Jventer Posted March 17, 2017 Author Share Posted March 17, 2017 Thank you very much to all the contributors to my request for assistance. An issue I would like to state is that my speakers are still not optimally placed according to the manufacturer. They are 32cm from rear wall instead of at least 50cm and 82 cm from a corner in stead of 1m (and that is if I understand the distances correctly). I can move the speakers forward and further from corners for special listening purposes. I have now started from scratch. Using my tape measure I have made sure that speakers are the same distances from side an back walls and that the listening position is exactly the same length from each speaker. Also have taken lots of photos. I have changed measurement levels and microphone sensitivity levels as advised. My room's SPL unfortunately is high at about 44db. I have therefore ensured that the measurements are up to a level of 85db to ensure a difference of 40db. In other words, everything in the room is symmetrical where possible. There is a bay window on the left and another room on the right which is not. See next post for new graphs. Link to comment Share on other sites More sharing options...
Jventer Posted March 17, 2017 Author Share Posted March 17, 2017 @davewantsmoore Thanks for your input so far The following are graphs at 10 to 400hz based on your recommendations. (Great to see physical speaker movements etc etc has made the 50hz issue dissappear). So guys, what is my issue and what do i need to do, please? Link to comment Share on other sites More sharing options...
davewantsmoore Posted March 17, 2017 Share Posted March 17, 2017 10 minutes ago, Jventer said: what do i need to do, please? You need to confirm what the issue between 90 and 300 is. It is very likely to be delayed reflected (from eg. the floor, or wall) sound that is arriving out of phase at these frequencies. To investigate this. Take some more measurement. If your current measurement is 4m back from the speakers. Take one at 3m, 2m, 1m, and 10cm distance. If/how the dip changes, will help confirm (or at least narrow down) what is causing it. If that sort of thing is the cause. The the recommendation could easily be anything from "do nothing", to move your speakers around some. 2 Link to comment Share on other sites More sharing options...
andyr Posted March 17, 2017 Share Posted March 17, 2017 On 17/03/2017 at 2:27 AM, almikel said: What is minimum phase and what isn't is very hard to get your head around - the simplest explanation is the one on REW, which I've read lots but still struggle with (and way simpler than any other definition I've found) https://www.roomeqwizard.com/help/help_en-GB/html/minimumphase.htm The best thing is you don't need to understand what minimum phase is, because REW can show you via the "Excess Group Delay" function/graph - apply PEQ in the flat areas only and you should be fine - also useful for showing time alignment issues. Hi cheers Mike Hi Mike, The REW paper shows an "Excess Group Delay" graph - which, as you say, identifies 'minimum phase regions' which can successfully have PEQ applied. Can you tell me how to generate this graph? I can't see which tab at the top of the REW screen would do this. Thanks, Andy Link to comment Share on other sites More sharing options...
davewantsmoore Posted March 18, 2017 Share Posted March 18, 2017 53 minutes ago, andyr said: Can you tell me how to generate this graph? On the SPL and phase tab. You will see that the line in the legend for 'excess phase' is there but unavailable (grey). Also you'll notice the minimum phase line is also there but grey. Open the controls menu, and choose generate minimum phase. What this does is generate a new phase curve that observes the relationship to the amplitude (ie. for every wiggle in the amplitude there is a wiggle in the phase). Now you can enable (tick it in the legend) the excess phase. The excess phase line is showing you the difference between the measured phase, and theoretical 'minimum phase'. ie. where it is non-zero the measured phase is rotating more than expected. However it is an oversimplification to say that you can apply shed loads of EQ everywhere, except where the excess phase rises (where you should apply none). Treat rising excess phase like a warning to investigate this region of your measurement further as it likely contains reflections. 1 Link to comment Share on other sites More sharing options...
