Rough pop/rock vocals

[tengiz]

I like how my setup sounds, and I’ve enjoyed it for years. However, there are occasional — and very annoying — wrinkles. Specifically, some vocals in pop and rock recordings don’t sound very pleasing. Classical and jazz vocals, on the other hand, sound spectacular.

I believe this may have to do with the fact that many pop and rock recordings use microphones processed through various distortion and compression tools or algorithms. But these flaws seem to become apparent only through my main speakers. I don’t hear them as clearly in my car, headphones, or other playback systems.

I’ve tried various things — swapping amplifiers, matching the frequency response to other setups, and more — but nothing has helped.

My listening room is acoustically treated, with absorbers and diffusers behind the speakers and on the side walls. I also have absorption and diffusion panels on the rear (listener-side) wall and on the ceiling, and the floor is carpeted. As a result, all first reflections are attenuated.

My main speakers are MartinLogan Ethos. I know ESLs tend to have a bad rap when it comes to off-axis response, and under the assumption that off-axis irregularities — which still interact with the room despite acoustic treatment — might be one of the culprits, I’ve considered that possibility. But I find it bizarre that it only seems to affect some pop and rock vocals.

If there’s something I’m missing, I’d love to learn more.

Thank you.

 

[DVDdoug]

I believe this may have to do with the fact that many pop and rock recordings use microphones processed through various distortion and compression tools or algorithms.

Possibly... Or possibly because it's "cool" to strain your voice...

But these flaws seem to become apparent only through my main speakers. I don’t hear them as clearly in my car, headphones, or other playback systems.

It could be something in your speakers. It's hard to tell without trying different (good) speakers. I don't know how your speakers measure or sound, but I believe electrostatics are "different" from traditional speakers and preference is a matter of taste. "You never know what the problem is until it's solved".

Usually headphones tend to bring-out the details and they usually go louder which also brings-out details. So if it's in the recording you'd probably hear it on headphones.

 

[Philbo King]

It might be mic selection?
Sometimes a different mic can make a big difference.

 

[tengiz]

It could be something in your speakers. It's hard to tell without trying different (good) speakers. I don't know how your speakers measure or sound, but I believe electrostatics are "different" from traditional speakers and preference is a matter of taste. "You never know what the problem is until it's solved".

Right — but which speakers should I try? I haven’t used dynamic loudspeakers in my main setup since 2008. While it’d be fun to explore, I’d rather focus on a solution without a long trial-and-error process.

I chose electrostatics for their spatial precision, especially in complex vocal recordings. If I switch to dynamics, I don’t want to lose that.

I visited a few local Magnolia listening rooms to check out the KEF R Meta series. Only one was acoustically treated — and it featured the KEF Reference 5, their top floorstander per Spinorama.

The demo was underwhelming. No EQ was used, which might explain it, but if well-behaved speakers with flat on-axis and good off-axis response sound good in most rooms, something still seemed off.

 

[tengiz]

It might be mic selection?
Sometimes a different mic can make a big difference.

It might be, but then again, should I hear the difference with other speakers as well?

 

[ThatSoundsGood]

What songs by which artists are you hearing this? Let's get specific.

 

[komhst]

I like how my setup sounds, and I’ve enjoyed it for years. However, there are occasional — and very annoying — wrinkles. Specifically, some vocals in pop and rock recordings don’t sound very pleasing. Classical and jazz vocals, on the other hand, sound spectacular.

Have you ever seen the waveform of those pop and rock recordings in a sound processing program i.e. Audacity? Probably they are "brickwalled" fully compressed with clipped peaks and a ton of distortion. So, don't care about these recordings and don't try to fix anything. You can't! It's not your system's fault. Those ****** recordings is the problem.

Just try to find those recordings here:

Album list - Dynamic Range DB

And if they are in the orange or red range don't bother anymore.

 

[DavidMcRoy]

I have bemoaned the tendency for pop/rock vocals to sound distorted since the 1960s. Clean ones are pretty rare.

 

[somebodyelse]

Possibly... Or possibly because it's "cool" to strain your voice...

If it's actually a strain, not just good technique, you won't be a vocalist for very long! This discussion between clean and metal vocal coaches should give some idea why.

