Otherwise, that location for the speakers is a good location to minimize other kinds of boundary effects. For example, most of the boundary effects seem minimal.

    It is also possible that the bass traps may be making things worse, oddly enough. If placed in the wrong locations for combatting this problem, they may actually exacerbate this by eliminating modes that might otherwise offset these. If you don't think of modes as bad and instead as just a thing to work with (which is the reality here), then imagine a scenario where you have lots of modes throughout the room and placing you or the speaker in different locations excites those modes differently. Now imagine a scenario where a specific location causes enough modes to be excited that for every null there is a near by peak. The end response is still relatively flat. Now imagine taking that same scenario and eliminating most of those modes (or weakening them) but leaving the strongest ones which happen to be identical, and thus have the biggest nulls and biggest peaks. This is how adding absorption can make things worse, and it's possible that is going on here.

    From what I can tell, these nulls/peaks are being caused by length and width medial modes (those that run directly down the middle of the room. I wouldn't necessarily remove your panels to fix this, but it is possibel that more traps on the back wall and both side walls would fix this. My dual sub trick probably helps too. Also, anything to move your head out of that exact midpoint should make it less, but it's possible the null is over a wide area. Sounds like you tested that and didn't find small changes in position mattered. Unfortunately there is not much you can do to fix the symetric dimensions, so I suspect more treatment and subs might be the only option to fix this.

    Here is your room now:
    RoomSim.JPG

    -The position and effect is a little off but its unmistakable that the dip is there. The fact that the exact location differs is not important, its really common for that to be true. These simple simulators make a lot of assumptions that aren't true. Walls aren't perfectly stiff, rooms aren't perfectcly cuboid, etc.

    Romsim2.JPG
    Here is what happens when you add two subs in the left and right corners. They make a big difference, but to have this effect, you need the subs to operate up to about 150hz, and for that to sound acceptable, they would need to be operated in pairs in stereo.

    Roomsim3.JPG

    This even nicer looking graph is what happens with 4 subs and you can easily see the improvement. This simulator is too basic to allow full exploration of the possibilities, but if you did this, you would likely want to operate the rear subs at a lower frequency, say 80hz and below, but still operate the front subs up to 150hz.
     
  21. Matthew J Poes

    Matthew J Poes Staff Writer
    Staff Member

    Joined:
    Oct 18, 2017
    Messages:
    1,653
    Likes Received:
    307
    Actually I just moved the speakers closer together and realized that is how you have them, once I put the mic back into the triangle it matches your problem better.

    Romsim3.JPG

    So Try spacing the speakers farther out, maybe more like 60-70" apart and then sitting farther back if possible. If these sims are right, that free and easy tweak will lessen the problem.
     
  22. dwillis60

    dwillis60 Member
    Thread Starter

    Joined:
    Jan 2, 2019
    Messages:
    38
    Likes Received:
    0
    Just to clarify...The center of the room is 10 feet. My head (the mix position) sits around 7 feet. The speakers are about 4 feet apart (probably more like 52" tweeter-to-tweeter) and about 3 ft from the front wall. These figures are approximate, since my head moves and I'm never clear on how to measure speaker placement distances. For example, is the distance from the wall measured from the rear or the face of the speaker? Is speaker separation measured tweeter-to-tweeter or in some other way? Perhaps using the inner or outer edge.

    Yikes! That's an interesting perspective, although the logic is a little above my pay grade. The bass trap gurus and the manufacturers always say, "You can never have enough bass traps." Of course, these are the same people who suggest "diving" into the Waterfall. And I see that you are more of a resist-the-waterfall type of guy.

    This suggest that the floor-to-ceiling issue is less of a problem and that additional insulation on top of the existing ceiling clouds would do little to solve the issue. Is that correct?
     
  23. dwillis60

    dwillis60 Member
    Thread Starter

    Joined:
    Jan 2, 2019
    Messages:
    38
    Likes Received:
    0
    This might be of interest. I moved the measurement mic forward and backward. The red line shows the mic at 48" with the speakers spread at 48" (an equilateral triangle). The green line represents the mic at 36" with the speakers again located 48" apart. Finally, the blue line shows the mic pushed back to 60" with the same 48" speaker spread.

    The mic movement has reduced the null but with some tradeoffs. In particular, a dip at 80hz when the mic is at 60" and a further boost at 150hz when the mic is at 36". Moving the mic also raises that question about the importance of the equilateral triangle. I may need a variance from the rule granted by one of the home studio gods. Or at least an authority on room acoustics.
     

    Attached Files:

  24. Matthew J Poes

    Matthew J Poes Staff Writer
    Staff Member

    Joined:
    Oct 18, 2017
    Messages:
    1,653
    Likes Received:
    307
    I do think extra insulation on the ceiling won’t do much.

    As for “never too much bass traps” that is true but when you don’t have most surfaces covered in significant LF damping, then strategic placement becomes important. Sometimes you end up treating the low hanging fruit and it leaves the worst offenders bare. I’m also not suggesting this is what’s happening, just that it can happen and may be what we see here.

    As for the waterfall, people don’t seem to understand how low frequency acoustics work when they promote the waterfall. Or if they do, they say very confusing things. They talk about the ridges as ringing, which is true, but as if ringing is a time domain property detached from the steadystate. That isn’t true. It’s all minimum phase and as such anything in a waterfall is also in the steadystate. Get a flat response and you don’t have ringing anymore. So why look at the waterfall, If the response isn’t flat then there is still possible ringing.

    If your interest is LF decay, waterfalls are a very hard way to examine that. A much better way is to use things like filtered impulse response and wavelets. They make it possible to actually measure decay and see it more obviously. So I say no waterfalls because they aren’t useful.
     
  25. dwillis60

    dwillis60 Member
    Thread Starter

    Joined:
    Jan 2, 2019
    Messages:
    38
    Likes Received:
    0
    I'm not clear on how to setup a wavelet for proper interpretation or even if it's appropriate to analyze for my situation. Having said that, here goes...

    These are both based on my previous set up--48" speaker spread with mic positioned 48" out. All at the room position described above. Speakers about 3ft from front wall and mix position roughly 7 feet.
     

    Attached Files:

Share This Page