@vineethkumar01 - I'm learning too - all the time 😊 Thank you for keeping it open - Cause - If you got proof and a point - I'll adapt and pursue 👍
I just EQ'ed a set of JM-lab alto utopia BE yesterday for a friend. The speaker itself measure very good. But a bit of manual adjustment, removing a bit of the excessive presence around 5,5kHz and three rather big bass humps in this given room. Really made the speaker sound WAY nicer and smoother in a slightly reflective room. I measured several places in and around (50cm) the MLP, and then focused on the main issues. And by simply damping them around 70-80%... not fully.... because it is still an average of the total response - really made the speaker more coherent again. The speaker is smooth on and off axis to begin with, but all JM-lab speakers have a little bit extra energy in the 3-7kHz area - for some reason. And just a -2db with a Q around 2-3 in this area... really smooth out the overall response. Again, I can't change the speaker.... but an overall EQ can smooth out a bit of that excess total energy in this area. And after this... we listened to slow simple vocals from Leonard Cohen and Lisa Ekdahl, and now the voices were clear, relaxed and never stressing to listen to - even at high volumes. And this is that experience, where you finally can relax and forget about the system for a while, and just listen to music 🙂
I just EQ'ed a set of JM-lab alto utopia BE yesterday for a friend. The speaker itself measure very good. But a bit of manual adjustment, removing a bit of the excessive presence around 5,5kHz and three rather big bass humps in this given room. Really made the speaker sound WAY nicer and smoother in a slightly reflective room. I measured several places in and around (50cm) the MLP, and then focused on the main issues. And by simply damping them around 70-80%... not fully.... because it is still an average of the total response - really made the speaker more coherent again. The speaker is smooth on and off axis to begin with, but all JM-lab speakers have a little bit extra energy in the 3-7kHz area - for some reason. And just a -2db with a Q around 2-3 in this area... really smooth out the overall response. Again, I can't change the speaker.... but an overall EQ can smooth out a bit of that excess total energy in this area. And after this... we listened to slow simple vocals from Leonard Cohen and Lisa Ekdahl, and now the voices were clear, relaxed and never stressing to listen to - even at high volumes. And this is that experience, where you finally can relax and forget about the system for a while, and just listen to music 🙂
Earthworks are very good mics. I used them a lot as they are as useful for measurements as they are for tracking/recording when you want 'absolute' transparent sound... anyway i didn't say buy a binaural mic ( Neumann or Schoeps 'dummy head') i sqid you could use a binaural technique to record your clip.
It doesn't need a binaural dummy head: you can diy a pair of intraaural capsule and use your own head as a 'dummy' one.
Iirc this is how Mitchba ( the author of the article linked ) recorded the Kleinhorns ( soundclips availlable which obviously make headphone use mandatory).
Siegfried Linkwitz (RIP) used same thing to evaluate his own design: goes to a concert, listen/record it at his favourite location, then playback on his loudspeakers to evaluate how 'true' they were to his memory.
I think he described his own intra ear mic in his site.
@krivium: Thanks a lot.. This is all new information for me.. I just read the article by mitchba and took a look at the mics he used. I haven't heard about intraaural capsules.. Let me search and find out more details about them and get one if possible... 😀
I am still reading through all the information about DRC FIR and trying out things.. Hopefully, by the time I move into the next house (in a month), I have a better understanding of how all these works and apply some of these techniques there to recalibrate the system and if possible get some sound samples captured.. 🙂
I am still reading through all the information about DRC FIR and trying out things.. Hopefully, by the time I move into the next house (in a month), I have a better understanding of how all these works and apply some of these techniques there to recalibrate the system and if possible get some sound samples captured.. 🙂
Ah ah... you can approach it with a pair of glasses and some electret capsule:
https://www.linkwitzlab.com/Recording/AS_creation.htm
I think Griesinger used same technique too ( dumb i am... of course he did! He is one of reference used by Linkwitz!).
Vineeth that is a lot of info to digest. Just wanna point in direction you could follow if you are motivated enough about sharing things loudspeakers related through recording.
About FIR and 'tuning'/'voicing': you can't make reliable judgement on small time scale. I mean yes you can have huge improvement fast but when you want to fine tune things only comparison for long period will give you useful insight. Mater of weeks (with regular comparison of different profiles) ime.
https://www.linkwitzlab.com/Recording/AS_creation.htm
I think Griesinger used same technique too ( dumb i am... of course he did! He is one of reference used by Linkwitz!).
