How do you use a spectrum analyser

[ForTozs]

Hi. Newbie here. I am currently trying to sweeten up a live show that my band did a few weeks ago. We had a great setup with a beautiful Mackie 32 channel 8 bus console using 20 channels recorded on an Alesis ADAT-style HD recorder. We had a sound guy in charge of it all and the raw tracks turned out beautifully. Anyway, I have only played around with mixing/mastering for fun, but I want to make a semi-professional sounding (but not overproduced) CD that I can show to bar managers and such to book gigs. I have good ears and I'm good with computers and software. I am using Cubase SX and have a variety of VST plugins (Sonnox, IK, TC, etc.) and want to do as good of a job as I can on this. My mixdowns have turned out alright but I think they can be better. I think what I lack most is monitors. I've been using a combination of cheap headphones, computer speakers, home stereo system, and a PA system for my mixing/mastering. The PA system probably represents my mix the best since I have EQed it with its multiband EQ. I check my mixes in the car, since I trust Honda stock sound systems and they are what I am most familiar with. I am still having some translation problems. Here is my question. I have been using T-racks on the main mix for "mastering" and it has a multiband sprectral graph. How can I use this to its full potential? Am I basically trying to make the spectrum as flat as possible? For what frequencies (e.g. 50-20k)? Can good spectral analysis replace good monitoring? Are there any better plugins or programs out there for this purpose? I really have a million questions about all this. Thanks.

[Cellotron]

Originally Posted by ForTozs

I have been using T-racks on the main mix for "mastering" and it has a multiband sprectral graph. How can I use this to its full potential?

By mainly leaving it off and using your ears instead. The only time I use a spectral analyzer is to find specific frequencies of a single note or resonance that is sticking out that for some reason I am not able to hone quickly in on just by ear. 99% of the time here it is left off though.

Am I basically trying to make the spectrum as flat as possible?

No - not at all - you are trying to make things sound good. What sounds "good" can result in lots of variance in the way its spectrum is displayed - there is no correct way for good sound to "look"!!!

Can good spectral analysis replace good monitoring?

Absolutely not, never, in no way at all, not even a little bit.
Music is appreciated through the ears - not the eyes!
Leave the pretty animations off - and listen!

Best regards,
Steve Berson

[bobwarren]

Listen to the man. Steve is telling the truth!

[ForTozs]

Thanks. I'm glad I didn't spend all afternoon trying to learn how to discern a spectral graph. Any advice to someone without monitors? Is there anything better than my current burn, listen, burn, listen method? I am going through CDs like nobody's business.

[bachconnelly]

Its generally tough to EQ something you are unable to hear.. I would suggest returning the remaining blank CD's you've purchased and buy any type of monitoring system.. Without SOMETHING to listen on your method seems to be the best..

[ONDRAY]

for FUN i played around in mastering a tune poorly using a spectrum analyzer, a phaze scope and meters... ... . Seeing and hearing are two totaly different senses, there's just NO replacing one for the other.

[sHOWpONY]

Originally Posted by ForTozs

Hi. Newbie here.... Thanks.

Hi ForTozs
True wisdom has already been spoken, but here's a suggestion you might find useful with your SA.
Find some great sounding tracks and poke them thru the SA and see how they look. Conversely, poke some shitty sounding mixes thru and see how they look. Set your SA to 300ms response time, turn on a peak hold and have the graph set at 1/3 or 1/6 octave. Generally, you'll see differences between really good and bad sounding material from an EQ perspective.
You might find it harder to 'see' transients.
It's also useful to look at the SA with the mix divided into mid and then side modes separately. Run great sounding mixes thru this way and see how they look. Keep in mind this is only a rough guide to point a 'newbie' in the right direction. I'm also simplifying this, but I think you'll find this process useful.
Chuck our SA as the last plug on your master buss and refer to it when mixing.
Dothis AND use your ears.

Hope this is useful for you.

[Storyville]

ABSOLUTELY DO NOT USE YOUR SPECTROGRAPH. DO NOT, DO NOT, DO NOT.