andyr Posted March 18, 2017 Share Posted March 18, 2017 3 hours ago, davewantsmoore said: On the SPL and phase tab. You will see that the line in the legend for 'excess phase' is there but unavailable (grey). Also you'll notice the minimum phase line is also there but grey. Open the controls menu, and choose generate minimum phase. What this does is generate a new phase curve that observes the relationship to the amplitude (ie. for every wiggle in the amplitude there is a wiggle in the phase). Great - thanks very much, Dave. 3 hours ago, davewantsmoore said: Now you can enable (tick it in the legend) the excess phase. The excess phase line is showing you the difference between the measured phase, and theoretical 'minimum phase'. ie. where it is non-zero the measured phase is rotating more than expected. I have ticked both 'Minimum phase' and 'Excess phase', separately. See attachment - Minimum phase is the first graph. But what do these graphs tell me? 3 hours ago, davewantsmoore said: However it is an oversimplification to say that you can apply shed loads of EQ everywhere, except where the excess phase rises (where you should apply none). Treat rising excess phase like a warning to investigate this region of your measurement further as it likely contains reflections. Understood. Thanks, Andy REW Phase Plots.pdf Link to comment Share on other sites More sharing options...
Jventer Posted March 18, 2017 Author Share Posted March 18, 2017 As recommended by @davewantsmoore I have done more measurements. So, from the seating position at 3.9 m I have gone forward until I got stuck right in front of the TV screen at.35m. Link to comment Share on other sites More sharing options...
almikel Posted March 18, 2017 Share Posted March 18, 2017 41 minutes ago, Jventer said: As recommended by @davewantsmoore I have done more measurements. So, from the seating position at 3.9 m I have gone forward until I got stuck right in front of the TV screen at.35m. Remind me again what the dimensions of your room are? Are these all just with a single speaker driven? Mike Link to comment Share on other sites More sharing options...
Jventer Posted March 18, 2017 Author Share Posted March 18, 2017 @almikel Room is: 4,080 W 4,660 L 2,420 H. Thanks Link to comment Share on other sites More sharing options...
almikel Posted March 18, 2017 Share Posted March 18, 2017 1 hour ago, andyr said: Great - thanks very much, Dave. I have ticked both 'Minimum phase' and 'Excess phase', separately. See attachment - Minimum phase is the first graph. But what do these graphs tell me? Understood. Thanks, Andy REW Phase Plots.pdf Hi Andy Above the transition zone (around 250Hz) it will never be minimum phase. Change the limits for below 250Hz (shouldn't have to run the measurements again unless you want to) Get rid of the Mic/Cal curve (just untick) Just show the FR and the Excess Phase plots below 250Hz Just above the vertical axis there's a little camera icon to capture JPGs if you want to post them directly into this thread. based on what you've posted, the only region approaching minimum phase is below 80Hz. Do you run a sub crossed to mains around 90-100Hz? Note that large dips in the FR always correspond to large non-zero Excess Phase (which is why the rule of thumb "don't try to EQ dips" is valid). How many speakers driven for this? Run a measurement just with 1 sub running (I think you have more than 1?) Run another measurement with just 1 of your mains running. The current graph looks like you have sub and mains integration suckout. cheers Mike Link to comment Share on other sites More sharing options...
almikel Posted March 18, 2017 Share Posted March 18, 2017 8 minutes ago, Jventer said: @almikel Room is: 4,080 W 4,660 L 2,420 H. Thanks measurements with single speaker driven? Link to comment Share on other sites More sharing options...
Jventer Posted March 18, 2017 Author Share Posted March 18, 2017 2 minutes ago, almikel said: measurements with single speaker driven? Yes, left Link to comment Share on other sites More sharing options...
almikel Posted March 18, 2017 Share Posted March 18, 2017 1 hour ago, Jventer said: As recommended by @davewantsmoore I have done more measurements. So, from the seating position at 3.9 m I have gone forward until I got stuck right in front of the TV screen at.35m. I need a mode calculator in metric here's your theoretical modes based on room dimensions I've forgotten what the difference between a circle and a dot is The peak in all FR curves around 40Hz is that 1st and 2nd axial mode. Looks like another around 70-80Hz Pretty consistent suckout just above 100Hz - could be modal - would be interesting to see if that occurs close miked - run a measurement very close to the woofer (like 1cm away). Do you know what the Xover freq of your speakers are? cheers Mike 1 Link to comment Share on other sites More sharing options...