Vineeth that is a lot of info to digest. Just wanna point in direction you could follow if you are motivated enough about sharing things loudspeakers related through recording.
About FIR and 'tuning'/'voicing': you can't make reliable judgement on small time scale. I mean yes you can have huge improvement fast but when you want to fine tune things only comparison for long period will give you useful insight. Mater of weeks (with regular comparison of different profiles) ime.
Last edited:
@krivium: Indeed, there are a lot of things to learn for me... These days, every day, the only thing I really learn is that there are more new things to learn. 😀
You are very right about taking time to evaluate things..
At the moment, I can only differentiate between the really bad-sounding speakers (to my liking) and better-sounding speakers. I can detect whether the overall bass level is to my liking or not or whether the overall treble level is to my liking or not. Trying to differentiate between a good-sounding speaker and a really good-sounding speaker has been very very hard for me till now.. 😀
I will take time and keep learning things.. especially about the 'tuning' part. Audio-related stuff is very fascinating. The side advantage that I have is a lot of the signal processing involved is directly applicable to my day-to-day work too.. 🙂
You are very right about taking time to evaluate things..
At the moment, I can only differentiate between the really bad-sounding speakers (to my liking) and better-sounding speakers. I can detect whether the overall bass level is to my liking or not or whether the overall treble level is to my liking or not. Trying to differentiate between a good-sounding speaker and a really good-sounding speaker has been very very hard for me till now.. 😀
I will take time and keep learning things.. especially about the 'tuning' part. Audio-related stuff is very fascinating. The side advantage that I have is a lot of the signal processing involved is directly applicable to my day-to-day work too.. 🙂
Soft may be misleading as that is just a ready made preset, it falls in between minimal and normal on the authors scale. Measuring further away will mean that the direct to reverberant ratio in the measurement will be different. If it is windowed then you can control what is let in or shut out.
Room correction or any EQ applied to the source as a whole is 1D, so you are changing the on axis or predominant axis's frequency response. Directivity is unaffected by this so you can't turn a wide radiating tweeter into a laser beam. DRC should be viewed as being able to make improvements not turn water into wine.
You can view this like baffle diffraction. A speaker that has diffraction problems radiates differently in different directions and makes equalization tricky. Trying to fix an on axis dip results in an off axis peak and that does not usually make things better. The on axis may be better being compromised to improve the overall radiation behaviour. DRC can do this in a room. Some surfaces will cause a peak or dip in the frequency response and you might be able to find a better overall result in what you perceive at the listening position by changing the frequency balance. If you went back into an anechoic measurement that EQ would make the on axis look wonky but the combination of the speaker and the room together make that wonky response sound better. The problem with this is it is alchemy, the right answer is never the same twice so no hard core objectivist can accept it 🙂
The directivity of a speaker will determine what the on axis or listening axis should measure relative to it's overall power and directivity response. A speaker with consistently rising directivity sounds good with a flat on axis response. A speaker with a flat DI does not always sound good with the same flat on axis response. Changing that with global EQ to shelve the high end down could make it sound better and less bright.
So the other day, I was hearing some music on the speakers. I noticed some sibilance in female vocals in some tracks. I played many more songs and found out that this was highly recording dependent. Also switching off the room EQ slightly reduced the level of sibilance in some case. Even without room EQ there was easily easily hearable amount of sibilance on some tracks.. I have some questions. 😀
1) What is the reason behind sibilance?
2) Is it more speaker dependent or recording dependent..?
3) What are the mitigation strategies that we can put in place in the speaker design, if any techniques are available?
1) What is the reason behind sibilance?
2) Is it more speaker dependent or recording dependent..?
3) What are the mitigation strategies that we can put in place in the speaker design, if any techniques are available?
Difficult to say, attached an overview where in the spectrum amongst others are the sibilants.
It can be recording dependent for sure, but also some anomality in your speakers in that range, a small resonance f.i. , that gets excited by some sibilants.
Your speakers is what you can analyse, recordings not at all 😎
It can be recording dependent for sure, but also some anomality in your speakers in that range, a small resonance f.i. , that gets excited by some sibilants.
