Rough pop/rock vocals

[tengiz]

Your tinnitus might be contributing as well here then. To me all these tracks or close to harshness but not becoming harsh on a balanced system. You would probably benefit from a low Q reduction of the 1-4khz area, if you have access to some manual EQ, try putting a 0.7Q -2dB EQ at 2.5 or 3khz and see how you perceive the difference.

Yes, I can do that with Audyssey MultEQ-X - it can add a manual PEQ filter to its automatic correction, I will give it a try. Thank you!

 

[574stereo]

ABBA vocals... they can sound pretty harsh. A lot of music from this era does.

Don't forget on one hand there are those of us listening on good systems, but the mixes also had to 'cut' on small radios, background sound systems, etc back in the day. Car audio systems... this is part of what we in the industry call 'translation' and it's very important. Mixing something to sound spectacular on a big sound system is not actually that difficult, making something which still has impact across all systems is much harder.

The tools we had to mix back then were exceptionally simplistic in comparison to what we have nowadays. The standard technique back then was simply to boost 3k (the presence range). It still works nowadays, but we can use dynamic EQs etc to make sure these ranges don't get too harsh. Plus there are a million other ways to help with translation (large, up-front vocals being one of them).


FWIW, The Visitors is one of my favourite sounding albums, although yes, in places the vocals can be a bit harsh. They didn't have the tools back then to get surgical, you had to rely on the performer and the vocal chain (microphone, preamp, EQ, compressor). The drums on The Visitors, and sound staging, it's incredible.

 

[Mart68]

ABBA vocals... they can sound pretty harsh. A lot of music from this era does.

Don't forget on one hand there are those of us listening on good systems, but the mixes also had to 'cut' on small radios, background sound systems, etc back in the day. Car audio systems... this is part of what we in the industry call 'translation' and it's very important. Mixing something to sound spectacular on a big sound system is not actually that difficult, making something which still has impact across all systems is much harder.

The tools we had to mix back then were exceptionally simplistic in comparison to what we have nowadays. The standard technique back then was simply to boost 3k (the presence range). It still works nowadays, but we can use dynamic EQs etc to make sure these ranges don't get too harsh. Plus there are a million other ways to help with translation (large, up-front vocals being one of them).


FWIW, The Visitors is one of my favourite sounding albums, although yes, in places the vocals can be a bit harsh. They didn't have the tools back then to get surgical, you had to rely on the performer and the vocal chain (microphone, preamp, EQ, compressor). The drums on The Visitors, and sound staging, it's incredible.

Can't say I agree with that. Struggling to think of any music I'm familiar with from that era that has issues with recording quality, at least not recordings from major labels. There are some where a 'lo-fi' production was intentional - I'm not counting those.

Yesterday I was listening to 'Off The Wall' which was made the same year as 'The Visitors'. Despite being a highly commercial recording for a mass market it sounds great, with zero harshness. On about £30K worth of equipment.

That's not to say no recording has any issues, but reported 'Harshness' problems are almost always problems with the playback system and/or the acoustics of the listening space.

The idea that some recordings can sound better on poorer quality playback equipment has gained some traction over the years, despite not making any logical sense.

 

[tengiz]

I don't know how much science it behind this, but I do recall a lot of music sounding fantastic on tiny transition radios when I was a kid, like Oxygene by JMJ.

FWIW, The Visitors is one of my favourite sounding albums, although yes, in places the vocals can be a bit harsh. They didn't have the tools back then to get surgical, you had to rely on the performer and the vocal chain (microphone, preamp, EQ, compressor). The drums on The Visitors, and sound staging, it's incredible.

The "harshness" of the vocal on "Should I Laugh or Cry" might also be intentional: I mean, I can’t imagine a harsher message to the guy in the song that she’d lost all respect for him.

 

[574stereo]

Bruce had a Harrison console when he did Off The Wall… it’s about as high budget as it gets. £30k wouldn't have even touched the sides with the budgets they were blowing on recording back then.

First three/four Steely Dan albums - harsh, pokey vocals.
Pretty much anything MPB - harsh, pokey vocals
Some of Joni's vocal recordings - harsh and pokey as hell in places.
Pet Sounds... pokey
10cc... again, harsh vocals


I'm not complaining, it's part of the sound of the era, and if anything continues all the way through into the late 90s, although the frequencies of choice moved lower and lower (by the mid/late 90s, pop music is all about 1.2k on the vocal). In fact some of my favourite sounding albums suffer from these issues, you just have to accept it as part of 'the sound'.