andyr Posted March 18, 2017 Share Posted March 18, 2017 (edited) 2 hours ago, almikel said: Hi Andy Above the transition zone (around 250Hz) it will never be minimum phase. OK, thanks, Mike - but what does that mean about doing PEQs above 250Hz? Quote Change the limits for below 250Hz (shouldn't have to run the measurements again unless you want to) Get rid of the Mic/Cal curve (just untick) Just show the FR and the Excess Phase plots below 250Hz Just above the vertical axis there's a little camera icon to capture JPGs if you want to post them directly into this thread. Have done this after spending the day re-doing PEQs after your posts - lets see if these 'captures' post. Quote based on what you've posted, the only region approaching minimum phase is below 80Hz. Do you run a sub crossed to mains around 90-100Hz? Yes, 2 subs in the front corners - crossing to the 'mains' at 100hz. 24dB L-R HP & LP. I believe I have the delay for the subs "perfectly" set (they are further away than the mains - Maggies). Does REW have any way to confirm or disprove this? Quote How many speakers driven for this? 2 Maggies + 2 subs. Quote The current graph looks like you have sub and mains integration suckout. You are absolootely correct! I thought I had already tested this but after changing the invert setting for the subs in the miniDSP plug-in, I found that 100hz dip disappeared. Now, re. the latest waterfall plot, can you comment on the below?: this still has the 50hz horizontal 'bar' ... but this appears to be over 25dB down, right? there is another bar at 100hz - is this the mains harmonic? what causes the bars at 60hz & 80hz? Room modes (5.8m x 3.6m x 2.8m)? decay seems to be acceptably steep ... all the frequencies have decayed before 240mS - except the bars from 50-120hz ... and below 30hz? so is this waterfall a good result? Thanks, Andy Edited March 18, 2017 by andyr Link to comment Share on other sites More sharing options...
Jventer Posted March 18, 2017 Author Share Posted March 18, 2017 17 hours ago, Jventer said: As recommended by @davewantsmoore I have done more measurements. So, from the seating position at 3.9 m I have gone forward until I got stuck right in front of the TV screen at.35m. @davewantsmoore - your thoughts please Link to comment Share on other sites More sharing options...
ghost4man Posted March 19, 2017 Share Posted March 19, 2017 For reasons relating to mains power issues and amplifiers that are not up to my needs I made the decision to suspend testing until these essential basics were dealt with. I have to say the way in which this discussion has continued it has brought back memories of the frustration that I felt after quite literally hundreds and hundreds of sweeps, staring at graphs and trying to figure out where next to go and the sense that one is going deeper down the rabbit hole. As good as the software is supposed to be there I feel for Andy and Jventer in that here we are and it comes down trying to work out what it all means. More graphs from every distance imaginable, seemingly no end in sight. This isn't meant to be a criticism either in any way. Let me make that clear. But it does feel awfully frustrating I have to say. Link to comment Share on other sites More sharing options...
davewantsmoore Posted March 19, 2017 Share Posted March 19, 2017 21 hours ago, andyr said: But what do these graphs tell me? Non-zero excess phase, tells you that your measured phase is different from the phase you would expect to see when considering the amplitude response. A likely reason for this could be that the response includes reflected sound. The default settings in REW have a very long window (this appropriate for looking at << 200hz).... this means that the microphone is listening for a long time after the stimulus (and thus hears the reflections bouncing around). Comparing various window lengths can give you an idea of what spectral contribution the reflections are making. Link to comment Share on other sites More sharing options...
andyr Posted March 19, 2017 Share Posted March 19, 2017 3 minutes ago, davewantsmoore said: Non-zero excess phase, tells you that your measured phase is different from the phase you would expect to see when considering the amplitude response. Aah - thanks, Dave. 3 minutes ago, davewantsmoore said: A likely reason for this could be that the response includes reflected sound. The default settings in REW have a very long window (this appropriate for looking at << 200hz).... this means that the microphone is listening for a long time after the stimulus (and thus hears the reflections bouncing around). Comparing various window lengths can give you an idea of what spectral contribution the reflections are making. I will play with window length, next time I set up my computer in the 'music room' (2nd floor!). Regards, Andy Link to comment Share on other sites More sharing options...
