Rough pop/rock vocals

[Chris A]

Where did you get this track to find out it was clipping?

You quoted the source...right? Is this a trick question?

Can you tell me your process here?

Here is a fairly ancient resource that explains how to use Audacity freeware to do similar operations without using for-profit plugins (i.e., freeware only):

Demastering Your Music Tracks (Part 2) - diyAudio

Nowadays, I would use a Fabfilter Pro-MB or perhaps Pro-Q4 plugin to target and compress the two identified narrow bands where indicated at ~3.1 and ~4.6 kHz (...sort of like using a de-esser...but using a spectrogram to target the offending overheated synthesized vocal harmonics where they occur--then listening again).

But that's an unnecessarily complicated procedure for this thread (i.e., without someone else trying to put on a prior-life mastering hat to imply that "you can't do that")...

We wouldn't want to muddy the water any more than necessary, would we...?

At this point - I am done indeed...

This is the real goal.

Chris

 

[ThatSoundsGood]

You quoted the source...right? Is this a trick question?

It's not a trick question at all. I know that different platforms can use different conversions. I also know that it can be fairly easy to make a mistake with it at any point in the signal chain. 2 or 3 db of clipping is kinda wild and I'm just wondering how that happens with a source that was done in the 70's and would likely not be anywhere near clipping when it was original. They just didn't master anything that hot in those days because it was going onto vinyl. So, for me, the source is important. I wasn't questioning your integrity here. I was just wondering how you got it into MP3 or WAV form to import into Audacity because that makes a big difference. I would be willing to take a look at that WAV file in Pro Tools to see if it clips in the same way. I'm a bit of a cynic about this stuff because people are constantly talking about Masters being too loud and clipping when they are not actually clipping. But I'm not going to accuse you of being wrong unless I look at it myself, because I have seen some terrible stuff and it very well might be the disaster you're showing here.

Here is a fairly ancient resource that explains how to use Audacity freeware to do similar operations without using for-profit plugins (i.e., freeware only):

Demastering Your Music Tracks (Part 2) - diyAudio

This is an interesting guide.

Nowadays, I would use a Fabfilter Pro-MB or perhaps Pro-Q4 plugin to target and compress the two identified narrow bands where indicated at ~3.1 and ~4.6 kHz (...sort of like using a de-esser...but using a spectrogram to target the offending overheated synthesized vocal harmonics where they occur--then listening again).

But that's an unnecessarily complicated procedure for this thread (i.e., without someone else trying to put on a prior-life mastering hat to imply that "you can't do that")...


We wouldn't want to muddy the water any more than necessary, would we...?


This is the real goal.

Chris

 

[Chris A]

I wasn't questioning your integrity here.

Right.

(To other readers here: I think my suspicions were actually correct in this instance.)

2 or 3 db of clipping is kinda wild and I'm just wondering how that happens with a source that was done in the 70's and would likely not be anywhere near clipping when it was original. They just didn't master anything that hot in those days because it was going onto vinyl.

Because the version I found clearly wasn't remastered in the 1970s. Did you miss that part above in post #85?

I'm actually trying to make it as easy as possible for the readers here. Is that what you're doing?

Chris

 

[ThatSoundsGood]

Right.

(To other readers here: I think my suspicions were actually correct in this instance.)

Fine. If you want to argue this then let's have a go.

Because the version I found clearly wasn't remastered in the 1970s. Did you miss that part above in post #85?

I saw that and this is why I was asking where you got it from. Was it remastered or was it the conversion to MP3/WAV that caused the clipping problem? Was it clipping when you imported it into Audacity or after you "Normalized" it? If you tell me which version you have then I can check the work for accuracy, but you still haven't answered where you got the MP3/WAV from.

I'm actually trying to make it as easy as possible for the readers here. Is that what you're doing?

I read the De-Mastering document. It was clearly written by someone who knows very little about actual mastering, which I generally find to be the case in the hifi community. So, I'm not trying to make it simple for the readers. I'm trying to make it accurate, which is what this site is about.

Chris

 

[Chris A]

Fine. If you want to argue this then let's have a go.

No.