If you are trying to eq your mix and you compare one spectrograph image to another song's spectral image you will f*ck up your mix. Guaranteed. The only thing a spectral image is good for is finding a problematic frequency band really fast.

My suggestion, mix on the best headphones you can get. Do not try to be as loud as a commercial record. Just try to imitate the sound. Or, pay an ME as a consultant, or just have an ME do it. ME's tend to be pretty kind toward people in your shoes. You could probably get a three song demo mastered up for two hundred bucks.

[Online-Mastering]

Quote:
Can good spectral analysis replace good monitoring?
Absolutely not, never, in no way at all, not even a little bit.
Music is appreciated through the ears - not the eyes!
Leave the pretty animations off - and listen!

That's so true. BUT: For anybody who is mixing at home under NOT so good listening conditions and with small (multimedia) speakers or little speaker that roll of at 60 or 75 Hz already, the use of an analyser could indeed be helpful...

Not mixing-wise (Everything about that was said in this thread already ), but to check some very basic things concerning the low end (rumble, hum and ultra deep bass) going on. Frequency content that cannot be heard on a specific (small) home studio set-up. But this low end takes energy out of the mix (you will not get it to good loudness) and leads to distortion on different types of playback devices. Therefore it might be helpful to have one running, in this case...

2) The headphone tip is a very good thing: It gives you an easy and cost-effective way to start and get to good results. Listen to reference tracks on your headphones and compare them soundwise with your mix (in terms of: sounds it equally balanced or does my mix sound MUCH brighter or does it have A LOT more low end?). If your own new metal EP sounds much more powerful than Linkin Park or way better than the Black Metallica album, just for example, than someting may be particularly wrong

The rest is training your ears, learning how GOOD mixes translate on your headphones. The better you their frequency response, the better mixes you will achieve in the end. Try to listen to the fine details as well as to the sound 'in general'. Try to listen to the stereo information (what's going on in the side channel). Listen, listen, listen. And if you think you're done: Listen again

Arne

[Edward_Vinatea]

I knew this was coming sooner or later. RTAs are powerful tools often misunderstood and dismissed by many audio engineers who all they see is a 'Christmas lights' show for their studios. The truth is it will take you months if not years to really grasp its power. I have spectrum analyzed over 20,000 songs (I've lost my count) and always run one one every mixing and mastering session.

So - what is the presumed advantage over someone who doesn't read a RTA? I can master something in minutes with dead accuracy and unless someone doesn't approve of the 'color' or the amount of 'crispiness' a master gets - I usually nail the sound on the first try. I can smooth out harshness exactly where is needed, allow high transients to pass when necessary and other benefits that I will skip. However - these decisions are not arbitrary but based on sound engineering experience (in my case 25 years).

Thus - the power of the RTA is not complete without experienced ears to match. You can either start now and develop the skill simultaneously or if you are already an advanced audio engineer begin the process of learning to interpret sound on vision. One of the main and immediate advantages would be creating a perfect bass at the sub sonic level and never have to jump into your car to play it on the stereo system.

Regards,

[24-96 Mastering]

To judge the amount of bass, imo, better than an RTA or listening in the car is having a monitoring system and a room that can reproduce the lows sufficiently and accurately. An RTA or a listen in the car may be useful as an additional pointer anyway, but that's more for the engineer's confidence than it is really a superior "technique".

It often gets written that RTAs are powerful tools that actively enable you to do better work than you possibly could without them, that it opens up a world of possibilities (think Har-Bal marketing, for example). I disagree with that statement to some extent, but most of all with the statement's intent.

I personally think they can be very useful, but mostly to provide pointers, not really to empower you in entirely new ways. What is often written (and sold as exciting or controversial) with regards to analyzers, imo seems either incredibly trivial (been there, done that) or misleading (like an RTA can make up for a lack accurate monitoring). Using RTAs for additional pointers is exactly what everyone (or let's say most people) use RTA for. Many MEs like to have an RTA running - me included, by the way. Some don't. That doesn't mean they don't understand the idea or are in any way old-fashioned. It's not rocket science, no RTA is reinventing the wheel those that don't run an RTA aren't necessarily fools that haven't seen the light.

