I have noticed that the default recordings in BIAB seem to use a +6 dB per octave slope (which also seems to be the slope a lot of jazz music use. Also, the Yamaha synth I use also defaults to about +6 dB per slope. And even some rock-recording multi-tracks I have remixed was delivered at +6 dB per octave)

Meanwhile a lot of modern pop (mastered for radio) seem to go for a slope as bright as +3 dB per octave.

For a long time I used +4.5 dB per octave for my own music, but have lately tried mastering a few at +3.

Personally I find +6 dB too dark for my taste, but not being as young as I was I'm wondering if perhaps my ears are just losing their ability to hear the bright frequencies as good as they used to.

Which slope do you generally prefer?
+6, +4.5 or +3 (or something else perhaps)

Hi JEL,

I'm sure there are people on here who know exactly what you mean. Unfortunately, I'm not one who has such a level of audio engineering proficiency. I would like to learn, though. What do you mean by +6 dB slope per octave and in what program?

Is it related to the -6 dB frequency rolloff found in Ozone5? I always use this reference line. It's represented by the grey line below (the pink line is a -3 dB roll off). I don't know of anything that has a +ve slope.



Regards,
Noel

@Noel96;

Hi Noel

It's the slope relative to the slope of white-noise.

White noise is the same dB level through the entire frequency spectrum, which makes the slope look flat when viewed at '0 dB per octave' (0 dB change in how you project the slope per octave, meaning a signal with the same dB value will look the same height across the entire frequency spectrum. For example; a -18 dB bass signal is as high as a -18 dB cymbal signal)

For music that slope is too bright though, so music (viewed on a '0 dB per octave' slope) tends to look like it goes downward as you move from low to high frequencies (the high frequencies look like they have less power, because they do have less power normally)

So on spectrum plugins for music you can usually adjust how you want the slope projected on the screen, so that you can more quickly/easy see if your music fits your desired slope.

So to have the music-slope appear flat, you actually make it go upwards as it moves to the higher frequencies. Hence it's expressed as a plus dB value.

A +6 dB per octave slope means that 2 signals displayed an octave apart will appear at the same height if the signal one octave higher is 6 dB less powerful. 2 octaves above it will have the same height if it is 12 dBs less powerful.

So since treble signals are less powerful, this +6 dB slope will make music look flat across the spectrum if the music is mixed/mastered toward that specific brightness.

If you project a white noise signal on this slope, it will look like it's rising as the frequency increases.

So basically, the slope is used to determine how bright your sound should be.

Music that appears flat on a '0 dB per octave' slope sounds too bright, if it looks flat on a '+6 dB per octave' slope it sounds a lot darker.

The slope on your image is probably (I have never used Ozone, so I don't really know how it works) set to a specific value somewhere in Ozone's settings.

I'm not good at explaining, but I hope it makes sense

Thanks, JEL.

Your explanation made perfect sense to me. I was not aware of this relationship before although when I consider it in relation to Fletcher-Munson equal loudness curves, I can see what you are saying.

Thank you for taking the time to offer some added insight. I appreciate it greatly.

Regards,
Noel

JEL,

I hope you don't mind me hijacking the thread for this post but I thought that I'd better explain what I mean in relation to the Fletcher-Munson curves for anyone who hasn't come across them before.


Fletcher-Munson Curves
The graphs below show a collection of Fletcher-Munson curves. In the study of psychoacoustics, these are also known as Equal Loudness Curves.



The above show how the average human hears the loudness of music when different pitches are compared. I've chosen the 80 phon line (circled in blue) as this would be a fairly standard level of loudness.

• For comparison: normal conversation is around 60 dB; city traffic noise is around 80 dB.

I've also marked two 2-octave jumps (in green). These two sections can be regarded as almost linear. These are 50 - 200 Hz and 200 - 800 Hz. (The x-axis scale is logarithmic and unless one has had practice reading log scales, it's not always clear to see where the values lie. That's why I added the green lines.)


What the 80 phon curve above shows is how many decibels different sound frequencies need to be set at so that they are heard as equally loud. The reference frequency is 1000 Hz set to 80 dB - hence the name 80 phon.


How to interpret Fletcher-Munson Curves
For a frequency of 200 Hz, the 80 phon curve gives a y-value of approximately 85 dB. This is saying that, for the average person, a 200 Hz sound at 85 dB will sound equally loud to a 1000 Hz sound at 80 dB

If a bass note of 50 Hz is played, it will need to have a loudness of around 102 dB to sound as loud as a 1000 Hz sound at 80 dB.