 

[goat76]

The best master found on Tidal of the ABBA song Should I Laugh or Cry is the one on the Deluxe Edition, at least out of the two versions of the Visitors album they have. The same is most likely the case with the versions found on Spotify.

Deluxe Edition:

The "regular" version:

 

[ThatSoundsGood]

No.

In the interest of Science, clarity and truth, I could not let it go.
I downloaded 3 different versions from Qboz in WAV format (44.1K/16bit). They are all 3 very different. One is mastered quietly (I'm guessing more like an original), one is medium and one is mastered fairly loud. The only one that clips is the medium one and it only clips on snare hits by .1db in the digital realm, which you cannot hear. The loudest master (RMS is much louder and it clearly has more compression and limiting) never clips. It stays 1db BELOW digital clipping.
Now, we can talk about the limiting on it. You can see it is making a bit of a square waveform when it gets loud, which is mostly the bass guitar and kick drum hitting it. This master has a bunch more low end than the other ones. This is a pretty standard way of mastering and while it might look like a square waveform when zoomed in, I would bet a whole paycheck that it would sound better to most people on most systems, which is why the mastering engineers do it. It certainly sounds far better to me.
The medium compressed one (the middle track in the pictures) does barely hit digital zero and technically clips but it's only by .1db. This is inconsequential.
The vocals never come close to being the loudest thing in any of these mixes and never hit the mastering compression or limiting. The kick drum, snare and bass are the loudest and they are the only thing really hitting any compression so we can easily rule out the mastering as the culprit for anything you hear on the vocals except for maybe an EQ thing but that is the mix engineer about 95% of the time. And I will say it again, that it really sounds like an effect thing more than an EQ thing to me. I think it's an early Eventide Harmonizer.
I have attached a couple of pictures of the 3 songs in pro tools. One is zoomed out and one is zoomed in. When you zoom out, the first twol look like square waveforms. When zoomed in, only the loud one looks a bit square in a few spots but most of it isn't even close to hitting the limiters.

The 3 versions are as follows and are in the same order in the pictures:
The top one and loudest master is from Abba Gold (Anniversary Edition)
The middle one with the loud snare is from Thank You For The Music
The bottom and quietest one is from The Visitors Edition (Intro Version)

I rest my case. Please stop blaming the mastering engineers. They're really good at what they do. I'm thinking about writing a piece on how badly the Hifi community misunderstands Mastering because it comes ups all of the time. Do you really think one of the best selling artists of all time would have a master that clips by 2db? Not a chance.

 

[ThatSoundsGood]

The best master found on Tidal of the ABBA song Should I Laugh or Cry is the one on the Deluxe Edition, at least out of the two versions of the Visitors album they have. The same is most likely the case with the versions found on Spotify.

Deluxe Edition:
View attachment 453484


The "regular" version:
View attachment 453483

I believe this is the third and quietest of the 3 versions I downloaded. It certainly has the least compression and limiting. It's so much quieter than the top one! Ha! I'm going back and forth in real time and it just sounds like I'm turning it down by 6 or 8db. I would say that this is probably the most accurate one and those stats pretty much match what I'm seeing in Pro Tools.

 

[goat76]

I believe this is the third and quietest of the 3 versions I downloaded. It certainly has the least compression and limiting. It's so much quieter than the top one! Ha! I'm going back and forth in real time and it just sounds like I'm turning it down by 6 or 8db. I would say that this is probably the most accurate one and those stats pretty much match what I'm seeing in Pro Tools.

Yes, less loud is almost always a good sign. The Deluxe Edition is the quietest one out of the five versions of the song found on Tidal, compilation album versions included.

 

[ThatSoundsGood]

Yes, less loud is almost always a good sign. The Deluxe Edition is the quietest one out of the five versions of the song found on Tidal, compilation album versions included.

While I respect that, I have a different preference on some things. I will say though, I think the least loud and the loudest of the 3 versions I heard are the best ones here. The middle one is not great. For me, it depends on the kind of music I'm listening to. I don't want to hear a jazz record that is mastered really loud, but electronic music is awesome when it's squashed to bits!

