Spectrum Analyzers. Great tool!

Status
Not open for further replies.

Mr. ROUSH

Active member
I’ve seen a lot of negative criticism on the use of spectrum analyzers. That’s really unfortunate. I would encourage newbies to take full advantage. Our ears are the best analyzers we have at the end of the day. That being said, our ears need a reference to establish the ability to identify specific frequency ranges. Frequency analyzers offer visual confirmation. They can be used to help in training your ears.

Even the most experienced professionals should be using spectrum analyzers, especially when experiencing listening fatigue during a long session. They are such a valuable tool with many great free analyzers available, such as Voxengo Span.

Don’t let anyone discourage you from using this most helpful tool!!
 
I'll play devil's advocate here... what is the spectrum analyzer supposed to tell us? That all frequencies are the same level? Is that a requirement for good music? People recorded music for over 50 years without looking at the spectrum, just listening and adjusting levels based on the way the artist wants things to sound.
 
What exactly do good sound look like? What does a good mix have to look like in order for it to look like it's supposed to after its mastered?
 
I'll play devil's advocate here... what is the spectrum analyzer supposed to tell us? That all frequencies are the same level? Is that a requirement for good music? People recorded music for over 50 years without looking at the spectrum, just listening and adjusting levels based on the way the artist wants things to sound.
A spectrum analyzer can tell us a lot of things. It can tell us when our ears or our monitors are deceiving us. It can tell us when the mix is balanced, yes. It can tell us where frequencies are building up and likely causing problems such as masking. It's certainly not a requirement for good music. That being said, it's certainly helpful in making good music. People also used to manually reset the bowling pins, too. But, they figured out a better way to do it that made bowling much more efficient. The technology has changed a great deal in 50 years. Music has changed a lot in 50 years. It's very helpful to compare your mix to a target reference in an analyzer. It can show you glaring inconsistencies that might have otherwise taken some guess work and added time to figure out.

Furthermore, for the number of people making music for that 50 years, how many people were successful at it? A relatively small percentage. How many people would have benefited from having a reference like a spectrum analyzer? I'd venture to guess that a significant percentage of people who are currently making music in the world do not have well trained ears. And, that especially applies to newbies with zero experience. An analyzer can certainly speed up that learning curve and can very much make mixing more efficient.
 
What exactly do good sound look like? What does a good mix have to look like in order for it to look like it's supposed to after its mastered?
Are you suggesting that there are no consistencies between popular recordings relating to the curve you observe in an analyzer? Your argument is anecdotal to an extent. Fact is, many popular recordings have very similar curves, especially within the same genre. Fact is, there is math and science behind it. It's not necessary. It's not that you can't do without it. It's not that it should be relied on or trusted over your ears. It should be used in conjunction with your ears, not in place of.

Point is, it's a useful tool that people shouldn't be discouraged from using. There's no shame at all in taking advantage of the reference. It's no coincidence that most EQ's now come standard with spectrum analyzers built in.
 
I put a SA on a couple of songs I was using as reference mixes - wow, were they smooth. It sure gave me a good way to get my mixes more balanced that my old tinnitus-plagued ears couldn't hear well.
 
While I agree that you mix with your ears, and not your eyes, it never hurts to look at available tools. I don't think a real spectrum analyzer is going to help most folks, but, then, I've never used one of those.

iZotope's "Tonal Balance Control" is a kind of analyzer and it has a few built-in curves (broad bands actually) that you can check your mix against, and they've been set by the folks at iZotope after analyzing lots of music. More interesting is the fact that you can create your own curves from your own reference tracks. I do occasionally pop that plugin into the final (pseudo-mastering) step just to make sure my old ears are not missing something, which they frequently do, and so I probably will try to use it more often - primarily to check for stuff at the ends of the spectrum I have to admit I simply cannot hear anymore...

 
Comparing the spectrum of a raw mix vs. a professionally mastered song is kind of an apples to oranges comparison. It becomes even less useful when used on individual instruments, since you don't have a clean reference.

(Puts on old man hat) since when has the goal been to sound like everyone else? For example, Pantera's Cowboys from Hell and Metallica's Black Album are both the same genre from about the same time, both professionally done, both multi-platinum...don't sound anything like each other.

Using it to find problem frequencies is useful of course.

There are a lot of interfaces that can record at 192k, which is completely useless.
 
I have all sorts of test gear. Scopes, analysers and other gizmos and they are very good for giving me a visual representation of all sorts of things, but they are for me the occasional solution for problems my ears cannot sort - so something not quite 'right' can be looked at.