By comparing the 200 Hz and the 50 Hz values, it's possible to see that a 50 Hz sound needs to be 17 dB (102 - 85 = 17) louder than a 200 Hz sound at 85 dB if it is going to sound equally loud. Since 50 to 200 Hz corresponds to a jump of two octaves, 50 to 100 Hz is one octave and 100 to 200 Hz is the second octave, then the decibel adjustment per octave could be approximated by 17/2 = 8.5 dB.

Applying this information...

Let's say that my sounds are all electronically set to 85 dB. The above information tells me that if I want my bass to sound equally loud to my mid-frequency harmony (around middle C), I'll need to increase the low frequency volumes by around 8.5 dB per octave (from 200 to 100 Hz and then from 100 to 50 Hz).

Using a similar argument, if I have a 200 Hz note at 85 dB (instrumental or sung), then to get a note at 800 Hz to sound equally loud, it would need to be at around 79 dB. This corresponds to a reduction of 6 dB over the two octaves. This means that I would need to lower my upper frequencies by around 3 dB per octave. (I suspect that these explanations are part of the reason why Ozone versions 5 and 6 have roll-off guides for 6 dB and 3 dB.)

Using a parametric equaliser, the above information now allows me to begin devising a strategy for shaping my overall sound so that every frequency can sit comfortably in the mix.


Regards,
Noel

@Noel;

Thank you

Great detail

Blending the 6dB and 3dB slope from your measurements would explain the 4.5dB slope I am seeing in the plugins I use. I guess it may have been chosen as a rough average between the two. It certainly fits the math of it.

I am still not sure which brightness people would prefer though, if a single linear slope is to be used.

I suppose the ideal would be if the spectrum-display could weigh the signal according to the equal-loudness curve, such that a signal that fits that curve would appear as flat across the spectrum-display.

But of course that would only work for one single sound-level, and since people listen to music sometimes loud sometimes quiet we need to find some middle-ground that covers both situations. A slope that sounds good (not too dark, not too bright) both when played loud and not so loud.

According to the equal-loudness curves, music played at lower volumes should have a darker curve than music played at louder volumes. That might indeed explain why modern pop is using +3dB while jazz is using +6dB (people who listen to pop (kids and young people, crude generalization) play louder than people who listen to jazz (people with a bit more 'life experience' Again; crude generalization))

Thanks again Noel. Reading your post gave me some more perspective on this
It would appear there is no 'single-slope-fits-all' solution out there

Great question and conversation.

Unless I am understanding the coversation incorrectly one item neither JEL or Noel has mentioned is how each music genre has its own generally accepted, or standard, mix frequency curve. One piece of advice repeatedly given to beginner mixers is to compare and match the frequency curve of their song to the frequency curve of a hit song of the same genre.

Though RealTracks may not exactly match the frequency curve of the genre that inspired each RealTrack, there are audible differences in the sound of an electric "jazz" guitar RealTracks and a "metal" guitar RealTracks.

@Jim;

Interesting observation

I have only been through realtracks of the genres that had the style/sound I wanted, so I haven't actually listened closely enough to some of the other realtracks to notice the differences in brightness that you mention.

I AM actually getting a bit worried now that this is turning more into science than 'emotion'.

It's easy to get caught in the science parts of sound, and lose sight of what music is fundamentally about: emotions

But at the same time, the science part DO make a difference. A good song may get great if the brightness is spot on for that type of music, or a great song may be considered only 'good' if the brightness is off what it ideally should be.

Maybe it's true; sound-engineers and musicians need to be 2 different people, rather than being 1 person trying to master both disciplines.

Hi Jim,

You're absolutely correct. Different genres will certainly apply the information in different ways. For example, Electronic Dance Music (EDM) is bass-heavy so the bass boosting would be above the value determined by the equal loudness curves. A trio of jazz musicians, on the other hand, tend to be bass-light and so the bass boost (if any) would most likely be less than determined by the above set of curves.

As I said in my post, I'm only a novice. My above thoughts only deal with the concept of 'equal loudness'.

All said and done, the curves simply provide a starting place because it's easy to read approximate decibel values from a track using VU meters and then relate that information to the curves. In addition to this, PG Music provides a frequency analyser plugin with BIAB/RB that can be used to get additional useful information.