 

[komhst]

In the interest of Science, clarity and truth, I could not let it go.
I downloaded 3 different versions from Qboz in WAV format (44.1K/16bit). They are all 3 very different. One is mastered quietly (I'm guessing more like an original), one is medium and one is mastered fairly loud. The only one that clips is the medium one and it only clips on snare hits by .1db in the digital realm, which you cannot hear. The loudest master (RMS is much louder and it clearly has more compression and limiting) never clips. It stays 1db BELOW digital clipping.

Many tracks (especially after 1995) are clipped during the mastering process but then the level is brought down by i.e. -0.1dB or -0.2dB
So, if you extract the track and play it in Audacity you THINK that it's not clipping because the VU meters don't hit 0dB level. In fact they hit the maximum level of -0.1dB or -0.2dB
But if you zoom and take a look at the peaks you will see that every peak at this maximum level of -0.1dB or -0.2dB is clipped.

I rest my case. Please stop blaming the mastering engineers. They're really good at what they do.

No, they AREN'T. Especially after 1995 they DESTROY the music.

 

[Arindal]

Please stop blaming the mastering engineers. They're really good at what they do.

1,000% agree.

The funny aspect is that seemingly many self-proclaimed audiophiles disparaging the mastering engineers, who have themselves never fine-tuned the ratio, threshold and attack of a good analogue compressor themselves. They follow some strange cult of purity and show each other DR figures and waveform screenshots in order to declare certain recordings and mastering versions as ´haram´.

I'm thinking about writing a piece on how badly the Hifi community misunderstands Mastering because it comes ups all of the time.

I have tried so several years ago, and can tell you from own experience that it is the single most futile aspect of audio technology to explain and demo. Even if you practically show people that well-dosed usage of a compressor will make the sound particularly of acoustic (jazz) recordings perceived as more dynamic, they refuse to understand. They have chosen their personal satan in the sphere of audio, and nothing in the world will change that belief.

So better do not waste your time.

The vocals never come close to being the loudest thing in any of these mixes and never hit the mastering compression or limiting.

I did not find any version showing audible signs of clipping in the vocals either. But the lead vocals are recorded and processed with some strange effect I cannot clearly define. As if the voice was recorded with 2 microphones simultaneously with both mixed together leading to some high-frequency phasing or flanging effects. Maybe this is causing a similar issue on some systems. Pure speculation, admittingly.

 

[ThatSoundsGood]

1,000% agree.

The funny aspect is that seemingly many self-proclaimed audiophiles disparaging the mastering engineers, who have themselves never fine-tuned the ratio, threshold and attack of a good analogue compressor themselves. They follow some strange cult of purity and show each other DR figures and waveform screenshots in order to declare certain recordings and mastering versions as ´haram´.

Yes! Exactly! If you have never mastered a record or cannot articulate the difference between mixing and mastering then please stop with the opinions on something you know nothing about.

I have tried so several years ago, and can tell you from own experience that it is the single most futile aspect of audio technology to explain and demo. Even if you practically show people that well-dosed usage of a compressor will make the sound particularly of acoustic (jazz) recordings perceived as more dynamic, they refuse to understand. They have chosen their personal satan in the sphere of audio, and nothing in the world will change that belief.

Totally agree. It's so nice to hear that someone else understands this.

So better do not waste your time.

I landed to that conclusion also.

I did not find any version showing audible signs of clipping in the vocals either. But the lead vocals are recorded and processed with some strange effect I cannot clearly define. As if the voice was recorded with 2 microphones simultaneously with both mixed together leading to some high-frequency phasing or flanging effects. Maybe this is causing a similar issue on some systems. Pure speculation, admittingly.

The effect is an Eventide Harmonizer. It adds 2 more vocals to the original signal and makes them just a tad out of tune with the original to make it thicker. It sounds robotic. I use this kind of effect all of the time on vocalists live to make their voice sound bigger. It can sound like a flanger or phaser since it's probably using something in the time domain to make the slight pitch shift. This would have been one of the earliest models of it so when she sings harder it starts to sound robotic and actually sounds like it's distorting in the 2khz-5khz range. But really, it's just the robotic artifacts of the effect.