Your speakers is what you can analyse, recordings not at all 😎
Attachments
1) I do have an opinion on that, but it may not be a common view: I blame the Stereo cross talk for that. It exaggerates the ~3.7 KHz and ~7.2 KHz frequencies due to speaker position and head shape. A reflective room solves most of it, but if your speakers are more directional (less early reflections) this can become more noticable. Feel free to ignore my opinion though. Hearing sibilance in songs had always bothered me, I have no problems with it anymore after using the solution mentioned in No. 3.
2) I'd say it's very recording dependent, taking away that sibilance with EQ might make it sound as if someone has a lisp, it happens in songs... Some Michael Jackson tracks come to mind. Other examples do exist.
3) I use mid/side EQ with a bit of cross talk cancellation.
You could always insert a De-Esser if you're playing digitally. But keep in mind what I said about overdoing the use of a de-esser.
2) I'd say it's very recording dependent, taking away that sibilance with EQ might make it sound as if someone has a lisp, it happens in songs... Some Michael Jackson tracks come to mind. Other examples do exist.
3) I use mid/side EQ with a bit of cross talk cancellation.
You could always insert a De-Esser if you're playing digitally. But keep in mind what I said about overdoing the use of a de-esser.
Excessive S sounds occur around the 8th octave, around 8KHz usually related to singer being too close to the mic, lack of mic filter or windscreen or mic type and placement. Obviously it’s better to fix it at the source but if it was baked in, in the mastering stage a de-esser aka multi band compressor or EQ may be used to reduce it.
For a speaker designer the main thing you can do is to sure there’s not too much around 4 to 8KHz. For those who don’t design speakers, toeing speakers out may help.
The interesting thing you will find is that just a dB and a half can reduce the sibilance, but if you go too much but then other recordings feel like they lack that but clarity or can the speakers can even sound a bit dull.
A lot of audiophoolery goes into switching cables and amps etc, which is a very expensive way to do it.
I would turn on the REW RTA mode and click REC snd play the offending music and see what you’re hearing. If the treble looks highly, even by a dB, and it sounds high, then it probably is high…
For a speaker designer the main thing you can do is to sure there’s not too much around 4 to 8KHz. For those who don’t design speakers, toeing speakers out may help.
The interesting thing you will find is that just a dB and a half can reduce the sibilance, but if you go too much but then other recordings feel like they lack that but clarity or can the speakers can even sound a bit dull.
A lot of audiophoolery goes into switching cables and amps etc, which is a very expensive way to do it.
I would turn on the REW RTA mode and click REC snd play the offending music and see what you’re hearing. If the treble looks highly, even by a dB, and it sounds high, then it probably is high…
Last edited:
Or use Audacity to plot the spectrum 🙂
You can select just the portion of the track where the problem is to make things easier
You can select just the portion of the track where the problem is to make things easier
If it's not something you experience continously then your loudspeakers are not to blame, 'au contraire' if they display sibilance on tracks you never noticed before then they may be better than your previous one.
Reasons? There is a lot: first the source. Like tktran303 pointed recording techniques can be involved, mixing technique too ( too much compression used usually makes sibilance being proeminent in signal, obvious in radio/tv). Some singers have sibilance too, it's part of their natural voice.
Mitigation strategies? Well i see things differently, i do not try to compensate for flaws into recording: as i've been part of the production process i know it can be an artistic choice to let them go thru, or a technical compromise/tradeoff. If the recording is like that then it is how the artist wanted it to be, who am i to counteract their own vision of their art?
Wouldn't it be strange to repaint 'la Joconde' in redish color because i don't like how Leonardo Da Vinci used too much brown and this hazy effect on background? 😉
I consider it's the artist/recording/mixing/mastering engineer which could do something to mitigate the effect.
I will give you a famous reference where it was successful: listen to Massive Attack's 'Protection' track.
Tracey Thorn have a real issue with sibilance, it can be really bothersome in his own band Everything But The Girl. In this Massive attack track the mixer decided to use this to his advantage ( because the track and the way she sing gives the opportunity too): instead of using a de-esser he made the choice to use the sibillance to be sent into a big lush reverb and trigger them relative to the level the ss have.
It just 'open' the scene in width and depth...
Very very clever.
Try to find a good quality recording of it mp3 tend to eat the reverb details.
This track:
Look for an evaluation copy ( bootleg) of Izotope's RX.
There is dedicated plug ins too but RX is the industry workhorse. In my experience can be even better than Cedar.
There is dedicated plug ins too but RX is the industry workhorse. In my experience can be even better than Cedar.