I believe that RTAs are not really misunderstood by most of those opposed to their use. In the multiple threads on the topic, most of the posters (it's safe to assume) have been there and done that. And whatever their personal take on RTAs is, whether they like to work with it or without it: when someone asks "how a song should look" (which is the usual question that starts this kinda thread), it's very, very much worth pointing out that asking that question indicates that they're on a wrong course.
If you want to learn to read a meter, you have to learn what correlation the displayed information has to what you're hearing. It doesn't work the other way around. Your hearing will pick up much more than an RTA could display. This will be a continuous process. You're not done after 100 or 1000 albums. An RTA makes no technique to be mastered, it is simply an indicator that can be read and interpreted. But what you're hearing is the real thing, so if in doubt, trust your ears.
Use RTAs/meters for reassurance, for pointers, for security. IMO, only if your monitoring is far from ideal, an RTA may provide information that you can't at all get from your hearing - and then it may be beneficial to blindly rely on it. But in such a scenario, the RTA's importance is only given by your sub-standard monitoring, so rather than buying a better RTA, buy some better monitoring.

[leddy]

Ears and eyes do not have to be mutually exclusive. Good monitoring is obviously critical. A spectrum analyzer is a tool that can be helpful. Does it need to be more complicated than that?

[Storyville]

Originally Posted by Edward_Vinatea

I knew this was coming sooner or later. RTAs are powerful tools often misunderstood and dismissed by many audio engineers who all they see is a 'Christmas lights' show for their studios. The truth is it will take you months if not years to really grasp its power. I have spectrum analyzed over 20,000 songs (I've lost my count) and always run one one every mixing and mastering session.

So - what is the presumed advantage over someone who doesn't read a RTA? I can master something in minutes with dead accuracy and unless someone doesn't approve of the 'color' or the amount of 'crispiness' a master gets - I usually nail the sound on the first try. I can smooth out harshness exactly where is needed, allow high transients to pass when necessary and other benefits that I will skip. However - these decisions are not arbitrary but based on sound engineering experience (in my case 25 years).

Thus - the power of the RTA is not complete without experienced ears to match. You can either start now and develop the skill simultaneously or if you are already an advanced audio engineer begin the process of learning to interpret sound on vision. One of the main and immediate advantages would be creating a perfect bass at the sub sonic level and never have to jump into your car to play it on the stereo system.

Regards,

Congrats, but notice the part where you mention the 20,000 songs. Obviously you know the correlation between what you see and what you hear. And you know that what you see can be misleading. The guy who started this thread clearly has not mastered 20,000 songs, and relying on a visual representation when you are not familiar with what you are looking at is a really really fast way to mess up your mix.

[Edward_Vinatea]

Originally Posted by Storyville

Congrats, but notice the part where you mention the 20,000 songs. Obviously you know the correlation between what you see and what you hear. And you know that what you see can be misleading. The guy who started this thread clearly has not mastered 20,000 songs, and relying on a visual representation when you are not familiar with what you are looking at is a really really fast way to mess up your mix.

I welcome valid points when given. But it's really frustrating to advise the OP and for that matter anyone who cares to respond if they fail to read all my comments. I believe I did say that it will take months if not years to grasp it. Anything less than that and all you'll only see is flapping lights with no musical correlation.

Regards,

[PatrikT]

Originally Posted by Edward_Vinatea

One of the main and immediate advantages would be creating a perfect bass at the sub sonic level and never have to jump into your car to play it on the stereo system.

Say again?

[Edward_Vinatea]

Originally Posted by PatrikT

Say again?