 

[ThatSoundsGood]

Many tracks (especially after 1995) are clipped during the mastering process but then the level is brought down by i.e. -0.1dB or -0.2dB
So, if you extract the track and play it in Audacity you THINK that it's not clipping because the VU meters don't hit 0dB level. In fact they hit the maximum level of -0.1dB or -0.2dB
But if you zoom and take a look at the peaks you will see that every peak at this maximum level of -0.1dB or -0.2dB is clipped.

Where on earth are you getting this information? Why do you think they would digitally clip a master then turn it back down? Have you ever mixed or mastered a record?
I literally zoomed in on the peaks of these songs and said that I can see that the bass is hitting the compression but it's not clipping. Do you understand the difference between hitting a limiter and clipping?
If you cannot answer these questions then you're not at all qualified here and you are displaying the misunderstanding that I'm talking about in regards to mastering. Thanks for proving my point.

No, they AREN'T. Especially after 1995 they DESTROY the music.

What happened in 1995? Was it a consortium of mastering engineers getting together to plot their world domination by secretly clipping recordings then turning them back down?

 

[komhst]

Where on earth are you getting this information?

I have seen it many times in countless tracks extracted from original CDs. I'm not talking about mp3s or unknown CD-R copies of audio CDs. I'm talking only about original audio CDs

Why do you think they would digitally clip a master then turn it back down?

I have no idea why...

Have you ever mixed or mastered a record?

Of course I have!

I literally zoomed in on the peaks of these songs and said that I can see that the bass is hitting the compression but it's not clipping. Do you understand the difference between hitting a limiter and clipping?

Of course I know the difference between hitting a limiter and clipping. I'm talking about clipping. Zooming in many tracks, instead of a peak you see a flat horizontal line. And many times this horizontal flat line is not at 0dB, it may be at -0.1dB, -0.2dB or sometimes even lower.
I'm surprised that you have never seen this!!!

 

[ThatSoundsGood]

I have seen it many times in countless tracks extracted from original CDs. I'm not talking about mp3s or unknown CD-R copies of audio CDs. I'm talking only about original audio CDs

Extracting tracks from a CD is much more likely the cause of any clipping than the actual master. It could also be inaccurate metering on the program you're looking at it on.

I have no idea why...


Of course I have!

Can I hear it?

Of course I know the difference between hitting a limiter and clipping. I'm talking about clipping. Zooming in many tracks, instead of a peak you see a flat horizontal line. And many times this horizontal flat line is not at 0dB, it may be at -0.1dB, -0.2dB or sometimes even lower.
I'm surprised that you have never seen this!!!

I have seen this and I have already explained that this is not clipping. If you've mastered anything then you should understand this. The "flat" part you're talking about could be any number of things. It's most likely the limiter keeping it from clipping. Essentially, the limiter is stopping it from getting louder, but it's not "clipping" and this is clear because they keep it at -.01 or -.02 which is more evidence that they know what they're doing. It could also be the nature of the waveform in which the loudest source in the recording is not transient in nature so it gets square where it is loud. This happens with synthesizers and other effects. It could also be a compressor somewhere in the mix chain, which is long before the mastering. But yeah, mastering can cause certain round waveforms to be more square in nature. That's the process of getting the RMS to be louder so that the songs sound better on a wider variety of speakers that they will be played on. Even when the waveform is square there, it still moves the speaker in and out which is what we hear.

 

[komhst]

Extracting tracks from a CD is much more likely the cause of any clipping than the actual master. It could also be inaccurate metering on the program you're looking at it on.

Both EAC and CDex create the same waveform. There is no problem with the software.

I have seen this and I have already explained that this is not clipping. If you've mastered anything then you should understand this. The "flat" part you're talking about could be any number of things. It's most likely the limiter keeping it from clipping. Essentially, the limiter is stopping it from getting louder, but it's not "clipping" and this is clear because they keep it at -.01 or -.02 which is more evidence that they know what they're doing.

If it was from the limiter preventing clipping the shape of the waveform would be a curve or an angle. Since it is a flat horizontal line it's clipping. You can name it as you like, but it is clipping.

It could also be the nature of the waveform in which the loudest source in the recording is not transient in nature so it gets square where it is loud. This happens with synthesizers and other effects. It could also be a compressor somewhere in the mix chain, which is long before the mastering. But yeah, mastering can cause certain round waveforms to be more square in nature.

