IEMs, Singers and Occlusion

Continuing on from this excellent thread...
I think this is a great subject that could use a lot more discussion and solutions.

So, I only recently started seriously mixing for In-Ears. I've been a FOH engineer forever, and my monitor work typically was for wedges and only occasional IEMs. Never really studied different players mixes, just gave them what they want, or created something if they weren't specific about what they wanted.

So what I've been noticing is the huge variance in how much vocal each IEM user likes to have to feel comfortable. Some like their vocal just a little over the mix as you would expect. Others want 6 to even 12db or more of their vocal over top of their mix. It's almost unlistenable to cue up their mixes and there's no way I could listen to it at the volume level they like it. These are all players with custom moulded IEMS, mostly good ones like UE-11 or the like. I'm not including anyone who has unmoulded IEMs in this discussion as they aren't really taking it seriously and their mix could be all over the place depending on the fit.

So, since there's no way for me to REALLY hear what they're hearing, I gotta assume that they simply can't hear their vocal within their mix due to occlusion? I'm not a singer, but I know if I put in my IEMs and talk into a mic, I hear a mix of my voice and what I call my "inside head" voice which is all low frequency. The same sound you hear if you just plug your ears.
Now, if I push my In-Ears in as far as they will go (touching what feels like bone, and it's not very comfortable) I can get to a point where the low "inside head" voice goes away and I hear a brighter sound from the mic. Does this mean that my IEMs are not long enough? They don't sit in this "tighter" position on their own and will fall back to a more comfortable position in the ear once I stop pressing on them.
Is this what occlusion sounds like?

My concern is the huge volume level of these singer's vocals in their ears that they're "not hearing" and the damage that can do. Additionally, their mix isn't as clean as it could be because of all the ambient sound their vocal mic is picking up having their vocal mic so much hotter.

I've searched around and actually the linked thread has the most relevant information about this, and there's not much else on the net about it.

I don't really think that latency is the cause here as I've done some experiments with a straight analog-only path for the vocal mics to the IEMs and still hear the same effect.

My questions are:
Is this Occlusion that I'm hearing when I talk into a mic, and is it being eliminated by pressing my In-Ears in further into my head?
Are the players who need their vocal ridiculously loud suffering from a poor IEM mould?
 
Re: IEMs, Singers and Occlusion

I am a singer and a mixer. I have small ear canals and can't get a fit deep enough to kill the occlusion. Same with ear plugs. I have a hard time mixing with anything but small flanged plugs just slightly inside the ear. I have wondered what kind of process IEM mixers go through to get a solid reference on the different performers mixes. I am thinking I would almost prefer being totally isolated and listen through studio monitors.
 
Re: IEMs, Singers and Occlusion

Continuing on from this excellent thread...
So what I've been noticing is the huge variance in how much vocal each IEM user likes to have to feel comfortable. Some like their vocal just a little over the mix as you would expect. Others want 6 to even 12db or more of their vocal over top of their mix.

So, since there's no way for me to REALLY hear what they're hearing, I gotta assume that they simply can't hear their vocal within their mix due to occlusion?

Now, if I push my In-Ears in as far as they will go (touching what feels like bone, and it's not very comfortable) I can get to a point where the low "inside head" voice goes away and I hear a brighter sound from the mic. Does this mean that my IEMs are not long enough?
Is this what occlusion sounds like?

I don't really think that latency is the cause here as I've done some experiments with a straight analog-only path for the vocal mics to the IEMs and still hear the same effect.

My questions are:
Is this Occlusion that I'm hearing when I talk into a mic, and is it being eliminated by pressing my In-Ears in further into my head?
Are the players who need their vocal ridiculously loud suffering from a poor IEM mould?
Andrew,

From my experience mixing studio headphones, 6 dB of "more me" vocal seems fairly normal, 12 dB sounds like a problem is being overcome. That said, even in an all analog format, there are many musicians I've worked with that listen at a level 10 dB or more louder than I would, not happy until the headphone amp is clipping.

Pushing your in ear in further will reduce the cavity size between the ear drum and transducer, which may in your case may result in more HF due to less cancellation, as well as being louder due to the inverse distance law.
I'm no expert on ear mold fit, so won't say if that means yours need to be longer, obviously they should not be long enough to puncture your ear drums, but too much space probably is not good either.