If you meant that you want me to copy an paste the quoted comments - I have no problem with that. But many on GS know that I prefer real audio examples as opposed to engaging in lengthy useless discussions that clog this or other threads. So unless you have a different point to make I'll suggest you upload the best piece of music you feel you've ever mastered in your life {Something challenging like dance or hip hop}. If the bottom end needs to be rolled off - I will and I will also upload it back for your and everyone's review. If it's perfect I will confirm it for you an give you a screenshot of the audio spectrum where you'll see it for yourself. Up to you. I am just trying to help you 'understand' best as I can what I am saying.

Best regards,

[PatrikT]

Originally Posted by Edward_Vinatea

If the bottom end needs to be rolled off - I will and I will also upload it back for your and everyone's review.

What if I hate how everything else turns out with that roll off?


Regards
Patrik

[Edward_Vinatea]

Originally Posted by PatrikT

What if I hate how everything else turns out with that roll off?


Regards
Patrik

Well Patrik - I think that you could state your dislike as much as you want but if other users say that my version was an improvement then the honorable and manly thing to do would be to admit that you are wrong. Don't you think? Again, it's up to you if you want to test this out - the only thing that can happen {if what you did is on point} is to see what it looks like in the spectrum analysis.

Regards,

[Edward_Vinatea]

Originally Posted by studioman22

I am sure you have seen the triangle or "bump" that happens almost always centered around 2.5 to 3 khz. Kinda fun trying to reproduce that without the guitars sounding too harsh, eh? You know what I am talking about? I'd be curious to hear your approaches to dealing with that.

It's fun to talk to someone who can relate with all this without being suggestive or sarcastic - so thanks -- In response to your question it really varies from RTA to RTA display and while it appears to be always the case for the kind of mixes you described - you may not need to reproduce that as it may be undesirable in others {Yes this is when you need to use your ears} But then again - that bulge maybe the result of other harmonics in the mix and be perfectly fine. There is also the issue of context. Is a bit of harshness needed for the song? Only the context in which you are mastering to can give you the answer. Some mixes present excessive density between 2.5 and 5k and you guessed it - it usually always correlates to elec. guitars- that said you need to pay attention at what the vocals are doing and at the elec. guitars' position in the stereo field. You could easily reduce excessive electric guitars level by working the sides with a mb comp or a parametric eq (whichever sounds best for the application}. All of this is easily done without a RTA though - and there is no way you couldn't hear a badly mixed rock song. Usually the drums are badly mixed , overheads loud and too sharp, the guitars are too loud or way too harsh (usually around 2.5k yes), same thing with vocals being too loud or too low in the mix, the bass guitar is buried or overpowering the bass drum, etc, etc. Yeah - I can discern the sound of garbage but I don't need a RTA for that. Thanks for the chat --- thumbsup

I wouldn't want to upset the ultraslutters too much with RTA talk. (Actually, I'm joking, because really I could give a flying ****)

Besides, they would not care to hear your answer, but I would. While not 20,000, I have analyzed probably over 100 hits, mostly rock, some hip hop as well.

Don't worry you'll get there just do it always. -


Best regards,

[Online-Mastering]

My personal view: Yes, an RTA IS definitely useful. And: To talk about dance tracks again, it can be VERY helpful to judge the bass with ONE view. As long as your ears and your brain have the experience to interpret what's going on there. A good bass drum, that works in the clubs, usually 'rolls' down like a 'waterfall' starting with their 'main amount‘ of bass energy (I am talking just about that, not the transients / Attack phase, that surely can cover the whole frequency spectrum) in a specific area (e.g 125 Hz to 200 Hz and USUALLY rolls out to their deepest low freq content without 'bumps' in between this spectrum. And yes: You CAN actually see it. It's one view that gives you a good indication WHERE to look into... Your ears and your experience than have to do the rest (hearing + processing).

The quality of this visualization depends on the quality of the RTA, for sure. A non-precise RTA will SHOW NON-PRECISE results and is thusfor not useful at all. This quality ALSO depends on HOW the data is generated:

We -just for example- use the RME Digicheck analyser in our studios, which comes with the RME products. It has one major advantage compared to many RTAs out there. It works with 31 bandpass filters. Every independent band displays with the SAME RESOLUTION. That makes judgements in the low end much more precise than with many other software tools out there.