This is part of the reason OP should be specific about which bit of which song - otherwise we won't know whether it's a well recorded example of a vocal technique they don't like, a distorted recording, sibilance that's getting emphasised by a peak in the playback equipment or something else.

 

[tengiz]

What songs by which artists are you hearing this? Let's get specific.

For instance, on many ABBA tracks (usually from before 1979), the vocals sound harsher through my electrostats than on any other reasonably good speakers I’ve tried.

 

[ThatSoundsGood]

For instance, on many ABBA tracks (usually from before 1979), the vocals sound harsher through my electrostats than on any other reasonably good speakers I’ve tried.

That does not sound like it's the recording, the music, the microphone, or the singer. It sounds like a problem with the speaker. I believe there are measurements of those electrostatics somewhere and they are dismal. Absolutely horrific.

 

[ThatSoundsGood]

Have you ever seen the waveform of those pop and rock recordings in a sound processing program i.e. Audacity? Probably they are "brickwalled" fully compressed with clipped peaks and a ton of distortion. So, don't care about these recordings and don't try to fix anything. You can't! It's not your system's fault. Those ****** recordings is the problem.

Just try to find those recordings here:

Album list - Dynamic Range DB

And if they are in the orange or red range don't bother anymore.

Have you ever mixed or mastered a record? There is so much misunderstanding about it out there. If you zoom into many of these recordings closer, there are actually not clipped peaks on a lot of them. They look really compressed when you're zoomed out, but less so when you zoom in. Yes, they are compressed and ultimately limited to get the RMS louder, but a lot of that is not really adding much or any distortion unless the mastering engineer chose that. When they did, choose it, it's because most listeners thought it sounded better. We respond well to 2nd harmonic distortion and sometimes clipping the input of a limiter adds that special touch. The vocals, however are often not hitting the limiters unless they are mixed really loudly. It's mostly kick, snare and bass. And if the vocals hit the limiter then it would release so fast you would be unlikely to register it as distortion.
There are exceptions to what I'm saying........I know of some albums that are so horribly compressed that you can really hear them distorting (The Mars Volta's first record) but I often find that the distortion in mastering is far overstated. Also, have you ever listened to how much distortion is on old tape masters? Led Zeppelin 2 might as well have been run through a Boss distortion pedal.
Distortion on vocals usually subtle and pleasing. The old Journey records sound amazing and you can clearly hear the tube microphone on that.
But in the OP's post, he's saying that he only hears it on his electrostatics, not on any other speakers. To me, it's either the electrostatics are more revealing and we figure out what records have the problems, or the electrostatics themselves measure horribly (as they often do) and the speakers are the problem. All it would take is a rise in magnitude or distortion in the 2Khz-5Khz region to make you hear this problem.
Check these stereophile measurements on a different electrostatic. Horrible and that 2k-5k region would stick out like a sore thumb.

Quad ESL-989 electrostatic loudspeaker Measurements

Sidebar 3: Measurements Although the different kind of interaction between a panel speaker and the room can make it sound louder than its measured sensitivity would suggest, the Quad ESL-989 is still on the low side, at an estimated 83dB(B)/2.83V/m. Given its strict 100W power handling, this...

www.stereophile.com

 

[tengiz]

But in the OP's post, he's saying that he only hears it on his electrostatics, not on any other speakers. To me, it's either the electrostatics are more revealing and we figure out what records have the problems, or the electrostatics themselves measure horribly (as they often do) and the speakers are the problem. All it would take is a rise in magnitude or distortion in the 2Khz-5Khz region to make you hear this problem.

To my ears, it does often sound like the harmonics are exaggerated (or the fundamental suppressed) when certain vocal parts are played through electrostats. The speakers themselves have very low distortion — in the 0.5–4 kHz range and at my usual listening levels, THD is below 0.3%. But they are not flat - I never bothered, assuming a ±3 dB variation was acceptable.

In other words, if the fundamental is down by 3dB and the 2nd and 3rd harmonics are up by 3dB due to a non-flat frequency response, resulting in 6db raise in harmonics, it might indeed sound as though the recording is more distorted than it actually is.

If that’s the case, would the fix simply be to flatten the vocal range (with the first few harmonics) at and around the main listening position? That would be easy enough with my setup.