Some of the tools we have available to us in mixing had primitive counterparts back in the 70s and 80s, but they were astronomically expensive and often not as good. Even simple tasks like de-essing were hard to do effectively until the 1980s...

Can't say I agree with that. Struggling to think of any music I'm familiar with from that era that has issues with recording quality, at least not recordings from major labels. There are some where a 'lo-fi' production was intentional - I'm not counting those.

Yesterday I was listening to 'Off The Wall' which was made the same year as 'The Visitors'. Despite being a highly commercial recording for a mass market it sounds great, with zero harshness. On about £30K worth of equipment.

That's not to say no recording has any issues, but reported 'Harshness' problems are almost always problems with the playback system and/or the acoustics of the listening space.

The idea that some recordings can sound better on poorer quality playback equipment has gained some traction over the years, despite not making any logical sense.

 

[Arindal]

Specifically, some vocals in pop and rock recordings don’t sound very pleasing. Classical and jazz vocals, on the other hand, sound spectacular.

If that result is consistent over a vast number of tracks, there must be a reason for that.

I mean, what is the difference between pop/rock vocal recordings and classical/jazz? One might think on the former some things like EQ or effects are more likely to applied, but that is not a consistent thing. One really consistent difference is the fact that reverb on the recording differs. Pop and rock recordings tend to have artificial, decorrelated reverb, while acoustic recordings of classical and jazz catch the whole reverb in the concert hall including the footprint of the room and dispersion pattern of the voice in that room. The latter phenomenon helps our brain during stereo playback to ´contruct a virtual image´ of the whole scenario.

So I would try to find the answer in the acoustic situation in the playback room and how it interacts with the speakers. My experience with electrostatic planar speakers in a treated/overdamped room is limited, but I have encountered a similar scenario many years ago with some Quad ESL. Like your MLs, these are dipoles with a very high directivity index particularly in the upper mids and presence bands which is unusual for home speakers (Particularly the KEFs you mentioned are pretty broad-dispersion loudspeakers in that band).

So if the room suppresses early reflections from the sides as well as behind the listener, and the speakers tend to deliver a narrow beam, recordings without a natural reverb footprint were tending to sound ´like a laser beam flashlight´, particularly vocals. Overly thin tonality, obtrusively direct, bright, pronounced, artificial. The reverb from the recording was there, but detached from the direct sound impression of the voice. It might be an explanation that our brain cannot bring direct sound and resulting indirect reflection pattern together.

The same happens when listening to dry rough mixes in a studio on constant directivity monitors which are unforgiving regarding tonality (like the popular Kii Audio). Dry voices tend to sound overly bright, extremely narrow, very direct and ´like mono birds on a wire´, as a former professor of mine would put it. Maybe your ears are very sensitive to this scenario and you do not just perceive it as overly bright and direct, like me, but as annoying.

The moment the recording contains a meaningful reverb pattern, like classical music, this effect vanishes, and everything falls back into place. That was the case with the Quads back then. In a more lively room with decreasing RT60 towards higher frequencies hence ´warm reverb´, this problem was basically non-existent.

Note: This is just a theory. To verify it, an alternative scenario with higher and decreasing RT60 in the room might be a good idea.

 

[tengiz]

If that result is consistent over a vast number of tracks, there must be a reason for that.

I mean, what is the difference between pop/rock vocal recordings and classical/jazz? One might think on the former some things like EQ or effects are more likely to applied, but that is not a consistent thing. One really consistent difference is the fact that reverb on the recording differs. Pop and rock recordings tend to have artificial, decorrelated reverb, while acoustic recordings of classical and jazz catch the whole reverb in the concert hall including the footprint of the room and dispersion pattern of the voice in that room. The latter phenomenon helps our brain during stereo playback to ´contruct a virtual image´ of the whole scenario.

So I would try to find the answer in the acoustic situation in the playback room and how it interacts with the speakers. My experience with electrostatic planar speakers in a treated/overdamped room is limited, but I have encountered a similar scenario many years ago with some Quad ESL. Like your MLs, these are dipoles with a very high directivity index particularly in the upper mids and presence bands which is unusual for home speakers (Particularly the KEFs you mentioned are pretty broad-dispersion loudspeakers in that band).

So if the room suppresses early reflections from the sides as well as behind the listener, and the speakers tend to deliver a narrow beam, recordings without a natural reverb footprint were tending to sound ´like a laser beam flashlight´, particularly vocals. Overly thin tonality, obtrusively direct, bright, pronounced, artificial. The reverb from the recording was there, but detached from the direct sound impression of the voice. It might be an explanation that our brain cannot bring direct sound and resulting indirect reflection pattern together.