Last edited:
Thanks @JanRSmit, @wesayso,@tktran303,@fluid, and @krivium
This is a lot of new information I have to process and apply to learn more about the sibilance thing.. 😀
I will start this exercise today itself.. I checked the quasi anechoic measurement of the speaker. This is how it looks
Horizontal polar:
There is definietly something (probably having a minute effect??) going on around 7.2kHz with the BMS 4550 driver I think. It looks like a small resonance as it is visible even in the sound power response.
Looking at the EQed speaker response at MLP and 30cm away in the horizontal plane to the left of MLP doesn't show much of an effect, I think (The speakers are toed in about 40 ish degrees though):
@wesayso: Could you please explain a bit more about 'mid/side EQ with a bit of cross-talk cancellation'. What is it, and how is it implemented?
In the meanwhile, I am going to hunt the spectrum of the song I heard and see if something shows up.. 😀
This is a lot of new information I have to process and apply to learn more about the sibilance thing.. 😀
I will start this exercise today itself.. I checked the quasi anechoic measurement of the speaker. This is how it looks
Horizontal polar:
There is definietly something (probably having a minute effect??) going on around 7.2kHz with the BMS 4550 driver I think. It looks like a small resonance as it is visible even in the sound power response.
Looking at the EQed speaker response at MLP and 30cm away in the horizontal plane to the left of MLP doesn't show much of an effect, I think (The speakers are toed in about 40 ish degrees though):
@wesayso: Could you please explain a bit more about 'mid/side EQ with a bit of cross-talk cancellation'. What is it, and how is it implemented?
In the meanwhile, I am going to hunt the spectrum of the song I heard and see if something shows up.. 😀
^ you open a can of worms! 😉
Pano have a thread about it where Wesayso exposed most of his view and experiment ( as well as other members of course).
Edit:
https://www.diyaudio.com/community/threads/fixing-the-stereo-phantom-center.277519/
I know his own solution is not easy* to implement but worth an experiment/try about it. There is more simple way to implement than his own way ( which is highly dependent of his own loudspeaker and room so probably not easily implemented in other situation).
Fluid could explain too as 'mon petit doigt me dit' ( my little finger tells me ) they exchanged about it extensively.
Anyway from Wesayso's description gave me, i'm sure he managed some impressive results mimicing a center ( as in L/C/R) with a 2.2 setup ( well not really 2.2 as there is 'ambiance' loudspeakers too... it's a whole class aside other approach in my view) .
* it's not difficult per se, but need a complex digital routing implementation before the amps ( software). Even if i like this kind of thing i got headache looking at the general implementation he gave. And scratched my head trying understand what he does. Really advanced technique in my view. 🙂
Pano have a thread about it where Wesayso exposed most of his view and experiment ( as well as other members of course).
Edit:
https://www.diyaudio.com/community/threads/fixing-the-stereo-phantom-center.277519/
I know his own solution is not easy* to implement but worth an experiment/try about it. There is more simple way to implement than his own way ( which is highly dependent of his own loudspeaker and room so probably not easily implemented in other situation).
Fluid could explain too as 'mon petit doigt me dit' ( my little finger tells me ) they exchanged about it extensively.
Anyway from Wesayso's description gave me, i'm sure he managed some impressive results mimicing a center ( as in L/C/R) with a 2.2 setup ( well not really 2.2 as there is 'ambiance' loudspeakers too... it's a whole class aside other approach in my view) .
* it's not difficult per se, but need a complex digital routing implementation before the amps ( software). Even if i like this kind of thing i got headache looking at the general implementation he gave. And scratched my head trying understand what he does. Really advanced technique in my view. 🙂
Last edited:
Good points everyone 🙂
While I suspect its just in the tracks, some are more sibilant than others, its an opportunity to make some experiments and possibly learn something. You are now clearly aligned to sibilance so motivation is there.
You could try sibilant track with single speaker mono setup to test how it compares to stereo. You could try zone in to the frequency(ies) you are sensitive to with EQ. I seem to be quite sensitive to ~12kHz for some reason and found myself always dipping that if mixing, for some reason. You could try with headphones if the same tracks are sibilant there, is it nicer or worse? You could try with multiple playback levels, it might be non-linearity in your hearing? It might be some resonance in the driver that is audible, even though frequency response looks kind of clean, is it the same with other compression driver / waveguide? You could EQ the system, play with a lowpass filter, or narrow notch filter. You could try small inductor in series to make passive low pass filter, or just a resistor (pad) to test if (now somewhat reduced) distortion has something to do with it. The anechoic response seems to peak >15kHz so harmonic distortion of the sibilant bandwidth around 5-8kHz would land there and get amplified. Are the tracks sibilant to you if you listen with your cellphone speaker?
