Hi everyone, this is my first post... Hope to find help here!
My questions are related to why almost every audio system over-emphasizes every "s" and "t" and similar "sounds" when listening to voices... I find this quite annoying! For example someone sings "...what's this all about..." and it sounds like "...whatZZZ thiZZZ all about". "p and "t" sometimes sound strange, too. I've heard ONE single system which didn't do this and so I cannot blame it on my records.
1. Is there a special kind of distortion / lack of resolution that is responsible for this annoying sound? Has someone done research into this and can explain how it happens?
2. Is there a correct technical term to better describe what I hear (cold / harsh / sharp / ?)
3. How must an amplifier be designed to get rid of this problem?
Thank you very much!
My questions are related to why almost every audio system over-emphasizes every "s" and "t" and similar "sounds" when listening to voices... I find this quite annoying! For example someone sings "...what's this all about..." and it sounds like "...whatZZZ thiZZZ all about". "p and "t" sometimes sound strange, too. I've heard ONE single system which didn't do this and so I cannot blame it on my records.
1. Is there a special kind of distortion / lack of resolution that is responsible for this annoying sound? Has someone done research into this and can explain how it happens?
2. Is there a correct technical term to better describe what I hear (cold / harsh / sharp / ?)
3. How must an amplifier be designed to get rid of this problem?
Thank you very much!
Hi EE, welcome.
I think the problem comes mostly from the recording, especially cheap microphones and singers that don't know how to handle them. Doesn't answer your questions, though.
I know exactly what you mean and find it very annoying, too. Don't know the english terms for it. Usually you hear these artifacts on lo-fi systems only. A high quality tweeter will improve the situation. As for amplifiers, I found tubes and the Pass Labs Cass-A designs are the best in that respect.
I think the problem comes mostly from the recording, especially cheap microphones and singers that don't know how to handle them. Doesn't answer your questions, though.
I know exactly what you mean and find it very annoying, too. Don't know the english terms for it. Usually you hear these artifacts on lo-fi systems only. A high quality tweeter will improve the situation. As for amplifiers, I found tubes and the Pass Labs Cass-A designs are the best in that respect.
I agree with AMT...
Recordings often over-emphasize sibilants. This is often due to the fact that they are heavily compressed. There is a process used in the studio called "de-essing" that is supposed to take care of this. I find a lot of recordings that still have this too.
Have you listened to an SACD? I find the sibilants and cymbal sounds to be much more realistic and musically accurate. CD often transforms these sounds into a "white noise" effect.
Other than that, I agree that the use of certain types of equipment will make this sound more musical. Whether you are getting an accurate presentation is something else entirely. Maybe musicality is better?
Recordings often over-emphasize sibilants. This is often due to the fact that they are heavily compressed. There is a process used in the studio called "de-essing" that is supposed to take care of this. I find a lot of recordings that still have this too.
Have you listened to an SACD? I find the sibilants and cymbal sounds to be much more realistic and musically accurate. CD often transforms these sounds into a "white noise" effect.
Other than that, I agree that the use of certain types of equipment will make this sound more musical. Whether you are getting an accurate presentation is something else entirely. Maybe musicality is better?
This high frequency emphasis on some recordings is an effect that has become more prominent as the resolution of my hi-fi has increased. Whilst it never sounds harsh or unpleasant it is conspicuous on some of the brighter recordings.
Changing amps and/or speakers to try and cure the problem is an expensive and futile pursuit since the quality of recordings varies so much, many still sound perfectly ok.
Perhaps the solution is some kind of (dare I say) switchable tone control to cover the top-end.
Changing amps and/or speakers to try and cure the problem is an expensive and futile pursuit since the quality of recordings varies so much, many still sound perfectly ok.
Perhaps the solution is some kind of (dare I say) switchable tone control to cover the top-end.
I've discovered that silk tweeters are much "smoother" than metal domes and switching to them can remove much of the sibilance during playback.
I also suspect that poor phase response in the high-pass filter of the speaker cross-over is partly to blame. My theory is that the changing delay over the HP bandwidth is causing an effect that is similar to noise. Since removing some of the harmonics can decrease this noise-like effect, I think that what is known as the "full" sound of tubes is actually the removal of sibilance with the loss of some harmonics.
ensen.