The same happens when listening to dry rough mixes in a studio on constant directivity monitors which are unforgiving regarding tonality (like the popular Kii Audio). Dry voices tend to sound overly bright, extremely narrow, very direct and ´like mono birds on a wire´, as a former professor of mine would put it. Maybe your ears are very sensitive to this scenario and you do not just perceive it as overly bright and direct, like me, but as annoying.

The moment the recording contains a meaningful reverb pattern, like classical music, this effect vanishes, and everything falls back into place. That was the case with the Quads back then. In a more lively room with decreasing RT60 towards higher frequencies hence ´warm reverb´, this problem was basically non-existent.

Note: This is just a theory. To verify it, an alternative scenario with higher and decreasing RT60 in the room might be a good idea.

The result is somewhat consistent, but I wouldn't say it broadly applies to most pop/rock. I usually don’t pay close attention to sound quality as long as it’s within reason. But when I notice something irritating or distracting, I focus more. I don’t have the vocabulary or training to use proper audio terminology, as you noticed—but I am very well-versed in the physics behind it and in signal processing.

There’s much less reverb in the room after treatment. I’m not sure what the “right” amount is; I just followed a rule of thumb I found that recommends covering no more than a certain percentage of surfaces. I placed the treatment “strategically” to handle the most important reflections.

It improved my listening experience a lot. The most noticeable change is that, as they say, the speakers completely “disappeared.” Unlike in the untreated room, the illusion of recreating the acoustics of the recorded space is much more convincing. I’d say my low-reverb room is no longer getting in the way.

I think I get what you mean about artificial reverb—especially the effect you describe as “mono birds on a wire.” But isn’t there a flip side? To me, it’s often pleasant, as it helps resolve spatial detail—for example, separating voices and parts in polyphonic vocal recordings. I really enjoy “dissecting” multi-part harmonies. It also helps with intelligibility.

I probably made one mistake—rather than hanging the acoustic panels, I glued them to the walls, so I can’t move them to experiment.

But I will experiment with the KEF R5 Meta, which seems to be Spinorama’s top floor-standing speaker when scored with EQ and a subwoofer. Luckily for me, they’re reasonably priced and arriving in a day or two.

 

[obsessive9119]

.But I will experiment with the KEF R3 Meta, which seems to be Spinorama’s top floor-standing speaker when scored with EQ and a subwoofer. Luckily for me, they’re reasonably priced and arriving in a day or two.

But the R3s are bookshelves, aren’t they?

 

[tengiz]

But the R3s are bookshelves, aren’t they?

Sorry, my mistake, I meant R5 - fixed on the original message

 

[richard12511]

This would be easier for us to read if you screenshot it from the "All SPL" tab, and change the Y-axis so that it goes from 45-105dB. Hopefully this also increases the resolution so that it shows every 5dB instead of 10dB on the Y-axis.

Beyond that, this overall looks (way) too bright, as in not enough bass. This can give the perception of unpleasant highs. Too much highs or too little lows is just two ways of saying the same thing.

Auddessy makes every speaker way too bright imo and less you use the mobile app and use a custom target curve and limit it to pretty low in frequency.

This very well could be an Auddessy problem, and not the speakers. OP do you have the Auddessy Mobile app? Try limiting it’s correction to just the subs and maybe manually adjusting it a bit.

 

[tengiz]

Auddessy makes every speaker way too bright imo and less you use the mobile app and use a custom target curve and limit it to pretty low in frequency.

This very well could be an Auddessy problem, and not the speakers. OP do you have the Auddessy Mobile app? Try limiting it’s correction to just the subs and maybe manually adjusting it a bit.

I use MultEQ-X, the Windows App that requires purchasing the Audyssey license for each receiver you want to use with it. Despite offering much more control than the mobile App, it still lacks very basic features in my opinion.

 

[Mart68]

Bruce had a Harrison console when he did Off The Wall… it’s about as high budget as it gets. £30k wouldn't have even touched the sides with the budgets they were blowing on recording back then.

First three/four Steely Dan albums - harsh, pokey vocals.
Pretty much anything MPB - harsh, pokey vocals
Some of Joni's vocal recordings - harsh and pokey as hell in places.
Pet Sounds... pokey
10cc... again, harsh vocals


I'm not complaining, it's part of the sound of the era, and if anything continues all the way through into the late 90s, although the frequencies of choice moved lower and lower (by the mid/late 90s, pop music is all about 1.2k on the vocal). In fact some of my favourite sounding albums suffer from these issues, you just have to accept it as part of 'the sound'.


