davewantsmoore Posted October 4, 2018 Posted October 4, 2018 19 hours ago, davewantsmoore said: Move the speaker (just one speaker) into the 'middle' of the room.... away from as many things as you can. Move at least one of the couches. Move 1 speaker into the middle of the room... take 2 measurements on tweeter axis, one from 20cm ish, and one from 1m+ ish ... use 1/6oct smoothing (or less, ie. 1/12) .... provide rough pic to show speaker and microphone position in room, and how far the couches got moved away, etc. This type of 'baseline' will then start to point at 'what could it be' (ie. what to investigate next).
audio_file Posted October 5, 2018 Author Posted October 5, 2018 I thought a good way to test wall reflections behind the speaker is to open the sliding door behind them, that whole wall is a window with a sliding door in the middle. This way there can be no reflections at least from the middle of that wall.
audio_file Posted October 5, 2018 Author Posted October 5, 2018 The ceiling is 251 mm high, and has a rough texture, full of little stones. The floor has is carpeted.
davewantsmoore Posted October 5, 2018 Posted October 5, 2018 14 minutes ago, audio_file said: I thought a good way to test wall reflections behind the speaker is to open the sliding door behind them, that whole wall is a window with a sliding door in the middle. This way there can be no reflections at least from the middle of that wall. Don't worry about any of this stuff yet. Move one speaker as far away from walls as you can.... shut the doors, etc. FWIW - The dips seem too high in frequency to be from path length cancellations from walls or floor, etc.... but you need to get a baseline first, before there is any idea which way to step next. 8 minutes ago, audio_file said: The ceiling is 251 cm high Just rough distances is all that's needed at this stage, when you do a new measurement (or two measurements, 20cm and 1m+ distance) Mic (on tweeter axis) height above floor Mic distance from tweet Speaker rough distances to walls Also remove mic and soundcard calibrations for now You may also need to test running the input and output through the same sound device (ie. leaving the chord dac out of it) and using the behringer inputs and outputs.
audio_file Posted October 5, 2018 Author Posted October 5, 2018 11 hours ago, almikel said: so all those different speakers had the same dip at 500Hz? with the mike in the same position?...would have to be SBIR - too high for modal - but I would have expected the frequency to change with mike/speaker position if SBIR??? Too tricky for me... @davewantsmoore this is why I mentioned opening the sliding door, not because of what you asked.
audio_file Posted October 5, 2018 Author Posted October 5, 2018 38 minutes ago, davewantsmoore said: Also remove mic and soundcard calibrations for now You may also need to test running the input and output through the same sound device (ie. leaving the chord dac out of it) and using the behringer inputs and outputs. I had a look at the Behringer documentation and it looks like I can use the output and input this way after all. I'll try it. There was never a mic calibration, only a sound card calibration. 1
davewantsmoore Posted October 5, 2018 Posted October 5, 2018 10 minutes ago, audio_file said: @davewantsmoore this is why I mentioned opening the sliding door, not because of what you asked. Yes, I understand .... I'm saying don't worry about that for now.
Ando Posted October 5, 2018 Posted October 5, 2018 5 minutes ago, audio_file said: I had a look at the Behringer documentation and it looks like I can use the output and input this way after all. I'll try it. There was never a mic calibration, only a sound card calibration. I have the same set up using Behringer in and out. W hat do you use for mic pre amp ? My mic preamp has some EQ functions. Maybe worth checking ? Cheers Mike
audio_file Posted October 5, 2018 Author Posted October 5, 2018 1 hour ago, Ando said: I have the same set up using Behringer in and out. W hat do you use for mic pre amp ? My mic preamp has some EQ functions. Maybe worth checking ? Cheers Mike I use the UD202HD as mic pre amp, nothing else.
Ando Posted October 5, 2018 Posted October 5, 2018 42 minutes ago, audio_file said: I use the UD202HD as mic pre amp, nothing else. Apologies for my useless comment. I completely misread what you wrote. 1
Jventer Posted October 5, 2018 Posted October 5, 2018 @audio_file I am going through a similar process - a bit derailed as I had to move house. There are fantastic guys here trying to help, e.g @davewantsmoore. I am unfortunately confused. Are you still referring to your original measurements or have you done new measurements as suggested, e.g. Move at least one of the couches. Move 1 speaker into the middle of the room... take 2 measurements on tweeter axis, one from 20cm ish, and one from 1m+ ish ... use 1/6oct smoothing (or less, ie. 1/12) .... provide rough pic to show speaker and microphone position in room, and how far the couches got moved away, etc. or have you done tests outside? If you have not done measurements per single speaker inside or outside we cant move ahead. Thanks
audio_file Posted October 5, 2018 Author Posted October 5, 2018 I haven't had the time to do any more.
audio_file Posted October 5, 2018 Author Posted October 5, 2018 This is the sound card calibration graph that was used for reference:
Jventer Posted October 5, 2018 Posted October 5, 2018 11 hours ago, audio_file said: I haven't had the time to do any more. No worries, fully understand. I have 90kg each speakers and it is a pain to do, hence my own hiatus. It may be a room, calibration or measurement issue, but we dont really know. In my limited experience: make sure the microphone is calibrated properly find the expected performance of the specific loudspeaker (if possible) measure the speaker outside/inside where there is minimal influence taking care with microphone measurement check everything, e.g. is measurement per expectations, if not start fault finding. then once the previous bits are done satisfactorily go to next steps in meauring speakers one by one in room, with dfferent locations etc. Hope this make sense. All the best.