I also suspect that poor phase response in the high-pass filter of the speaker cross-over is partly to blame. My theory is that the changing delay over the HP bandwidth is causing an effect that is similar to noise. Since removing some of the harmonics can decrease this noise-like effect, I think that what is known as the "full" sound of tubes is actually the removal of sibilance with the loss of some harmonics.
ensen.
My impression is that there's a trend among singers, in particular female for some reason, to actually sing "schinging".
Possible reasons:
. I can hear it in ear phones too, so the cross-over network is not the single source, at least.
Rune
Possible reasons:
- It's a singing style trend, intented to signal "nearness"
- It's an artefact produced by the de-essers
- It's something singers do to defeat the de-essers, as they sound muffled otherwise
- It's CD-s mixed/mastered into digital clipping somewhere
. I can hear it in ear phones too, so the cross-over network is not the single source, at least.Rune
Sometimes it's bad recording. But this could also be deliberate.
Certain recordings, especially female vocals, have their sibilance boosted during mastering so that they sound better in low-fi systems in the mass market.
Unfortunately, in better systems, these excessiveness are highlighted to the point of being unbearable at times. I often wonder why recording companies fail to issue audiophile versions where the mastering is not targetted to the mass, but true music lovers.
Certain recordings, especially female vocals, have their sibilance boosted during mastering so that they sound better in low-fi systems in the mass market.
Unfortunately, in better systems, these excessiveness are highlighted to the point of being unbearable at times. I often wonder why recording companies fail to issue audiophile versions where the mastering is not targetted to the mass, but true music lovers.
Don't forget to post those shi*bilant recordings here for all to be wary of.
http://www.diyaudio.com/forums/showthread.php?goto=newpost&threadid=20845
ensen.
http://www.diyaudio.com/forums/showthread.php?goto=newpost&threadid=20845
ensen.
Is it live or is it re-mastered?
That doesn't add up for me... if the recording is to be played back on a lo-fi, then it's also likely the system is EQ'ed in a smiley face, and that's usually over and above the loudness being activated. This of course results in boosted highs and the playback being even more sibilant. Even the dumbest studio engineers must have learned that by the time they get to be top tech dog on a session. (Minor apologies to any studio engineers who have done this.) Now the producers, they are another story... I think many of them couldn't tell a Casio from a Yamaha grand and that includes many of the artists themselves.
Whatever the reason, I would pay extra money for audiophile masters. In particular I think Sony deserves a hand for all the re-masters they are issuing of the old Columbia recordings. I've not found one that I didn't like and it might be due to the extra love and care they put into stuff that will go to SACD. Never mind the higher resolution, that format is at least forcing some quality back into the process. Maybe if Memorex re-issued their old commercials in a new campaign, consumers will start asking for better again.
ensen.
mcp said:
That doesn't add up for me... if the recording is to be played back on a lo-fi, then it's also likely the system is EQ'ed in a smiley face, and that's usually over and above the loudness being activated. This of course results in boosted highs and the playback being even more sibilant. Even the dumbest studio engineers must have learned that by the time they get to be top tech dog on a session. (Minor apologies to any studio engineers who have done this.) Now the producers, they are another story... I think many of them couldn't tell a Casio from a Yamaha grand and that includes many of the artists themselves.
Whatever the reason, I would pay extra money for audiophile masters. In particular I think Sony deserves a hand for all the re-masters they are issuing of the old Columbia recordings. I've not found one that I didn't like and it might be due to the extra love and care they put into stuff that will go to SACD. Never mind the higher resolution, that format is at least forcing some quality back into the process. Maybe if Memorex re-issued their old commercials in a new campaign, consumers will start asking for better again.
ensen.Ah yes, the smiley face........
My guess is that they are trying to give it more "presence" by filling the hole in a 2-way low-end speaker as most of them use the woofer all the way up to the roll off before a tweeter is crossed with a cheapo cap. Usually, it is the frq between 4K - 8K that are boosted. Viewed with a spectrum analyzer, these boostings look like they were done with parametric EQs.