Occlusion (the sound you hear when your ears are plugged) and latency are two different but related issues, I explained the latency issue in the linked thread.
Both can be overcome with level, but for my taste, the latency issue would cause me to want to listen at far too high a level.

As far as your straight analog-only path for the vocal mics to the IEM, did you try this with your vocal or their vocals?
Does your IEM have latency?
Proper mic polarity is important for the in ear vocalist (no one else will notice a difference) and between console, RF mic and receiver, in ear transmitter and receiver, and in ear wiring there are a lot of places where a switch can happen.
Did you try switching polarity?

Art
 
Re: IEMs, Singers and Occlusion

Thanks, Art!

Just for clarity, I was thinking to 6-12db louder than what other "more normal" vocalists wanted, but wrote "over their mix"...

The funny thing was, listening to my own voice back and forth between an analogue path and digital, didn't get rid of the "inside head" voice. It may have sounded a little different, but I didn't spend enough time matching the eqs so it wasn't a really "scientific" experiment. I haven't tried the analogue path with a singer yet. It would be problematic to do a show that way due to the issue with accurately monitoring their mix. I tried switching polarity on my own mic and it made no difference. I also played with delaying the mic further and it changed the sound, but never got rid of that "inside head" voice.

I would have expected the latency to show up as a "flanged" sort of effect or hollowness, which is not what I'm hearing. The console I'm testing on is a PM5DRH (2.5ms latency @ 48k).

Again, I really don't know what the singers are actually hearing and why they need their vocal so much louder to "hear" it, but my tests lead me to believe they're experiencing a similar problem to what I'm hearing.
 
Re: IEMs, Singers and Occlusion

Thanks, Art!

Just for clarity, I was thinking to 6-12db louder than what other "more normal" vocalists wanted, but wrote "over their mix"...

The funny thing was, listening to my own voice back and forth between an analogue path and digital, didn't get rid of the "inside head" voice. It may have sounded a little different, but I didn't spend enough time matching the eqs so it wasn't a really "scientific" experiment. I haven't tried the analogue path with a singer yet.
I would have expected the latency to show up as a "flanged" sort of effect or hollowness, which is not what I'm hearing. The console I'm testing on is a PM5DRH (2.5ms latency @ 48k).
"Normal" (or "more normal") is relative, when working clubs I found 110 dBA at FOH was not uncommon, with stage levels of 120+ dBA, that's 30+ dB louder than what I want to hear now.
The vocalist may just like the loud "feel", or be deaf, or want to get over the occlusion/latency problems, or a combination of all three.

The "inside head" voice will always be there with in ear monitors, but matching their phase to the to the "inside head" voice is only possible with no latency, or less latency than is currently possible with the "state of the art", especially when digital RF latency is in series with the console latency.
Once 2.5 MS+ latency is introduced, I don't think mic polarity would be detectable.

People are sensitive to different things, many have grown up on digital and when hearing their voice in ears or headphones they have never heard it without latency, so that's the only reference. You can get used to anything, some artists still don't use monitors at all.

Others can't stand the sound of latency combined with their "inside head" voice, there is a good reason why Bono is using a Midas Venice for his in ears while they have a pair of XL8 serving the other guys and FOH.
 
Re: IEMs, Singers and Occlusion

"Normal" (or "more normal") is relative, when working clubs I found 110 dBA at FOH was not uncommon, with stage levels of 120+ dBA, that's 30+ dB louder than what I want to hear now.
The vocalist may just like the loud "feel", or be deaf, or want to get over the occlusion/latency problems, or a combination of all three.
Again, for the sake of clarity, I'm speaking about the BALANCE of their vocal vs. the rest of the mix, not the overall level. There's no correlation so far with overall level and how far above the mix they want their vocal. Limiters are off, BTW, before anyone asks...

The "inside head" voice will always be there with in ear monitors, but matching their phase to the to the "inside head" voice is only possible with no latency, or less latency than is currently possible with the "state of the art", especially when digital RF latency is in series with the console latency.
Once 2.5 MS+ latency is introduced, I don't think mic polarity would be detectable.

People are sensitive to different things, many have grown up on digital and when hearing their voice in ears or headphones they have never heard it without latency, so that's the only reference. You can get used to anything, some artists still don't use monitors at all.