That said - there is no way in the world that any analyser is an everyday tool for the perfect mix. That revolves around the music - the genre, the tempo, the instruments and the feel of the piece you are recording. Your mention of curve is also indicative of a misunderstanding of what the analyser is displaying. As an example - lets say you have 32 bands that are converted to a level on a display. Each band is then shown as a peak or an average or other commutative level - this is not the same as the instantaneous level, and means that if you play loudly a G and an A on a synth, those notes will be added together and a peak created that is NOT there. If you have an analyser that has a much finer solution and sample rate, the G and A will be two separate events, at lower levels. This ignores all the harmonic products, so while an analyser can show you 'holes' in the spectral assignment of the music - this observation needs to be treated with a link to the music being analysed.

The idea that there is some kind of spectral fingerprint that says good or bad is ridiculous. If you take the spectral content of 80s synth pop they are similar - because the content is similar. If you look at the Beatles does their spectral fingerprint differ from the stones? Yep - it does!

Having a mix that won't let the baseline show through can benefit from a look at the spectral distribution and to the skilled observer, reveal what is going on, but it's a fairly blunt tool and every analyser will give different results. I have a few on my phone - they all display the same music differently, because they work differently.

Newcomers need to develop their aural acuity - and learning by reliance on a tool, is a bad move because you are basing mix decisions on what you are seeing, not hearing, and your hearing might well be the correct decision. when you are learning, you need to gain experience and a gadget with uncoordinated results is bad.

For problem solving, they're great but I would NEVER use one for my first mixes. I might use my expensive one for real problems. I might use the one on my phone to check where the main energy is, but my primary tool is my ears. I do not need a LUFS meter while recording and I don't need to constantly check the tuning on the guitars. I don't need to use a display to tell me where that EQ boost is needed, because my ears come close enough.

This bit had me falling over laughing!
Furthermore, for the number of people making music for that 50 years, how many people were successful at it? A relatively small percentage.
I've been in recording studios all my working life - all genres, both sides of the glass and I have done it in TV too, and in my 45 odd years, it's only recently I hear people extolling the virtues of analysers, and until they started being built into the desks, I saw maybe a handful of people use them. why? because they did not need to and frankly - I do not believe you can remotely have a typical curve for music. In the seventies we started the smiley face in people's 7 or 9 band graphic EQs in their hi-fis.

If you want to use an analyser to produce music that's fine, but I'm very wary of using them as an everyday tools in production.
 
While I agree that you mix with your ears, and not your eyes, it never hurts to look at available tools. I don't think a real spectrum analyzer is going to help most folks, but, then, I've never used one of those.

iZotope's "Tonal Balance Control" is a kind of analyzer and it has a few built-in curves (broad bands actually) that you can check your mix against, and they've been set by the folks at iZotope after analyzing lots of music. More interesting is the fact that you can create your own curves from your own reference tracks. I do occasionally pop that plugin into the final (pseudo-mastering) step just to make sure my old ears are not missing something, which they frequently do, and so I probably will try to use it more often - primarily to check for stuff at the ends of the spectrum I have to admit I simply cannot hear anymore...

Izotope makes some great tools. I read that they recently improved the algorithms for the Tonal Balance feature on the low end. I've been wanting to try out the latest versions. Also, even young ears get worn out. It's a useful tool for all ages!
 
Comparing the spectrum of a raw mix vs. a professionally mastered song is kind of an apples to oranges comparison. It becomes even less useful when used on individual instruments, since you don't have a clean reference.

(Puts on old man hat) since when has the goal been to sound like everyone else? For example, Pantera's Cowboys from Hell and Metallica's Black Album are both the same genre from about the same time, both professionally done, both multi-platinum...don't sound anything like each other.

Using it to find problem frequencies is useful of course.

There are a lot of interfaces that can record at 192k, which is completely useless.
You may have misunderstood me. I don't recommend comparing an unmastered mix to a mastered reference. However, I would recommend comparing a mastered version of your mix to a target reference. In FL studio there is a "C" fader. It stands for "current." It monitors whatever track you have selected at any given time. I keep Span on this track. I find it useful for individual instruments, too. Gives a bit of a visual reference relating to effects that are added and what it's doing to the signal.