Re-reading this, I realized that I own some records that produce terrible sibilants on one system and sound smooth and warm on others. So probably we cannot blame it on the records alone.
Also, I don't think it is mainly related to manipulation of frequency response, whether done at home or in the studio. You cannot correct this phenomenon with EQs while playing back. I think it is some kind of "'s' to 'ZZZ' distortion" which occurs at both ends, while recording and again while playing back. To me it seems that "s" (which is actually noise of limited bandwidth) is very hard to reproduce accuratly for an audio system. It might be related to phase distortion or TIM.
It would be interesting to do some research into this, trying to get a very accurate recording of a good singer, and seeing (by wave form or spectrum comparison) what a bad system makes of it.
Also, I don't think it is mainly related to manipulation of frequency response, whether done at home or in the studio. You cannot correct this phenomenon with EQs while playing back. I think it is some kind of "'s' to 'ZZZ' distortion" which occurs at both ends, while recording and again while playing back. To me it seems that "s" (which is actually noise of limited bandwidth) is very hard to reproduce accuratly for an audio system. It might be related to phase distortion or TIM.
It would be interesting to do some research into this, trying to get a very accurate recording of a good singer, and seeing (by wave form or spectrum comparison) what a bad system makes of it.
1x1x1x1x1=1, 1.1x1.1x1.1x1.1x......... = infinity
In my experience, tweeters that are allowed to operate at their resonance frequency can cause this kind of hard 'essing' distortion.
Also high loop NFB amplifiers will react to reactive (non resistive) loads, and typically cause a ringing, essing condition.
In my experience, RC and RLC networks across the loudspeaker drivers dramatically reduces this effect.
Highly inductive speaker cable also contributes strongly to this effect when using a reactive high loop NFB amplifier.
When this amplifier/cable/speaker system is impedence compensated correctly, the 'essing' differences between recordings reduces dramatically.
The original recorded 'essing' sounds are not actually reduced, but the 'cascade' downstream effects are, and the resultant in-room sound is much more ear friendly, and more consistent from recording to recording.
By removing (reducing) this system 'cascade' reactivity, a large degree of masking is eliminated and then recordings can be heard correctly without final replay system embellishment (and consequent ear-bleed/tinnitus).
As Charles states, mic preamp behaviour (especially transient overload/recovery behaviour) is mission critical to capturing good sounding recordings, and this is even more important with modern overly close mic'd loud sound sources.
A clean non-reactive replay system will tell it like it is, and a reactive playback system will exponentially add HF dirt.
On first listen a non-reactive system may sound a litttle high mids/tops lacking but on extended listening will be revealled as correctly accurate.
After listening to such a system, a reactive system will sound rough and even irritating, and aparent high mid/highs detail will be revealled as false detail and embellishment.
Eric.
In my experience, tweeters that are allowed to operate at their resonance frequency can cause this kind of hard 'essing' distortion.
Also high loop NFB amplifiers will react to reactive (non resistive) loads, and typically cause a ringing, essing condition.
In my experience, RC and RLC networks across the loudspeaker drivers dramatically reduces this effect.
Highly inductive speaker cable also contributes strongly to this effect when using a reactive high loop NFB amplifier.
When this amplifier/cable/speaker system is impedence compensated correctly, the 'essing' differences between recordings reduces dramatically.
The original recorded 'essing' sounds are not actually reduced, but the 'cascade' downstream effects are, and the resultant in-room sound is much more ear friendly, and more consistent from recording to recording.
By removing (reducing) this system 'cascade' reactivity, a large degree of masking is eliminated and then recordings can be heard correctly without final replay system embellishment (and consequent ear-bleed/tinnitus).
As Charles states, mic preamp behaviour (especially transient overload/recovery behaviour) is mission critical to capturing good sounding recordings, and this is even more important with modern overly close mic'd loud sound sources.
A clean non-reactive replay system will tell it like it is, and a reactive playback system will exponentially add HF dirt.
On first listen a non-reactive system may sound a litttle high mids/tops lacking but on extended listening will be revealled as correctly accurate.
After listening to such a system, a reactive system will sound rough and even irritating, and aparent high mid/highs detail will be revealled as false detail and embellishment.