Because: Everyone, who has an mathematical idea of Fast Fourier Tansformation (FFT), where digital software rta's are usually based on, knows, that an FFT transformed signal has LOWER resolution in the lows getting to an extraordinary band precision / amount of bands in the highs. This is the reason, why many RTAs just bounce around quite 'nice' in the lows, but you can actually still not 'judge' anything on them.

Finally, to bring peace between you guys Who likes / loves using RTAs should use them. For good reason. Who don't likes them, should NOT use them. For good reason as well. Everyone has it's own methods, and the only thing that counts: how it sounds!!!

Because there is ONE THING that does not have to be discussed: We are all sure, that the most precise and detailed instrument for acoustic perception will always be our pair of ears

P.S.: Too bad I can not do a role reversal reversal with my dog for one day. Maybe his ears would make him the better mastering engineer, who knows

Arne

[macc]

Originally Posted by Edward_Vinatea

many on GS know that I prefer real audio examples as opposed to engaging in lengthy useless discussions that clog this or other threads.

[Online-Mastering]

The shape of the bottom end Online-Mastering describes just like the 'triangle' Studioman22 witnesses is not a constant but a variable and is also governed by many factors including the instrument's fundamentals, the instrument's sub-harmonics, the amount of instruments, their individual level and their behavior over time.

Exactly. So this is what I meant. Learn to see and hear the correlation / relation between certain parameters and their RTA displayed behaviour. This could be helpful for one, while others prefers NEVER to watch any of these tools. Which doensn't mean, as said, they are not useful... It's more a matter of the personal workflow...

Arne

[lucey]

Why would you want to put part of your brain power on looking while listening? It's not necessary to use a RTA, and it has so many negatives and misinterpretations ... so what's the point?

Seriously I'm asking .. what can you not do faster and better by listening?

[Edward_Vinatea]

Originally Posted by lucey

Why would you want to put part of your brain power on looking while listening?

IMO/E part of your brain power is not even active. I am no brain specialist but if we only use 10% or so of our brain capacity then doing both things simultaneously shouldn't be of much effort for our brains.

It's not necessary to use a RTA, and it has so many negatives and misinterpretations ... so what's the point?
Seriously I'm asking .. what can you not do faster and better by listening?

The point is it gives those less talented engineers 'unlike' you a weapon to improve or deliver a decent product. Whatever advantage one gets is not something that is given without extensive curiosity, discipline and sacrifice. You need to train your ears to what's being revealed to you visually. You obviously don't understand why some people can't move a knob without one but DO why some can't give a rat's ass about their utility. It is really up to you if you are curious and committed enough to making your work a little more precise and cohesive than the next ME. I know some immediately are going to post a reply to say "I can do all that without one!!". I say GOOD FOR YOU!!, not everyone is so darn talented like you. You are the stuff others only dream and write about, you are the GOLDEN EARS mastering guy so congratulations. Unless you have something positive to say or contribute to the experience of working with RTAs {I know I could write volumes} please go somewhere else to rant about other grievances like "why use ITB for mastering?" or whatever - there are plenty of silly threads to make your points - skip this one.

Have a good one Lucey,

[lucey]

Talent has nothing to do with this ... it's a matter of practice. We train ourseves to be the analyzer. The subtleties are more than any meter can see, and yet the non-subtleties a RTA shows are what we need to learn about the room anyway.

I just don't see any advantage in looking for anyone, as you are not relying on yourself and not increasing your experience of listening while looking. What are the advantages, long term?

[Edward_Vinatea]

Originally Posted by lucey

Talent has nothing to do with this ... it's a matter of practice. We train ourseves to be the analyzer. The subtleties are more than any meter can see, and yet the non-subtleties a RTA shows are what we need to learn about the room anyway.

I just don't see any advantage in looking for anyone, as you are not relying on yourself and not increasing your experience of listening while looking. What are the advantages, long term?