Others can't stand the sound of latency combined with their "inside head" voice, there is a good reason why Bono is using a Midas Venice for his in ears while they have a pair of XL8 serving the other guys and FOH.
Understood.

I think the next test I will do is play them back a recording of their monitor mix and see if they feel that the mix they hear is the same as when they're on stage or if they notice the vocal to be louder than normal. I at least need to see if they in fact are not hearing the mix that's being sent to them - i.e. some of their vocal level is being "cancelled out" due to latency, occlusion or other factors...
 
Re: IEMs, Singers and Occlusion

Okay, i'll weigh in. I'm a retired Sound Guy that now holds a job at a church and uses custom IEMs as a singer/guitar player. i also run the monitor rig for group and I have experienced the same things you describe in working with my various singers. my singers control their own mix, but if there is a technical issue, i will often take their beltpack from them so i can hear what they hear and try to fix whatever isn't working correctly. in doing this, i am often amazed at how wacky some of their mixes are. i'm also often amazed at how Quiet a lot of them have their volume levels.

i think a LOT of what you are describing is Occlusion Effect. Different people experience this differently depending on the shape of their ear canal as well as how deep their IEMs go into their ears. the tiniest difference will change what the resonance frequency of that space is considerably. The person's skull shape and other things also will change how this sounds.

in my experience, the only way to get over the Occlusion effect is to make the sound of your vocal louder in your ears than the muddy resonance. If you run your overall mix fairly loud, then your vocal might still be only 3 dB over the rest of the band. This is how i run my mix. I run it fairly hot, but pretty well balanced. I'm not killing myself. No ringing ears, etc. after playing. But i've got the level pretty well up there.

However, if you don't tend to run the level of your mix that loud, there might be a tendency to still want that vocal level hot enough to get over the occlusion. That can lead to the 6-12 dB over the mix scenario that you describe. Keep in mind that if you tend to run your mix fairly quiet, you're going to hear significantly more of the stage noise as well, so what you may hear as their wildly out of balance mix may be quite different to them since their hearing quite a bit of stage bleed mixing with that mix.

All of these comments assume you're talking about folks using RF IEMs, since those people ultimately set the level in their IEMs themselves, and you can never really know how loud that is, even if you take their beltpack from them and listen yourself as your ear physiology is likely quite different and your IEM drivers may be very different as well.

Now latency is a whole 'nother bucket of worms. But frankly, my experience has been that various latencies do little more than change the resonance frequency in my head. so long as i have the level up loud enough to overcome that resonance, it's really not a problem. But i'm sure there are folks for whom this is a MAJOR issue [and not just Bono]. I know a lot of folks that are running parallel analog mic channels [Mic goes into analog mixer, monitor mix minus Mic goes into analog mixer, output goes to IEMs] so that at least the vocal mic doesn't have any latency issues, even if the rest of the mix is coming from a digital console. i haven't felt like i needed to do that yet. But i may still try it just to see if i like it better.
 
Re: IEMs, Singers and Occlusion

So maybe inverting polarity on the singer's own voice, changing the cancellation wrt to the occlusional sound, could have a good effect on the perceived sound.

My experience with doing this has been totally random. For my voice, ears, and monitor rig flipping polarity made things sound different but not better or worse. i've a couple singers who really like their vocal mic polarity flipped [they don't know that's what i did. just that i did something]. and others have hated it. it all depends on how much latency, where the occlusion effect resonance is, and what the singer's perceptions are. other than those couple of variables it's pretty consistent. :)
 
Re: IEMs, Singers and Occlusion

My experience with doing this has been totally random. For my voice, ears, and monitor rig flipping polarity made things sound different but not better or worse. i've a couple singers who really like their vocal mic polarity flipped [they don't know that's what i did. just that i did something]. and others have hated it. it all depends on how much latency, where the occlusion effect resonance is, and what the singer's perceptions are. other than those couple of variables it's pretty consistent. :)

Good to know. I already figured that polarity changes would be in the toolkit of Things That Might Help. Too many variables in this equation!
 
Re: IEMs, Singers and Occlusion

Well, unfortunately, this subject is still not resolved for me yet.