davewantsmoore Posted October 6, 2018 Posted October 6, 2018 (edited) On 05/10/2018 at 10:21 AM, davewantsmoore said: remove mic and soundcard calibrations Also consider setting up using the Behringer only (understand this might not be easy to do immediately) .... as assuming the wiggles above 500Hz are still present in your 'baseline' measurement, then the next steps will be looking for other technical problems (which are easy to have when you have one audio device as the output and a completely different device as the input) 4 hours ago, Jventer said: speaker outside The wiggles in the measurement are high enough in frequency to be excluded from the data (if needed) using a window function. ... but for now, the goal is just to confirm that the measurement setup is sane - as the wiggles (which should not be there) are showing up for various different speaker and amps - so something is wrong. For now, just need to get one speaker (doesn't matter which) as far away from walls as possible, and take a sweep at a couple of different distances. Edited October 6, 2018 by davewantsmoore
audio_file Posted October 6, 2018 Author Posted October 6, 2018 I do plan to try this, but I have maybe 10 minutes free time a day with my kids and my work.It took me a long time to take the above measurements. I might have to take a day's leave. 1
audio_file Posted October 8, 2018 Author Posted October 8, 2018 On 03/10/2018 at 10:17 PM, audio_file said: Short vs long sweep: This graph remains the most interesting to me, position of speaker and mic remained the same, SPL, sound card, speakers, amplifiers, etc. all remained constant, yet the result is so different. The only difference was the frequency delay, due to using a longer sine wave to measure vs a shorter one. (256K vs 1M on REW) The blue line is the 256K one and the reddish line the 1M one. If you look at the 1M sweep, the upper mid and treble from around 1.3kHz to 10kHz are +- 1 dB, which is great. If you include the part from 650Hz then it gets to around +- 3.5dB. I have a feeling REW doesn't check the frequency picked up by the mic, but rather matches the level picked up by the mic to the frequency played at the same time by the REW sine sweep. If this is true, the explanation for the above would be that the bass lingers so long that those dips and peaks in the mid range might actually be partly due to leftover bass frequencies cycling back and forth in my room. By lengthening the sine sweep, the bass had time to dissipate more when REW got to the mid range. I'm going to test this by doing a sine sweep from 200Hz up, because that would exclude the bass frequencies from the equation. This is in addition to all the other tests I plan to do so far.
audio_file Posted October 9, 2018 Author Posted October 9, 2018 Looks like I might get a chance to do more measurements on Tuesday.
Hipper Posted October 10, 2018 Posted October 10, 2018 I think REW just listens to what arrives at its position. If for example you have some external sounds (e.g. traffic noise, amp hum) they will also be seen on the FR graph. Indeed you can record ambient noise without playing the sine wave (you can do this by measuring but turning the volume of your speakers off). You can check on reverberation by looking at the waterfall plot or spectrogram.
audio_file Posted October 10, 2018 Author Posted October 10, 2018 That makes sense and supports my suspicion. I think REW just listens to what arrives at its position. If for example you have some external sounds (e.g. traffic noise, amp hum) they will also be seen on the FR graph. Indeed you can record ambient noise without playing the sine wave (you can do this by measuring but turning the volume of your speakers off). You can check on reverberation by looking at the waterfall plot or spectrogram.
audio_file Posted October 10, 2018 Author Posted October 10, 2018 (edited) What does everyone else use for their sine wave? The 1M one, the 256K one, or a different one? Also is there a way to play it in the other direction, from high to low? Edited October 10, 2018 by audio_file
davewantsmoore Posted October 11, 2018 Posted October 11, 2018 3 hours ago, audio_file said: What does everyone else use for their sine wave? The 1M one, the 256K one, or a different one? Don't worry about this too much. Choosing the faster one is more convenient (stick with that for now) ..... if you are really looking for something in particular, then the longer sweeps can be used. 3 hours ago, audio_file said: Also is there a way to play it in the other direction, from high to low? No - what are you trying to achieve?
audio_file Posted October 11, 2018 Author Posted October 11, 2018 6 minutes ago, davewantsmoore said: No - what are you trying to achieve? It would just be interesting to see how it changes if it is indeed lingering frequencies affecting later frequencies.
davewantsmoore Posted October 11, 2018 Posted October 11, 2018 Ahh, I see.... I've replied out of order, and should have done this one first. On 08/10/2018 at 12:34 PM, audio_file said: This graph remains the most interesting to me, position of speaker and mic remained the same, SPL, sound card, speakers, amplifiers, etc. all remained constant, yet the result is so different. The only difference was the frequency delay, due to using a longer sine wave to measure vs a shorter one. (256K vs 1M on REW) The blue line is the 256K one and the reddish line the 1M one. No... It wasn't. Something else was different. The sweep length doesn't cause differences that large. If you think otherwise (which is good - don't take anyones word for anything) then you can carefully repeat the measurement just changing the length - and you will see that you can't reproduce this result On 08/10/2018 at 12:34 PM, audio_file said: I have a feeling REW doesn't check the frequency picked up by the mic, but rather matches the level picked up by the mic to the frequency played at the same time by the REW sine sweep. If this is true, the explanation for the above would be that the bass lingers so long that those dips and peaks in the mid range might actually be partly due to leftover bass frequencies cycling back and forth in my room. By lengthening the sine sweep, the bass had time to dissipate more when REW got to the mid range. No, that's not how it works... On 08/10/2018 at 12:34 PM, audio_file said: This is in addition to all the other tests I plan to do so far. Don't over complicate this - by testing for, and drawing conclusions from, things you don't understand. Just do a couple of basic sweeps carefully as suggested, and that information will tell us what to do next. 1
davewantsmoore Posted October 11, 2018 Posted October 11, 2018 2 minutes ago, audio_file said: It would just be interesting to see how it changes if it is indeed lingering frequencies affecting later frequencies. You need to forget trying to explain your data, and focus instead on whether your data is valid. It is very unlikely to be. As you can imagine, trying to explain flawed data will lead to incorrect conclusions. 1
Recommended Posts