Some of the tools we have available to us in mixing had primitive counterparts back in the 70s and 80s, but they were astronomically expensive and often not as good. Even simple tasks like de-essing were hard to do effectively until the 1980s...

I meant £30K of playback equipment lol - no doubt the equipment at Westlake was in the millions even back then.

Steely Dan - never noticed harsh vocals and been listening to those records for almost 40 years. Even on vinyl, no problem.

Joni Mitchell - was her own producer I think - 'Blue' is a very direct recording that gives issues on some set-ups - I'll concede that one.

MPB? Who is that? Been wracking my brain!

10CC never noticed any problems but been a long time since I listened to them. Pet Sounds I've no idea. Never rated it, don't own it.

Recording tech may have improved massively but I can listen to records from 1967 such as 'The Doors' and 'Music From Big Pink' and I think they sound fantastic. Then there's Sinatra recordings from the fifties which are pretty much perfect - and there's jazz recordings from that era I can't fault.

If you're a recording engineer this may be a case of a different perspective? I've not been in a studio in 25 years.

 

[sigbergaudio]

Auddessy makes every speaker way too bright imo and less you use the mobile app and use a custom target curve and limit it to pretty low in frequency.

This very well could be an Auddessy problem, and not the speakers. OP do you have the Auddessy Mobile app? Try limiting it’s correction to just the subs and maybe manually adjusting it a bit.

Not sure it has been my experience that Audyssey makes every speaker too bright, but I do agree that ideally you should limit it's correction to at least below 500hz. @tengiz if you purchase the audyssey mobile app as @richard12511 suggests, you could test that. Not sure it would help though, I suspect it may be a combination of your speakers and possibly the tinnitus. But through the internet it's all speculation on our part.

 

[sigbergaudio]

I incidentally wrote a guide for getting useful sound out of Audyssey five years ago or so. If Audyssey hasn't changed too much it might still be applicable:

Audyssey room calibration: Common mistakes

In this guide we go through the basics of running Audyssey room calibration and making sure it is configured for the best possible sound. This is a topic with lots of nuance and edge cases, but if you don't want to spend hours and hours on figuring it all out, this guide should get you off to a...

www.sigbergaudio.com

 

[Arindal]

I’m not sure what the “right” amount is; I just followed a rule of thumb I found that recommends covering no more than a certain percentage of surfaces. I placed the treatment “strategically” to handle the most important reflections.

Cannot tell you from afar. But if the voice reproduction appears to be overly direct and annoying, it might have been too much already given the fact that these speakers are high-direcitivty-index models requiring a certain minimum of RT60 in the room.

If you have problems with discrete reflections but none with dominating indirect field in the room, as a rule of thumb diffusion at the strategic positions is always better than absorption, as the latter is further reducing RT60.

The most noticeable change is that, as they say, the speakers completely “disappeared.”

That is a good indicator for localization precision and stability, but would not tell you anything about the level of indirect sound in the room and perceived distance to the projection plane. Recordings with proper natural reverb even work in an anechoic chamber in stereo.

To me, it’s often pleasant, as it helps resolve spatial detail—for example, separating voices and parts in polyphonic vocal recordings. I really enjoy “dissecting” multi-part harmonies. It also helps with intelligibility.

Agreed, but ideally it should be possible to achieve all goals at the same time: optimum of localization stability, detail resolution/transparency with complex recordings, and a reasonable distant projection plane hence depth-of-field. I prefer a slightly more distant and reverberant imaging and I am really into complex sacred music, polyphonic choir works and opera stuff. Could not get more complex.

Auddessy makes every speaker way too bright imo and less you use the mobile app and use a custom target curve and limit it to pretty low in frequency.

The problem with these automatic routines is they do not optimize for direct sound frequency response and have not possibility to distinguish direct and reflected sound. If both response curves deviate due to directivity index or uneven RT60, any attempts to get a flat overall curve will lead to pretty strange results, like the brightness you mentioned.

Joni Mitchell - was her own producer I think - 'Blue' is a very direct recording that gives issues on some set-ups - I'll concede that one.

Would not call it overly direct, but the lead vocals are an example of monaural ´bird on the wire´, reverb is somehow damped hence quickly decaying, and the seemingly short initial delay makes it sound like recorded in a very little chamber with the microphones pretty close to the instruments. But the vocals do not sound direct or pokey at all.