 

[tengiz]

Have you ever seen the waveform of those pop and rock recordings in a sound processing program i.e. Audacity? Probably they are "brickwalled" fully compressed with clipped peaks and a ton of distortion. So, don't care about these recordings and don't try to fix anything. You can't! It's not your system's fault. Those ****** recordings is the problem.

Just try to find those recordings here:

Album list - Dynamic Range DB

And if they are in the orange or red range don't bother anymore.

This is a useful list—thank you!

But I think the DR in this list refers to something slightly different?

No, I’ve never looked at the waveforms of the recordings where the vocals sound harsh on my speakers. I never bothered because the effect seems limited to the vocals, even though there’s a lot more going on in the background. I’d assume that anything approaching clipping would affect the entire mix, not just the vocals.

 

[sigbergaudio]

If old Abba vocals sounds displeasing to you, I would suggest you have frequency irregularities due to either speakers, room or a combination of both. If it is the speakers, it may not be easily fixed by EQ either.

With regards to Kef, did you listen to the Reference 5 or The R5 (Yes, the naming is confusing)?

 

[ThatSoundsGood]

To my ears, it does often sound like the harmonics are exaggerated (or the fundamental suppressed) when certain vocal parts are played through electrostats. The speakers themselves have very low distortion — in the 0.5–4 kHz range and at my usual listening levels, THD is below 0.3%. But they are not flat - I never bothered, assuming a ±3 dB variation was acceptable.

In other words, if the fundamental is down by 3dB and the 2nd and 3rd harmonics are up by 3dB due to a non-flat frequency response, resulting in 6db raise in harmonics, it might indeed sound as though the recording is more distorted than it actually is.

If that’s the case, would the fix simply be to flatten the vocal range (with the first few harmonics) at and around the main listening position? That would be easy enough with my setup.

The only way to determine this is to take good measurements from the listening position. Can you do this?

 

[DavidMcRoy]

Regarding speculation that speakers may be at fault, if you're lucky enough to find an old review with speaker distortion measurements you'll note that electrostatic designs as a group are pretty low in distortion. No one knows how OP's speakers perform without measurements. Anyway, distorted pop and rock vocals are definitely a thing, and it can be most easily heard on good headphones or IEMs.

 

[tengiz]

The only way to determine this is to take good measurements from the listening position. Can you do this?

I’ve done it many times while tuning the bass response, but the rest was left to Audyssey and its standard approach to EQ—averaging across multiple listening positions. The end result stays within a ±3 dB band at the main listening position, so I didn’t go beyond that. However, I can see that the vocal fundamentals fall into a dip, while the second and third harmonics land roughly in peaks—potentially boosting those harmonics for female vocals by around 6 dB.

I have a background in physical acoustics (I used to do research that involved building multi-beam phased-array sonars), but my knowledge of psychoacoustics is limited. So I have no idea whether a 6 dB boost in the second and third harmonics is enough to cross an audibility threshold if the vocal is already distorted in the recording.

Still, it does sound plausible to me.

 

[tengiz]

Regarding speculation that speakers may be at fault, if you're lucky enough to find an old review with speaker distortion measurements you'll note that electrostatic designs as a group are pretty low in distortion. No one knows how OP's speakers perform without measurements. Anyway, distorted pop and rock vocals are definitely a thing, and it can be most easily heard on good headphones or IEMs.

I measured the distortion for my speakers in the room - it basically stays below 0.3% pretty much anywhere above 100Hz or so at 90dB.

 

[obsessive9119]

I’ve done it many times while tuning the bass response, but the rest was left to Audyssey and its standard approach to EQ—averaging across multiple listening positions. The end result stays within a ±3 dB band at the main listening position, so I didn’t go beyond that. However, I can see that the vocal fundamentals fall into a dip, while the second and third harmonics land roughly in peaks—potentially boosting those harmonics for female vocals by around 6 dB.

I have a background in physical acoustics (I used to do research that involved building multi-beam phased-array sonars), but my knowledge of psychoacoustics is limited. So I have no idea whether a 6 dB boost in the second and third harmonics is enough to cross an audibility threshold if the vocal is already distorted in the recording.

Still, it does sound plausible to me.

To my naive eye, that looks very flat for an in-room response - usually you'd expect a downward sloping curve. That'd be pretty bright, wouldn't it?