davewantsmoore Posted March 19, 2017 Share Posted March 19, 2017 4 hours ago, Jventer said: thoughts If 80dB is the average level above 400.... then I would be inclined to do not much. If it were 85, then I might consider some EQ to bump up the octave from 100 to 200, and maybe a little higher ... by 3 to 6dB. It doesn't seem like much of this is room modes (I used a quick online mode calculator), especially seeing your room is semi open. How far are your speakers from the front a side walls? Path length differences in the ~1 or 2m range, would cause these dips. 1 Link to comment Share on other sites More sharing options...
davewantsmoore Posted March 19, 2017 Share Posted March 19, 2017 5 minutes ago, andyr said: I will play with window length, next time I set up my computer in the 'music room' (2nd floor!). The window length is applied to a measurement after it is taken.... so you do not need to take a new measurement. You just adjust the window length of an existing one. Basically what it is saying is..... I just captured 20 seconds worth of sound. How much of it should I use to calculate the curves. Link to comment Share on other sites More sharing options...
almikel Posted March 26, 2017 Share Posted March 26, 2017 apologies, been a bit busy to post @andyr - I suggest you start a new thread on your issues to stop polluting @Jventer's thread. @Jventer apologies for adding further pollution to your thread. On 3/18/2017 at 6:22 PM, andyr said: OK, thanks, Mike - but what does that mean about doing PEQs above 250Hz? Yes, 2 subs in the front corners - crossing to the 'mains' at 100hz. 24dB L-R HP & LP. 2 Maggies + 2 subs. your excess phase plot looks all messed up - likely 'cos you're running both maggies at the same time. Run another measurement of 1 Maggie and all subs. Here's a pic of FR and Excess Phase in my room taken ages ago (sub + 1 x mid bass speaker) - levels not high enough (learning all the time). IMO, PEQs above 250Hz based on listening position measurements are far too mike position dependant and should be avoided. If you can't resist then only apply broadly with low "Q". PEQs can be applied above 250Hz or so to correct speaker anomalies based on anechoic (ie outdoor) measurements - but not based on "in-room" listening position measurements. Unfortunately panels like Maggies are more difficult to measure outdoors than horns or box speakers, as the mike needs to be further away (eg min 3m) to capture the sound from the whole panel, plus you need a lot of space behind the panel (or lots of absorption) to avoid reflections from the backwave. Good measurement rigs are hard to come by, and I swapped out my amplifiers for my PSE 144's because @Listener took such great measurements on his PSE144s at a GTG and I wanted to use his "equal power amps across all drivers" measurements done on his DEQX (I run DEQX also - so it was a no brainer). In my setup I apply linear phase EQ speaker correction (via DEQX) on my PSE 144s, based on the measurements done by @Listener. For my TD18 mid bass, I've applied linear phase speaker correction (via DEQX) based on close miked measurements. I've not applied any linear phase EQ to the sub - but have tweaked both the TD18s and the sub with PEQ based on listening position measurements. On 3/18/2017 at 6:22 PM, andyr said: I believe I have the delay for the subs "perfectly" set (they are further away than the mains - Maggies). Does REW have any way to confirm or disprove this? My recollection is that Mr Spencer set up your subs - so I'm confident the delay was set correctly... ...FYI this means the delay was applied to your mains to allow the subs to catch up - the low pass filter required for the subs creates delay and means that your mains need delay to allow the subs to catchup. If you get a clean excess delay graph - re-read the REW Minimum Phase link above - the excess phase graph can show time alignment issues On 3/18/2017 at 6:22 PM, andyr said: Now, re. the latest waterfall plot, can you comment on the below?: this still has the 50hz horizontal 'bar' ... but this appears to be over 25dB down, right? there is another bar at 100hz - is this the mains harmonic? what causes the bars at 60hz & 80hz? Room modes (5.8m x 3.6m x 2.8m)? decay seems to be acceptably steep ... all the frequencies have decayed before 240mS - except the bars from 50-120hz ... and below 30hz? so is this waterfall a good result? Thanks, Andy 25dB down is not enough for mains interference - ideally it should be buried in the noise floor 100Hz could be a mains harmonic (full wave rectification etc etc) could be modes at 60 and 80Hz taken from Paul Spencer's bass integration guide "Decay rate of 20 dB in the first 150 ms from 40 – 300 Hz." - I haven't looked too closely at yours - does it meet that? if not then it's not acceptably steep re-run the waterfall below 300Hz only - way too confusing full range. cheers Mike 1 Link to comment Share on other sites More sharing options...