In speaking to one VERY experienced IEM engineer who also makes IEMs, he says that the occlusion effect can only be reduced by ambient moulds. I asked about the length of tube, and he doesn't feel that this is a factor. Only fit (unsealed moulds) will alleviate it. I was quite surprised to hear this from an expert in the field.

In the very little that is written about occlusion and in-ears, they all seem to claim while venting works, occlusion can also be eliminated by having the mould go in far enough into the ear. And without the low-frequency loss and lack of isolation you get from ambients.

Most research that I can find is regarding in-ear hearing aids. I couldn't find any real research, just anecdotal evidence for musician/singer IEM/Occlusion issues.

e.g. RERC on Hearing Enhancement - Dr. Ross Says...
 
Re: IEMs, Singers and Occlusion

Continuing on from this excellent thread...
I think this is a great subject that could use a lot more discussion and solutions.

Hi Andy,
I see from further down the thread, that you are doing your auditioning via a PM5D, which is introducing some latency in the signal path.
I distinctly recall an occasion when I was using an XL3 and setting up an XTA SIDD as a channel insert over the main vocal channel using a microphone and headphones. The difference between the insert out mode, i.e. full analog path with no latency, and the insert in mode with latency, was phenomenal. I would imagine that if you were to ask a vocalist using iems which sounded better, they would always choose the zero latency option.
I know that George Michael finds it so off putting that they use an XL4 just for his IEM mix, and the band and BV's are on a digital console.
I wonder if that's why some people ask for their vocal to be so loud. They're hearing the comb filtering introduced by the combination of bone conductance at zero latency , and iem signal with latency, and they don't like it. If the IEM signal were to be significantly louder than the bone conductance signal would the comb filtering be markedly reduced? Can someone chime in with the maths?
You can probably tell that I'm thinking out loud here.
I'm not sure how much budget you have to throw at providing a better solution for your client, ( possibly not a whole XL4's worth of budget), but a cost effective improvement might be to have an external analogue mic-pre and an analogue summing unit to combine the mic-pre output and a stereo mix from the PM5D. The output of the summing unit would go straight to the IEM transmitter.
That way you'd have all the benefits of recall for the mix, but the all-important analogue signal path for the vocal.

The ideal items would be a Midas XL42 for the preamp, and a Midas XL88 for the summing.

Enough late night musings from me.

Cheers,
andy.
 
Re: IEMs, Singers and Occlusion

If the IEM signal were to be significantly louder than the bone conductance signal would the comb filtering be markedly reduced? Can someone chime in with the maths?
You can probably tell that I'm thinking out loud here.

I believe that that is primarily why people ask for their IEMs to be so loud. Comb filtering only happens when the signals are nearly identical, and levels are nearly identical. A 6dB difference will greatly reduce the effect of the comb filter.

Mac
 
I believe that that is primarily why people ask for their IEMs to be so loud. Comb filtering only happens when the signals are nearly identical, and levels are nearly identical. A 6dB difference will greatly reduce the effect of the comb filter.

Mac

Yep, that is my take also.

Unfortunately, a level necessary to overcome occlusion also usually drags up all the other sources in level as well.

Sent from my DROID RAZR HD
 
Re: IEMs, Singers and Occlusion

Is it possible to EQ the channel to enhance or add what's missing to the occluded sound rather than trying to rise above it? I understand that it's somewhat of a rhetorical question, as without a dedicated vocal channel for a given mix it would be detrimental to every other mix, but with large channel count digital boards it wouldn't be out of the question.
 
Re: IEMs, Singers and Occlusion

This is a fascinating discussion, thank you Andy and others for starting and continuing it. The author has much credibility, he is a very senior and well respected audiology researcher working for many years with a multi-institutional team on multiple academic audiology research projects funded by the National Institutes of Health. The journal is, however, a lay publication not a peer reviewed scientific journal and thus leaves out some important scientific details and in my view oversimplifies some concepts.

For example, he speaks of "vibrations escaping through the open ear canal" rather than the reduction of pressure amplitudes in an enclosed space when the space is vented. That issue aside, I think the concepts are pretty well presented and make physical sense. By way of my background, I am a musician, I do sound for our group, and am also a physician and was a physics major undergrad. Although I am not either an ENT specialist or a monitor engineer, I do have a personal interest in both IEMs and hearing loss. I do not have any financial stake in the IEM or hearing aid business and my comments reflect only my own thoughts and end user interest in this topic.