For those interested, Ozone has a feature that lets it use a sample of music in a specific genre and approximately match a particular mix's frequency responses to that sample. See the link below. It's a pdf found at iZotope. (Read the section under "Snapshots" on page 24.)

Mastering with Ozone 5 (5.2 MB, found freely available at the iZotope website)

Those interested in this thread might like to investigate the document from iZotope. Pages 15 to 28 relate to this discussion on EQ'ing

Regards,
Noel

JEL,

With music, 'emotion' always sits in the driver's seat. Science is only beneficial inasmuch as it: (a) gives some analytical tools that may be useful; (b) provides someone whose hearing is not 'average' a way to approach mixing. At the end of the day, if it doesn't sound right to an average listener, it ain't right - simple as that

Noel

I think I agree with Jim here. I cannot say I am unfamiliar with the technical, curves and -/+db per octave. I just use this rule of thumb to mix/master. Does my mix, sound like/curve like, a Nashville hit, Pop radio mix, EDM etc. I don't have any luck with matching actual curves. I think its because I was never good with math. I'm also a little bit visually impaired. If you can do it I think you are much better than I am.

In case anyone is interested...

Here's a link to EQ matching using Ozone 5

• https://www.youtube.com/watch?v=bBHmeFAIKnY

And here's a link EQ matching using Ozone 6

• https://www.youtube.com/watch?v=1XNPJqQ_4NM

I think frequency analysis is just another tool. People are most comfortable with things they are familiar with. If I want to make a "classic, Nashville sounding" country song I would want the mix to resemble "the Nasville sound" as close as possible. Frequency analysis helps me to visualize that goal.

Other things I would need to think about are instrumentation, song arrangement and how effects are used. Creating a new Hank Williams sounding song would require a completely different approach than creating a new Blake Shelton sounding country song. Though I haven't looked I'm sure the spectrum slope would be different for each song.

For myself I find the preproduction stage to be the most pleasurable stage of song development. In the beginning there are no boundaries or limitations other than just how creative do you want to be. Anything is possible. The challenge is in duplicating what you planned.

Do any of you have a song-link you could post, to a song of your own that, in your opinion, is just the correct brightness (or at least the brightness you prefer for that genre)?


It would be interesting to see if you aim for the same brightness or if you have different preferences.

Feel free to have a listen to any of my songs found at the links in my signature. Be warned, though, that mixing properly is still a work in progress for me! I have heaps to learn!

I also have a link to my musical highlights in my signature line that ultimately links to Soundcloud uploads.

The majority of my musical postings are either historical or experiments. What I consider my most accomplished effort so far is ++ Down In The Valley ++

Though I have been mixing for some time the amount of time does not accurately reflect my lack of effort! I am a music production beginner. I still am learning the basics.

You guys are amazing. I doubt I could have heard many of these subtleties 50 years ago when I had good ears. I've long passed the threshold where MP3's sound great. Just give me some presets and I'll pretend I know what I'm doing

Bud

Interesting topic. I have never given it much thought as I always just used my ears to get the best mix out of the stuff I record. What (free) plugin do I need to be able to select the required spectrum slope. The frequency analyser I use at the moment (blue cat) doesn't allow for that.

I found Voxengo Span, which covers lots of adjustments on the slope.

Bud,

Whatever you do, you do it REALLY well! Your mixes sit very comfortably in the the group "best mixes in the forums". It's always a learning experience for me to listen to your work.

Noel

Sinbad,

The Voxengo plugin looks interesting! Thanks for posting about it.

Noel

That's very kind of you Noel. Off and on for 40 years I played upright bass in bluegrass bands standing in the middle of the stage looking at the backside of 4-5 guys/gals. My whole mixing perspective comes from that - trying to build a live sound with minimal tracks. Get me into more than a few tracks and I'm lost. I have huge admiration for those who can mix lots of tracks and get it all to sit right. Pardon the somewhat off topic ramble.

Yes, Voxengo Span is a really great tool




And then there's tinnitus That really doesn't make mixing/mastering any easier either.

Yep. Tinnitus and a high frequency melt down make it a challenge for me. Fortunately I have friends with great ears that I can get opinions from. Particularly my best best friend... my wife Janice who can hear a humming bird burp 50 feet away during a thunder storm.

How about a gnat fart in the woods?

You forgot to mention the distance. Could make a difference.

You're a lucky man Bud

"It's no fun when you're doing it solo", as a popular TVshow theme-song goes

Yep, I won life's lotto 35 years ago when she said "I do!"