Honestly, I think it's pretty common that mixing engineers strive to achieve results similar to popular recordings. Certainly doesn't have to be the goal. But, for those who do have that goal, a spectrum analyzer is simply useful. It's useful in general. It's not the greatest thing since sliced bread or anything. Just useful. As for your reference, those records were much earlier on in the "loudness wars." Not near as much dynamics in most modern recordings. For those trying to compete, the target is smaller now. Doesn't hurt to have the extra tool at your disposal. Also, better to have and not need than need and not have.
 
I have all sorts of test gear. Scopes, analysers and other gizmos and they are very good for giving me a visual representation of all sorts of things, but they are for me the occasional solution for problems my ears cannot sort - so something not quite 'right' can be looked at.

That said - there is no way in the world that any analyser is an everyday tool for the perfect mix. That revolves around the music - the genre, the tempo, the instruments and the feel of the piece you are recording. Your mention of curve is also indicative of a misunderstanding of what the analyser is displaying. As an example - lets say you have 32 bands that are converted to a level on a display. Each band is then shown as a peak or an average or other commutative level - this is not the same as the instantaneous level, and means that if you play loudly a G and an A on a synth, those notes will be added together and a peak created that is NOT there. If you have an analyser that has a much finer solution and sample rate, the G and A will be two separate events, at lower levels. This ignores all the harmonic products, so while an analyser can show you 'holes' in the spectral assignment of the music - this observation needs to be treated with a link to the music being analysed.

The idea that there is some kind of spectral fingerprint that says good or bad is ridiculous. If you take the spectral content of 80s synth pop they are similar - because the content is similar. If you look at the Beatles does their spectral fingerprint differ from the stones? Yep - it does!

Having a mix that won't let the baseline show through can benefit from a look at the spectral distribution and to the skilled observer, reveal what is going on, but it's a fairly blunt tool and every analyser will give different results. I have a few on my phone - they all display the same music differently, because they work differently.

Newcomers need to develop their aural acuity - and learning by reliance on a tool, is a bad move because you are basing mix decisions on what you are seeing, not hearing, and your hearing might well be the correct decision. when you are learning, you need to gain experience and a gadget with uncoordinated results is bad.

For problem solving, they're great but I would NEVER use one for my first mixes. I might use my expensive one for real problems. I might use the one on my phone to check where the main energy is, but my primary tool is my ears. I do not need a LUFS meter while recording and I don't need to constantly check the tuning on the guitars. I don't need to use a display to tell me where that EQ boost is needed, because my ears come close enough.

This bit had me falling over laughing!

I've been in recording studios all my working life - all genres, both sides of the glass and I have done it in TV too, and in my 45 odd years, it's only recently I hear people extolling the virtues of analysers, and until they started being built into the desks, I saw maybe a handful of people use them. why? because they did not need to and frankly - I do not believe you can remotely have a typical curve for music. In the seventies we started the smiley face in people's 7 or 9 band graphic EQs in their hi-fis.

If you want to use an analyser to produce music that's fine, but I'm very wary of using them as an everyday tools in production.
Your initial statement is more along the lines of the way I have suggested using analyzers. I never said it was the key to a perfect mix, but rather a helpful tool to have in the toolbox. You've assumed a lot about my understanding.

Feel free to point out at what point I said that it was in any way a definitive visual indication of good and bad. What I did say is that it can help newbies identify glaring inconsistencies. It can certainly give a visual indication of bad things happening in a mix. As I've said from the beginning, it's a useful tool.

Various analyzers may yield slightly different results, but I don't use multiple analyzers regularly. I have one or two that I like to use. And, I appreciate the plugins that include them in their GUI. Essentially, you're giving an example of how it can be useful to analyze the low end. A blunt idea is better than no idea, and that's all I've ever suggested.

Not to be repetitive, but again, you've suggested that I've in any way implied that anyone should be reliant on this tool over their ears. I have in fact said the opposite. Again, it's simply useful as a reference. You may not benefit from it. That's great. You don't need it! However, to suggest that it's not a useful tool is demonstrably false.

When used appropriately in each situation, there's no reason to discourage anyone from using them on any mix, first, middle, or last. Also, I'm just not a proponent of the notion that things should always be done the way they've been done. I mean, no reason to fix what isn't broken. However, that definition varies broadly for many different people across all experience levels in mixing. Again, probably not that useful for you. Very useful for many others and something that I would never discourage.
 
Here's another great way to use a spectrum analyzer. Turn on the "side" overlay to view the differences between the mid/side of the mix. It's very useful to ensure that stereo effects have not caused the side mix to eclipse the mids. It's also a good way to be sure that stereo information is staying clear of the low end. Again.. useful.
 