davewantsmoore Posted March 26, 2017 Share Posted March 26, 2017 7 minutes ago, almikel said: 2 Maggies + 2 subs. Ah. I missed that. If the chart posted was of more than one speaker playing together, then that is the cause (I thought we'd covered this before a few times). Link to comment Share on other sites More sharing options...
almikel Posted March 26, 2017 Share Posted March 26, 2017 2 minutes ago, davewantsmoore said: Ah. I missed that. If the chart posted was of more than one speaker playing together, then that is the cause (I thought we'd covered this before a few times). only andyr's graphs not JVenters - he took your advice 2 Link to comment Share on other sites More sharing options...
almikel Posted March 26, 2017 Share Posted March 26, 2017 just for fun - this is Steven's ( @Listener ) measurement rig, with Alan Langford himself (Mr DEQX) driving the laptop in the chair (and Jiri in the foreground) No way I can replicate that myself - so I just use the measurements - gotta love the SNA community! Mike 2 Link to comment Share on other sites More sharing options...
andyr Posted March 26, 2017 Share Posted March 26, 2017 1 hour ago, almikel said: apologies, been a bit busy to post @andyr - I suggest you start a new thread on your issues to stop polluting @Jventer's thread. @Jventer apologies for adding further pollution to your thread. Yes, my sincere apologies to Jv - but as my thread would also be called "REW and Acoustics - help needed" if I had started my own thread, Mike ... you would be repeating on two threads. On just one, it seems to me, both Jv and I can learn from each other's questions and your/Dave's responses. 1 hour ago, almikel said: your excess phase plot looks all messed up - likely 'cos you're running both maggies at the same time. Run another measurement of 1 Maggie and all subs. Here's a pic of FR and Excess Phase in my room taken ages ago (sub + 1 x mid bass speaker) - levels not high enough (learning all the time). Yes, very nice, smooth, excess phase curve. My first attachment is the same pair of graphs for my setup - except with min phase + FR ... up to 400hz. (Sorry, can't see how to drag the attachment here.) This looks nice and smooth to me! But when I instead tick "Excess phase" ... it gets very nasty! (Second attachment.) But the vertical lines don't tie up with XO frequencies - so I can't see what the graph is telling me to do? 1 hour ago, almikel said: My recollection is that Mr Spencer set up your subs - so I'm confident the delay was set correctly... ...FYI this means the delay was applied to your mains to allow the subs to catch up - the low pass filter required for the subs creates delay and means that your mains need delay to allow the subs to catchup. Also, the subs are about 1.3m further away from my ears than the centre of the bass panels. 1 hour ago, almikel said: taken from Paul Spencer's bass integration guide "Decay rate of 20 dB in the first 150 ms from 40 – 300 Hz." - I haven't looked too closely at yours - does it meet that? if not then it's not acceptably steep re-run the waterfall below 300Hz only - way too confusing full range. Waterfall to 400hz is the third attachment. My reading of the graph is that it's not quite 20dB down by 150mS ... but it is by 180mS. Except for below 30hz. I'm fascinated to know what the "cocky's comb" (which is pronounced at 300hz) signifies. 42 minutes ago, almikel said: just for fun - this is Steven's ( @Listener ) measurement rig, with Alan Langford himself (Mr DEQX) driving the laptop in the chair (and Jiri in the foreground) No way I can replicate that myself - so I just use the measurements - gotta love the SNA community! Mike Great pic! You said "Unfortunately panels like Maggies are more difficult to measure outdoors than horns or box speakers, as the mike needs to be further away (eg min 3m) to capture the sound from the whole panel, plus you need a lot of space behind the panel (or lots of absorption) to avoid reflections from the backwave." You forgot to say that, with true-ribbon Maggies, you would run the risk of blowing out the 1.5m ribbon if there was a breeze! So I would never take my Maggies outdoors. Regards - and thanks, Andy 1 Link to comment Share on other sites More sharing options...
Recommended Posts