I'm not talking about synthesizers and electronic music. In this kind of music the waveform could have strange shapes, even square.

That's the process of getting the RMS to be louder so that the songs sound better on a wider variety of speakers that they will be played on.

No. Highly compressed music with minimum dynamic range and high RMS sounds like sh*t in EVERY speaker. Even in the cheapest plastic computer speakers.

Even when the waveform is square there, it still moves the speaker in and out which is what we hear.

It still moves the speaker but it creates an awful and unbearable sound.

 

[ThatSoundsGood]

Both EAC and CDex create the same waveform. There is no problem with the software.


If it was from the limiter preventing clipping the shape of the waveform would be a curve or an angle. Since it is a flat horizontal line it's clipping. You can name it as you like, but it is clipping.


I'm not talking about synthesizers and electronic music. In this kind of music the waveform could have strange shapes, even square.


No. Highly compressed music with minimum dynamic range and high RMS sounds like sh*t in EVERY speaker. Even in the cheapest plastic computer speakers.


It still moves the speaker but it creates an awful and unbearable

We can disagree on the subjective stuff about how it sounds. We can also disagree on the definition of "clipping". Both clipping and limiters can square the wave, but you don't have to believe me. Just try it. It's pretty simple.
What you didn't answer was my question about what you've mastered. I would love to hear it.

 

[Arindal]

The effect is an Eventide Harmonizer. It adds 2 more vocals to the original signal and makes them just a tad out of tune with the original to make it thicker. It sounds robotic.

Thanks a lot for this explanation! I am of course aware of later versions of the Eventide Harmonizer but have never experienced such phasy, flanger-like robotic voice coming out of one. But it absolutely makes sense and is from production point not that far from my initial suspicion that 2 different microphones in close proximity to the singer were used and mixed. It creates a surprisingly similar effect, but not as ´robotic´.

I have seen it many times in countless tracks extracted from original CDs.

Could you name some examples please which are popular and publicly available please? I am into pro audio, recording and producing for decades and have never heard of such case. The last time I identified accidental clipping on a publicly available album, was in the parallel thread Properties of speakers that creates a large and precise soundstage and it was seemingly an amateur piano recording.

The first and most publicly discussed case of pushing the idea of brickwall compression to a degree it would sound like distortion, was Red Hot Chili Pepper´s album ´Californication´. That was done intentionally in this case, but seemingly not very many followers were trying to follow this path of sound aesthetics, at least not in (more or less) acoustically recorded music such as pop, rock and jazz. Surely electronic music is a different case and distortion might be applied to create a certain sound.

It could also be the nature of the waveform in which the loudest source in the recording is not transient in nature so it gets square where it is loud. This happens with synthesizers and other effects.

It happens with several instruments, like distorted electric guitar, or even some brass instruments like a piccolo trumpet.

Since it is a flat horizontal line it's clipping.

Did you see this ´flat horizontal line´ when zooming in to the level of single samples being visible? And how many subsequent samples were of absolutely identical, at which value?

Highly compressed music with minimum dynamic range and high RMS sounds like sh*t in EVERY speaker.

@ThatSoundsGood - q.e.d.!

 

[goat76]

Many tracks (especially after 1995) are clipped during the mastering process but then the level is brought down by i.e. -0.1dB or -0.2dB
So, if you extract the track and play it in Audacity you THINK that it's not clipping because the VU meters don't hit 0dB level. In fact they hit the maximum level of -0.1dB or -0.2dB
But if you zoom and take a look at the peaks you will see that every peak at this maximum level of -0.1dB or -0.2dB is clipped.

Technically, it's done the other way around than how you describe it.

Digital clipping is something I'm sure all mastering engineers avoid. What is technically happening is that the limiter "cuts" the transient peaks and creates a larger headroom, and that larger headroom makes it possible to increase the overall level of the music track. Most mastering engineers increase the level until the highest peaks reach -0.1 to -0.3 under digital clipping.

I don't see anything wrong with you calling the limiting "clipping", as it is a form of controlled clipping which can be made to sound "softer" than the digital clipping that would otherwise occur if the transient peaks were hitting digital zero.