I put a SA on a couple of songs I was using as reference mixes - wow, were they smooth. It sure gave me a good way to get my mixes more balanced that my old tinnitus-plagued ears couldn't hear well.
It has certainly helped me to bridge the gap, too! I haven't been great to my ears over the years! lol
 
It's a another good tool....not a be-all end-all solution. And I know no one here is saying it is. As with most of the audio products we use to listen to our mixes with......monitors...speakers....headphones......etc......"learning it" then using it is the way to go. Can you mix or master without it......sure.

Mick
 
I agreed with your first tool - a great way to identify frequency ranges for EQ for treatment. What I don't get are the bits about a target reference. This is where we move apart. I just can't see that there is such a thing. Much we agree on, but [quote[ Fact is, many popular recordings have very similar curves, especially within the same genre. Fact is, there is math and science behind it. It's not necessary. It's not that you can't do without it.[/quote] worries me - if you look at spectral plots - and I use Spectral Layers quite a lot, I just don't see these commonalities.
 
I agreed with your first tool - a great way to identify frequency ranges for EQ for treatment. What I don't get are the bits about a target reference. This is where we move apart. I just can't see that there is such a thing. Much we agree on, but [quote[ Fact is, many popular recordings have very similar curves, especially within the same genre. Fact is, there is math and science behind it. It's not necessary. It's not that you can't do without it.
I would use the example of "matching" features in some popular plugins such as Ozone and T-Racks. These features use spectral analysis in an effort to give your mix/master a similar curve to the reference track you provided. They work pretty well in some cases. Also, in some cases, it ends up sounding nothing like the reference. The ears should always be the final decision maker. Mainly, my point is that drastic differences can offer a lot of indications of where your mix is deviating from your target. End of the day, each person should decide if it's useful for them. I find it quite useful.
 
Izotope Tonal balance is practically a spectrum analyzer with an intelligent hold time, it's great though!

I use ProQ2 on the master as my spectrum analyzer, it's great for that! It has a tilt enabled by default for that meaty low end/smooth highs if you get a relatively flat line. It's a great middle ground, you can go either way a bit but at least the analyzer gets you bloody close.

I don't need to use it anywhere near as much anymore, I use it nowadays mainly just to confirm that I didn't miss the mark, after just using my ears only. But I still find it essential for internalising what a balanced mix should sound like..... it's a way more reliable way of learning your speakers than that old method (listen to lots of music you know well on your monitors). Because let's face it. We can't trust our ears really can we, not when we are learning this stuff. Many a time I have been working too long, think I have a tonally balanced mix and then use the analyzer to find that the 9khz pick attack is leaping out of the speakers. THEN! my ears will hear it as a problem. But the analyzer told me first. Happens quite often.

I'd pick up on that problem straight away if I listened to the mix on my laptop, or if I came back with fresh ears. But to me this is just another win for visual aid.

The Spectrum analyzer has been the single most important tool in helping me understand how to mix, and to what a good mix behaves like within proQ2 only. I used to use SPAN but ProQ2 is where its really at for me.
 
Why is flat, good? We all know that every system we listen on is different sounding. It sounds like you don't trust your ears and want the support and comfort of the analyser?

Live sound wise - I totally get it, when the analyser was overlaid on the screen on my live sound gear, my guess the problem was 2.5K would often reveal it as the harmonic on 5K - that was useful and being able to see it I totally understand. In the studio though, I'm doing subtle, not radical EQ in the main, and with a mix building up, I really need to concentrate on my ears and not get sidetracked by seeing some empty space at 6K and wondering if I should fill it.

I suppose we all have our own system. what worries me is that new people see these kinds of posts and believe they need to get one because without one, your mixes will always be lacking.
 
There is a tilt so it's not (really) flat, just visually.

when everything is going at once flat (visually) s good IMO. When the vocal is not playing then I am looking for a slight dip at 1k ish, when im listening to a bass/kick heavy song i'm looking for a low end bump. Orchestral/score i'm looking for mid heavy and hardly any highs. etc.

I have a love for metering/analyzing tools, and this is the only place where I don't mind spending money on plugins because it saves you hours and helps you understand what is going on. Otherwise I do try to favour stock for everything else.

I most certainly do not trust my ears no. I don't nail pro sounding mixes everytime at all.
 
Status
Not open for further replies.
Back
Top