Eric.
Re: 1x1x1x1x1=1, 1.1x1.1x1.1x1.1x......... = infinity
Okay, then... is this what you are saying?
1) A tweeter that crossed over too low will play back signals that cover it's Fr which contributes to the effect.
2) Using an amp with high damping factor (ultra low impedance output) can create the effect.
3) High efficiency speakers contribute by being reactive (non-resistive)
4) Higher order crossovers reduce the problem by having many R(L)C networks.
5) Mis-matched input and output impedances along the signal path add noise which can accumulate into audible "ess" artefacts.
6) Clipping at any stage can start the "cascade" error and especially at the microphone.
7) Bright-sounding speakers are actually noisy and thus can create the problem.
ensen.
mrfeedback said:
Okay, then... is this what you are saying?
1) A tweeter that crossed over too low will play back signals that cover it's Fr which contributes to the effect.
2) Using an amp with high damping factor (ultra low impedance output) can create the effect.
3) High efficiency speakers contribute by being reactive (non-resistive)
4) Higher order crossovers reduce the problem by having many R(L)C networks.
5) Mis-matched input and output impedances along the signal path add noise which can accumulate into audible "ess" artefacts.
6) Clipping at any stage can start the "cascade" error and especially at the microphone.
7) Bright-sounding speakers are actually noisy and thus can create the problem.
ensen.
I'm aware of the phenomonon. However, to the best of my ability to remember, it was a lot worse 20-30 years ago. Some of it is, I'm sure. created in the recording studio. Maybe most as I find it absent in SACD DVD-A and most other recordings made with the audiophile market in mind.
Two suspicions: A- nearly all audio divices (SS and tube/ digital and analog) have minimum distortion around 1k-2kHz and rises, sometimes rapidly, above that. Sibilants, cymbals etc are in this range of rising distortion. While nearly every audio component has a distortion rating at 1k-2k that is well below audibility, this is not necessarily the case as one moves up the spectrum.
B- The is a paper published on www.rane.com that explains how clipping is more common at high than low frequencies as a result of compression caused by a combination of recording levels and playback levels.
Perhaps one or both of these suspicions of mine are involved.
****
Finally there is the irony of "revealing". Audio gear (particularly the expensive kind) is often praised as "revealing". A reviewer says that there are previously unheard details that become audible with the new piece of equipment. The cudest way to achieve this is just to turn up the treble - you will indeed hear new details - details the artist and recording engineer didn't intend you to hear! Actually I think you can probably get the same effect by cranking up any part of the sound spectrum relative to the rest.
I mention this, since it ocurrs to me that ever since installing a Behringer EQ and taking a couple of weeks to get it set up right, I don't recall any cases of annoying HF artifacts. I was motivated out a a desire to clean up the bass but I may have reaped an additional benefit.
Actually, think about it more. I think this latter part of my post may be more germaine than the speculations at the beginning.
Two suspicions: A- nearly all audio divices (SS and tube/ digital and analog) have minimum distortion around 1k-2kHz and rises, sometimes rapidly, above that. Sibilants, cymbals etc are in this range of rising distortion. While nearly every audio component has a distortion rating at 1k-2k that is well below audibility, this is not necessarily the case as one moves up the spectrum.
B- The is a paper published on www.rane.com that explains how clipping is more common at high than low frequencies as a result of compression caused by a combination of recording levels and playback levels.
Perhaps one or both of these suspicions of mine are involved.
****
Finally there is the irony of "revealing". Audio gear (particularly the expensive kind) is often praised as "revealing". A reviewer says that there are previously unheard details that become audible with the new piece of equipment. The cudest way to achieve this is just to turn up the treble - you will indeed hear new details - details the artist and recording engineer didn't intend you to hear! Actually I think you can probably get the same effect by cranking up any part of the sound spectrum relative to the rest.
I mention this, since it ocurrs to me that ever since installing a Behringer EQ and taking a couple of weeks to get it set up right, I don't recall any cases of annoying HF artifacts. I was motivated out a a desire to clean up the bass but I may have reaped an additional benefit.
Actually, think about it more. I think this latter part of my post may be more germaine than the speculations at the beginning.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.