 

[tengiz]

I incidentally wrote a guide for getting useful sound out of Audyssey five years ago or so. If Audyssey hasn't changed too much it might still be applicable:

Audyssey room calibration: Common mistakes

In this guide we go through the basics of running Audyssey room calibration and making sure it is configured for the best possible sound. This is a topic with lots of nuance and edge cases, but if you don't want to spend hours and hours on figuring it all out, this guide should get you off to a...

www.sigbergaudio.com

Apparently, I already follow all your advice, with the exception of limiting the filters to 500 Hz. I tried it just now. And... to my astonishment, it did not change anything for the recordings that sounded fine before. They sound just as good.

But it did improve the perception of Frida's voice part in Should I Laugh…! Again, I don't have the vocabulary to describe the change "objectively," but it is better now . The best I can express it: it sounds a bit like coming from behind a veil, or something like that; and it did reduce the perception of harshness for me.

Thank you!

 

[tengiz]

Cannot tell you from afar. But if the voice reproduction appears to be overly direct and annoying, it might have been too much already given the fact that these speakers are high-direcitivty-index models requiring a certain minimum of RT60 in the room.

If you have problems with discrete reflections but none with dominating indirect field in the room, as a rule of thumb diffusion at the strategic positions is always better than absorption, as the latter is further reducing RT60.

The problem with these automatic routines is they do not optimize for direct sound frequency response and have not possibility to distinguish direct and reflected sound. If both response curves deviate due to directivity index or uneven RT60, any attempts to get a flat overall curve will lead to pretty strange results, like the brightness you mentioned.

Thank you, I find this very informative. I do have absorber/diffuser combo panels for the side reflections.

Regarding RT60 - I really have no clue how to assess it for my room. I mean, I looked at REW results, but I am not sure how to interpret it besides just making sure that the decay time for anything above bass remains in the recommended range for home theater between 300 and 600ms or something like that.

 

[Arindal]

Regarding RT60 - I really have no clue how to assess it for my room.

The average is rather on the low RT60, meaning dry side, so the room would be suitable for broad-dispersion studio monitors. A planar dipole might already bear the risk that the direct sound is dominating, at least in some bands.

What I find unusual is the significant absorption in the midrange, RT60 being down already at 200Hz and reaching a minimum between 400 to 800Hz which is both indicative of extensive use of pretty thick foam absorbers. As these dipoles also reach a pretty significant directivity index in the midrange and the room becomes more reverberant above 3K, this all might be contributing to a dominant presence/brillance band in the indirect sound field.

 

[sigbergaudio]

You could try moving some absorbers from the side wall and on to the wall behind you if you don't have anything there already.

 

[tengiz]

The average is rather on the low RT60, meaning dry side, so the room would be suitable for broad-dispersion studio monitors. A planar dipole might already bear the risk that the direct sound is dominating, at least in some bands.

What I find unusual is the significant absorption in the midrange, RT60 being down already at 200Hz and reaching a minimum between 400 to 800Hz which is both indicative of extensive use of pretty thick foam absorbers. As these dipoles also reach a pretty significant directivity index in the midrange and the room becomes more reverberant above 3K, this all might be contributing to a dominant presence/brillance band in the indirect sound field.

Thank you. This is again very useful.

I did re-calibrate, as suggested, with the 0.5 kHz limit on what Audyssey is allowed to correct—and it definitely made things better.

Also, I recall Mr. Toole's suggestion not to correct for the seats. Despite there is a measurable difference depending on where the microphone is placed. If you position the measuring microphone in front of the seat back versus above it, the results differ.

That’s exactly how I’ve been measuring. Since the seats have fairly high backs, placing the microphone at ear level means it's right in front of a relatively large, partially reflective/absorptive surface. For my most recent re-calibration, I added an additional measurement position (later ignored in the calibration)—aligned with the MLP but with the microphone above the seat back.

These are raw, uncorrected measurements captured by Audyssey:

Note, the room response of the speaker is pretty good. It hardly needs any correction above 0.2k.

However, the position in front of the back of the seat measures much worse:

The difference roughly matches the geometry—the distance from the microphone to the back of the seat varies between 4–5", resulting in a 0.6–0.8m propagation delay between the direct and reflected sound reaching the microphone. This would create a comb filter with the first dip centered around 0.6k and a peak around 1.3k.

I guess this adds another reason why Audyssey should not be allowed to correct above 0.5 kHz.