The obvious difference between the current forum discussion and Dr. Ross' article is that he is discussing hearing aids not IEMs. This is important because the difficulty with venting the ear canal for a hearing aid is that it makes control of feedback from the transducer inside the ear canal to the microphone embedded in the hearing aid outside the ear canal much more difficult to control. This is not an issue with most IEMs that have no embedded microphone, so the venting solution becomes much more viable.

I have not seen vented IEMs, however I have a pair of passive musician's earplugs with a -10dB filter that is vented. I have had major problems with those as the vent seems to open and close almost randomly for unknown reasons, greatly affecting the attenuation from minute to minute. I will work with the manufacturer to try to resolve that problem but for now, I have stopped using the vented -10dB filters and use only the unvented -17dB filters.

I agree with Dr. Ross that making the earmold go deeper, so it ends past the second bend in the external ear canal, is one potentially effective approach. This can be demonstrated, as one forum member posted, by the reduction in occlusion effect when you press the IEM into the ear canal so it goes deep enough. This little experiment, is potentially very risky and could possibly lead to damage or rupture of the eardrum if the end of the IEM contacts the eardrum, so I strongly recommend against people trying it. Trust me, it works, as some of us have found out by carefully, recklessly or thoughtlessly trying it.

I would state the mechanism of the occlusion effect a bit differently than Dr. Ross stated it. The walls of the ear canal are cartilage covered with skin in the parts of the canal closest to the outside, and are bone covered with skin deeper in the canal, beyond where the canal makes its second bend. When you speak or sing, the vibration from the larynx is transmitted throughout the skull and the cartilage vibrates with much more amplitude than the bone. So if you occlude the ear canal with your finger, or use an IEM that ends before reaching the bony part, the larger vibrations of the cartilage are in contact with the air in the enclosed ear canal, are effectively transmitted to the eardrum, and you can hear the sounds very loudly. Dr. Ross reports around 20 dB louder sound than that you hear via external transmission through the air to your open ears. He did not mention the very large variation of this amplification over different frequencies, so that the sound heard is mainly lows and mids and not suitable for pitch control or optimal control of singing tone.

If you occlude the ear canal deeper than the second bend, and the walls in contact with the air are the bony part that has a much smaller amplitude of vibration, there is much less sound transmitted to the air inside the sealed ear canal and you don't hear the sounds nearly as loudly. Incidentally, in this setting, the loudness is not governed by the inverse square relationship, which applies to freefield not nearfield or enclosed spaces, but rather how much air is moved by the walls of the ear canal and whether the canal is completely sealed or vented.

The difficulty with the deep earmold approach is that it becomes much more difficult and risky to make earmolds as you go deeper. A physical block is placed in the ear canal to prevent the injected earmold material from contacting the eardrum. Placing this physical block safely without contacting or rupturing the eardrum becomes more difficult as you get closer to the eardrum, so most audiologists do it very cautiously. This often leads to earmolds that do not go quite deep enough, thus the common problem of the occlusion effect. Plus you don't want an earmold that, if touched or pressed on the outside, will contact the eardrum and possibly rupture it. I am hoping the venting approach works for IEMs so that would alleviate the problem in a way that would provide a greater margin of safety than the deep earmold approach.
 
Last edited:
Re: IEMs, Singers and Occlusion

Excellent discussion guys.

On a somewhat related note, I have been more regularly doing monitor mixes which involves realtime pitch correction being applied to vocals (let's not debate the use of pitch correction). If you haven't already run into this, you likely soon will as it becomes more widespread through the TC Helicon boxes and similar products.

Interestingly, occlusion is a major factor here as well, as when the singer's pitch is off, the corrected signal combines with the occlusion effect that the singer hears and creates an effect which sounds like their vocal is being run through a Chorus pedal. And, similarly, only the singer hears this phenomenon. This can obviously be applied only to the FOH mix to avoid this circumstance if you want to go that route. Also, I have encountered some singers who actually WANT to hear this effect - they say it helps them by letting them know when they are out of tune, effectively training them to sing better.

Sorry for a bit of a tangent from the original subject matter.