It could be anything, your hearing, something in the playback system, something in the tracks or multiple things combined. If the issue draws your attention it takes the attention away from the music so I believe it would pay out to zone in what it is and then try to do something about it. I would just play with EQ if nothing else works and then decide if the issue deserves to be fixed or is fine. The goal going through it all is to get peace of mind so that it doesn't draw the attention anymore 🙂
While I suspect its just in the tracks, some are more sibilant than others, its an opportunity to make some experiments and possibly learn something. You are now clearly aligned to sibilance so motivation is there.
You could try sibilant track with single speaker mono setup to test how it compares to stereo. You could try zone in to the frequency(ies) you are sensitive to with EQ. I seem to be quite sensitive to ~12kHz for some reason and found myself always dipping that if mixing, for some reason. You could try with headphones if the same tracks are sibilant there, is it nicer or worse? You could try with multiple playback levels, it might be non-linearity in your hearing? It might be some resonance in the driver that is audible, even though frequency response looks kind of clean, is it the same with other compression driver / waveguide? You could EQ the system, play with a lowpass filter, or narrow notch filter. You could try small inductor in series to make passive low pass filter, or just a resistor (pad) to test if (now somewhat reduced) distortion has something to do with it. The anechoic response seems to peak >15kHz so harmonic distortion of the sibilant bandwidth around 5-8kHz would land there and get amplified. Are the tracks sibilant to you if you listen with your cellphone speaker?
It could be anything, your hearing, something in the playback system, something in the tracks or multiple things combined. If the issue draws your attention it takes the attention away from the music so I believe it would pay out to zone in what it is and then try to do something about it. I would just play with EQ if nothing else works and then decide if the issue deserves to be fixed or is fine. The goal going through it all is to get peace of mind so that it doesn't draw the attention anymore 🙂
Thanks, @tmuikku, for the suggestions.
I think the sibilance that I was thinking about is there in the tracks themselves. I can hear it clearly through my truthear crinnacle headphones. 😀
For example, there is one instance (and several other instances throughout the same track) in the video clip here.
The 'ssss' sound around 32-33 seconds:
I tried to plot the spectrum for the 3 second duration in which the above sound was heard. It looked like below:
I am not exactly sure what to interpret from the above graph.
There are several other songs too which suddenly draws my attention to these kinds of sounds.
Playing this track on the speaker is like hearing the same thing but all imperfections (?) slightly magnified. Its like I am hearing all sorts of imperfections or recording artefacts in the recording... I have no idea whether it is good or bad. But now I don't like playing this song on the speakers with their current settings. 😀
But there are several other songs where I actually like what I am hearing.. 😀
I need to try out the mono speaker experiment now
I think the sibilance that I was thinking about is there in the tracks themselves. I can hear it clearly through my truthear crinnacle headphones. 😀
For example, there is one instance (and several other instances throughout the same track) in the video clip here.
The 'ssss' sound around 32-33 seconds:
I tried to plot the spectrum for the 3 second duration in which the above sound was heard. It looked like below:
I am not exactly sure what to interpret from the above graph.
There are several other songs too which suddenly draws my attention to these kinds of sounds.
Playing this track on the speaker is like hearing the same thing but all imperfections (?) slightly magnified. Its like I am hearing all sorts of imperfections or recording artefacts in the recording... I have no idea whether it is good or bad. But now I don't like playing this song on the speakers with their current settings. 😀
But there are several other songs where I actually like what I am hearing.. 😀
I need to try out the mono speaker experiment now
Oh jeah lot of modern music sounds like this, artefacts from all the processing. Perhaps best listened from small bluetooth speaker 🙂 If you hear it and it distracts you but still like the music, I would just EQ it out, roll off or notch out the excess top end. Do you have presets in your DSP with easy switching?
edit. the spectrum analysis looks like there is deliberate boost with EQ around 11k, so the sound is in the "production"
edit. the spectrum analysis looks like there is deliberate boost with EQ around 11k, so the sound is in the "production"
Last edited: