davewantsmoore Posted August 5, 2018 Share Posted August 5, 2018 18 minutes ago, almikel said: I think you meant to say using EQ to clean up the frequency response only cleans up the phase response where the response is minimum phase? Yes, I hope that's the message I got across: "Fixing frequency, also fixed phase. This is true where the response is "minimum phase". 18 minutes ago, almikel said: Using treatment (not EQ) At the risk of jumping off into philosophical semantics.... I like to call 'room treatment' EQ.... lest we lose sight of the reality that speaker position, a bass reflex port, a horn, a bass trap, or a signal filter whether DSP, active, or putting a LCR in the speaker circuit ..... are all just "EQ". (heh, sorry) 18 minutes ago, almikel said: Eg using absorption to damp SBIR (which is never minimum phase) will reduce swings in frequency response and phase response. Yes.... but this example is a little different. In this example the treatment is removing the cause of the problem. ie. if you 'infinitely' treated the room, there would be no more SBIR... so of course, the frequency and phase (non-minimum phase) distortion never happens. With other forms of EQ, you are not "removing the cause" (and preventing the distortion from ever happening)... you are correcting it after it happens. 18 minutes ago, almikel said: @Hipper - I'd be trying some specialist treatment for that big dip at 50Hz - Helmholtz, or limp mass etc. Mmm. It doesn't change with the repositioning of anything. I wonder what the room dimensions are. Link to comment Share on other sites More sharing options...
Hipper Posted August 5, 2018 Share Posted August 5, 2018 Thanks for your comments, Mike and Dave. I don't really want to hijack this thread but....! My room, in a first floor flat, is 420cm (14') by 386cm (13') by 240cm (8') high with sonic insulation and carpet on the wooden floor, brick walls with plaster and wallpaper, and a concrete ceiling. Therefore my speakers, set up on the longer wall, are placed 140cm in from the side wall and 129cm in from the front wall, with my ears (when sitting in my chair) 129cm from the rear wall. The speaker distances were taken to the inside front corners of the speakers. They were toed in to face the centre mark on the rear wall. As a result I sit about 150cm from the drivers but they seem to integrate well at that distance. In order to sort out that dip at 50-70Hz I sought the advice of GIK (who supplied the room treatment) and also took lots of measurements around the room of various frequencies between 40-70Hz, using in SPL meter placed back, middle and front, and near the ceiling, floor and middle, but there was no obvious solution. I also checked out each speaker, and the gear, including the Behringer, to make sure there was nothing these were doing. It was also not a problem with REW or microphone calibration. I have two different makes of mic and both exhibited the issue. It's a mystery but as using EQ solved it I move on. GIK do sell frequency targeted traps but until I could find out where to place them it seemed pointless investing in them. Back to phase, I posted this a couple of years ago which may be of interest: https://pinkfishmedia.net/forum/threads/digital-equalisers-phase-and-other-possible-damage-to-sound.177219/ And a question please. It is clear that my efforts have improved the phase measurements in the room even though I didn't target them. Can I attribute the resulting improved sound to: 1. Smoother frequency response. 2. Lower decay times across the whole frequency range. 3. Less and much smoother phase changes Or of course a combination of any or all these, or do they all go hand in hand anyway? Floyd Toole in his seminal book Sound Reproduction writes about phase (pages 418-421) and concludes that we like smooth phase but not necessarily linear. Phase issues though, after recording, engineering and replaying in a room are so altered that we can't hope to get accurate phase information. However it does not seem to be important as our brains can navigate their way through all this. So, can phase improvements be heard and is it wise to target them (as DEQX does with it's speaker calibration algorithms), or should we just target frequency response and decay times, and accept that phase improvements may likely follow, whether heard or not? Link to comment Share on other sites More sharing options...
Hipper Posted July 27, 2019 Share Posted July 27, 2019 (edited) Following a discussion on this thread: https://www.stereo.net.au/forums/topic/285535-listening-room-acoustics-the-ongoing-challange/page/3/?tab=comments#comment-4353017 where the subject turned to Excess Group Delay (EGD) graphs (from the 20th July), and not wanting to hijack that thread I thought I'd ask my question here. I refer to my post on the August 4th 2018 and above on this page. Below is my EGD graph at the moment from my listening position with all room treatment installed and EQ as described above, namely up to around 180Hz, but no EQ above that: Firstly the EQ I used to flatten up to 180Hz made no change to the EGD over that range. It still contained that 25ms hump. The passive crossovers of my three way speakers are at 280Hz and 6.9kHz according to the designer. My understanding is that this graph represents where the non minimum phase areas are. I have three questions please? 1. What is that 25ms hump at 50Hz? Is it something I should be concerned about? How can I find out more? 2. There is a dip of about 8dB at around 325Hz in the frequency response in my 4th August post (green) which may correspond to this 300ms peak on the EGD. I'm not worried about it but just intrigued as to what it means, does it have any significance, and what, if I fancy, could I do to find out more about it? Actually closer inspection shows the large peak on the EGD as 318Hz and the smaller one at 3.18kHz. What's this mean? 3. I'm surprised at the overall flat nature of this graph. Does that mean there is little in the way of reflections going on in my room? Edited July 27, 2019 by Hipper Link to comment Share on other sites More sharing options...
almikel Posted July 28, 2019 Author Share Posted July 28, 2019 10 hours ago, Hipper said: where the subject turned to Excess Group Delay (EGD) graphs (from the 20th July), and not wanting to hijack that thread I thought I'd ask my question here. I started this thread, and the topic is relevant, so post away 9 hours ago, Hipper said: My understanding is that this graph represents where the non minimum phase areas are. the concept of minimum phase is only relevant in a room below the transition zone of the room - I would ignore any "room" measurements of Excess Group Delay above that (say above 250 - 350Hz) - the fact that the graph shows 0 Excess Group Delay is likely erroneous - rooms are never minimum phase above their transition zone. 9 hours ago, Hipper said: 3. I'm surprised at the overall flat nature of this graph. Does that mean there is little in the way of reflections going on in my room? No - there will be loads - I suspect the very large scale of the X axis is hiding it all - remeasure over 10 - 300Hz. 9 hours ago, Hipper said: 1. What is that 25ms hump at 50Hz? Is it something I should be concerned about? How can I find out more? not sure - but I'd avoid EQ in this region - what does the FR and phase look like in this region? 9 hours ago, Hipper said: 2. There is a dip of about 8dB at around 325Hz in the frequency response in my 4th August post (green) which may correspond to this 300ms peak on the EGD. I'm not worried about it but just intrigued as to what it means, does it have any significance, and what, if I fancy, could I do to find out more about it? 9 hours ago, Hipper said: The passive crossovers of my three way speakers are at 280Hz ... according to the designer. it wouldn't surprise me if the peak in Excess Group Delay just above 300Hz was crossover related - the likelihood of a commercial passive crossover being a bit off spec (280Hz) is pretty high. cheers Mike Link to comment Share on other sites More sharing options...
Hipper Posted July 28, 2019 Share Posted July 28, 2019 (edited) Thanks for response Mike. OK some more graphs. These are all as I listen now, with full room treatment and EQ in the 0-180Hz range. These are SPL/Phase graphs taken with identical circumstances within minutes of each other but, whilst the SPL is virtually identical the phase is not. It's a similar shape but seems to be a bit lower from 40Hz in the full range measurement. Is this an inaccuracy of REW but perhaps nothing to be concerned about? Here are EGDs with a higher scale, both from the above full range measurement. About crossovers, the designer of my speakers (VMPS RM30M) said he paid a lot of attention to them (I suppose they all say that!) so I would be surprised if he got it wrong. Edited July 28, 2019 by Hipper Link to comment Share on other sites More sharing options...
almikel Posted August 2, 2019 Author Share Posted August 2, 2019 key takeouts: in that room, don't use EQ below 100Hz - it will be too position dependant - if that FR was with EQ applied below 100Hz then remove it and re-measure...if you have issues <100Hz you can try speaker/LP position changes, but unfortunately you'll need specialist treatments if moving stuff doesn't fix things you won't see the detail in phase plots with unwrapped phase - it looks messier with wrapped phase, but it's better if focusing on specific issues completely ignore Excess Group Delay above around 300 - 400Hz - but based on the large narrow peak in Excess Group Delay and the massive dip in FR and the significant Phase Response wiggle all just above 300Hz and a passive crossover point meant to be at 280Hz - Occam's Razor would point towards a crossover issue... Do a close miked measurement of a single speaker - drag it to the middle of the room and have the mic on axis, 1m away and centred between the 2 drivers with the 280Hz Xover. Do the same with the other speaker. They should measure very similarly (but given speaker and mic won't be in exactly the same position there will be differences) If the close miked measurements (FR and phase) are nice and smooth across that same region (250 - 350Hz), then it's a room issue or SBIR etc. If the close miked measurements show the same dips/wiggles in FR and phase - then it's almost guaranteed as a Xover issue - simply swapping polarity on a driver would likely fix it - be careful which driver you switch in 3 way or more setup - you don't want to impact another Xover. Cheers Mike 1 Link to comment Share on other sites More sharing options...
davewantsmoore Posted August 18, 2019 Share Posted August 18, 2019 On 28/07/2019 at 9:01 PM, Hipper said: These are SPL/Phase graphs taken with identical circumstances within minutes of each other but, whilst the SPL is virtually identical ....but it is not identical. Remember 1dB is about +25% ... and 3dB is DOUBLE .... so even fractions of a dB are quite significant differences. It is these differences in the SPL, which are causing the phase differences (it is nothing to be concerned about) Link to comment Share on other sites More sharing options...
davewantsmoore Posted August 18, 2019 Share Posted August 18, 2019 On 28/07/2019 at 9:01 PM, Hipper said: About crossovers, the designer of my speakers (VMPS RM30M) said he paid a lot of attention to them (I suppose they all say that!) so I would be surprised if he got it wrong. The peak in your EGD at 300Hz is obviously related to the non-flat frequency response you've measured at that frequency. At ~3Khz .... don't know. If that corresponds to a XO frequency, then it's either a problem with the XO.... or a problem with your measurement (eg. you measured from a "bad" location). EGD can be used to help you time align the drivers, which is what I expect he "paid attention to" .... although it is a potentially misguided way to go about it (better just use phase). Link to comment Share on other sites More sharing options...
davewantsmoore Posted August 18, 2019 Share Posted August 18, 2019 On 02/08/2019 at 10:37 PM, almikel said: in that room, don't use EQ below 100Hz No. In the region << 100Hz .... is it prudent to essentially ignore phase, and correct the SPL. The phase will follow the SPL (no matter what a measurement might indicate - the measurement is misleading) On 02/08/2019 at 10:37 PM, almikel said: completely ignore Excess Group Delay above around 300 - 400Hz Why? (Perhaps there's a reason, sorry if I missed it) On 02/08/2019 at 10:37 PM, almikel said: wiggle all just above 300Hz and a passive crossover point meant to be at 280Hz - Occam's Razor would point towards a crossover issue... Maybe. I don't see how you can know that though. What we know is that we put in a flat response... and got out a non flat SPL. Why? (Do we know?) It could be you've measured the speaker from a location where the 280Hz XO doesn't sum properly.... but I wouldn't think that's likely if the measurement position is sane. It could be a reflection from the room (hint: it's a reflection from the room, eg. SBIR) Again, sorry if I've missed something discuss.... I'm rushing. Link to comment Share on other sites More sharing options...
almikel Posted August 19, 2019 Author Share Posted August 19, 2019 ...ignore EGD above 300-400Hz... On 18/08/2019 at 6:41 PM, davewantsmoore said: Why? (Perhaps there's a reason, sorry if I missed it) @Hipper came to this thread in the context of using EGD to determine regions of minimum phase to be able to apply EQ. Above the transition zone of a room, there are unlikely to be any regions of minimum phase. @Hipper's EGD graph was "too flat" above the spike between 300-400Hz - which could lead to incorrect interpretations that EQ would be effective based on the graph showing it's "minimum phase". On 18/08/2019 at 6:41 PM, davewantsmoore said: Maybe. I don't see how you can know that though. I don't - that's why I suggested to @Hipper to do a close miked measurement to rule out room/SBIR issues. cheers Mike Link to comment Share on other sites More sharing options...
andyr Posted August 19, 2019 Share Posted August 19, 2019 1 hour ago, almikel said: Above the transition zone of a room, there are unlikely to be any regions of minimum phase. @Hipper's EGD graph was "too flat" above the spike between 300-400Hz - which could lead to incorrect interpretations that EQ would be effective based on the graph showing it's "minimum phase". But, Mike - here is the EGD plot produced by REW for my system? Looks pretty flat to me? Andy Link to comment Share on other sites More sharing options...
almikel Posted August 19, 2019 Author Share Posted August 19, 2019 Hi Andy, that's exactly my point - it was your EGD graphs in the other thread that looked far "too good" that made me think about the relevance of EGD above the transition zone of a room...and you lose all detail with smoothing... ...This is the EGD you posted with no smoothing... A much more "expected" result - it doesn't look quite "flat" anymore without any smoothing. I would be cautious of reading anything into "in room" EGD plots above the room's transition zone. Speaker/driver measurements taken outside are likely to be "minimum phase" if there are no reflections. "In Room" measurements won't be "minimum phase" above the room's transition frequency due to reflections, and below the transition zone there will likely only be some regions of "minimum phase". cheers Mike Link to comment Share on other sites More sharing options...
Ittaku Posted August 19, 2019 Share Posted August 19, 2019 8 hours ago, almikel said: Hi Andy, that's exactly my point - it was your EGD graphs in the other thread that looked far "too good" that made me think about the relevance of EGD above the transition zone of a room...and you lose all detail with smoothing... ...This is the EGD you posted with no smoothing... A much more "expected" result - it doesn't look quite "flat" anymore without any smoothing. He did flatten it further by changing his REW setup and adding more delay from 300 up as well though so you're not comparing the same response any more. Link to comment Share on other sites More sharing options...
tripitaka Posted June 9, 2022 Share Posted June 9, 2022 Just found this, what an interesting thread it is! Bumped to make this wisdom from expert practitioners visible, for folks like me who missed it the first time around Link to comment Share on other sites More sharing options...
davewantsmoore Posted June 11, 2022 Share Posted June 11, 2022 (edited) On 19/08/2019 at 10:39 PM, almikel said: "In Room" measurements won't be "minimum phase" above the room's transition frequency due to reflections It depends on the time window of the measurement (ie. is any refelected sound included in the data, or not?)..... (but yes, there will always be diffictult to avoid things like cabinet diffraction) ... and what the window should be, depends on the purpose of the measurement. ie. what is the question? Edited June 11, 2022 by davewantsmoore 1 Link to comment Share on other sites More sharing options...
Recommended Posts