I won't mind answering endlessly your line of questions. I know that in the end I am not looking like a fool. I will give you just one example why 'talent' or I guess you would prefer the term "experience" then?? Fine! EXPERIENCE it is. Here is your example: An engineer hears a mix and knows something sounds wrong in the highs, thinks is the vocal but has no idea what frequency. The problem: he is reacting to the sound of sibilance but has no idea what frequency it is. Is it 5K? 6K? 8-9-10? No idea. So he kills the problem with eq shelving reducing everything from 5-20kHz . Now He has a mix whose vocals have no major annoying sibilance issues but sounds rather muffled.

Granted this is an extreme example and only a fool would treat the problem this way, but you'd be surprised. So now the less talented -- ERrr "experienced" engineer but armed with RTA looks at the problem and recognizes the sibilant transient at 7.5kHz - gets a narrow parametric eq (or a nb comp} and BAM! Its' gone!! Cha0!! CIAO!! Arrivederchi!! Alright? It's the difference of knowing NOT ONLY that you solved the problem but fixed it exactly where it was.

If you want to continue this useless talk I am going to politely ask of you to take the NB comp test .

Best regards,

[Ben F]

Originally Posted by Edward_Vinatea

If you want to continue this useless talk I am going to politely ask of you to take the NB comp test .

Best regards,

Hi Edward,

I don't think it's entirely fair that you point people to that test. Whist you did prove a point that narrow band compression can be transparent, there are other aspects of the test that we have discussed privately that are not transparent.

Regards,

Ben

[lucey]

Originally Posted by Edward_Vinatea

I don't mind answering endlessly your line of questions. I know that in the end I am not looking like a fool. I will give you just one example why 'talent' or I guess you would prefer the term "experience" then?? Fine! EXPERIENCE it is since you are now analyzing my comments word by word and you seem to suddenly understand RTAs perhaps better than most. Here is your example: An engineer hears a mix and knows something sounds wrong in the highs, thinks is the vocal but has no idea what frequency. The problem: he is reacting to the sound of sibilance but has no idea what frequency it is. Is it 5K? 6K? 8-9-10? No idea. So he kills the problem with eq shelving reducing everything from 5-20kHz . Now He has a mix whose vocals have no major annoying sibilance issues but sounds rather muffled.

Granted this is an extreme example and only and idiot would treat the problem this way, but you'd be surprised. So now the less talented -- ERrr "experienced" engineer but armed with RTA looks at the problem and recognizes the sibilant transient at 7.5kHz gets a narrow parametric eq (or a nb comp} and BAM! Its' gone!! Cha0!! CIAO!! Arrivederchi!! Alright? It's the difference of knowing NOT ONLY that you solved the problem but fixed it exactlty where it was.

If you want to continue this useless talk I am going to politely ask of you to take the NB comp test .

Best regards,

Does one follow up post from me make an "endless" line of questioning? Look I'm not trying to be a smart ass here ... I'm seriously asking you what's the positive for looking at a meter? Your response was not at all convincing. You have sibilance in your example ? ... here's an idea, cut narrow and roll it around in that range, or use a desser and roll it around. Play with the parameters and solve the problem with the least damage and the most gain.

The whole point to making records is that it's as much an art as a science and discovery is where you put your stamp on the work. If we have a clear taste, our stamp has value to some people. If we want cause and effect in nice provable quantities this is not that kind of work. Harmonics matter and overtones may push the meters to peak in places where it's the fundamental or some other tones that's are the issue. Ears, not meters are the way to:

a) learn the room
b) learn to think about problems from multiple directions
c) max out the musicality by listening to the options, quickly and with an intention

And no, our listening is never as sharp while we're looking. Close your eyes and prove it to yourself, every time. Looking without focusing is the way to open the ears. Following your eyes is a great way to trick your ears into what's "right" when really it's not the best solution. Eyes create a mental block and potential conflict to what listening tells us.

I'm not trying to insult anyone, I REALLY don't see the point in looking vs. all there is to gain by listening.

[macc]

24-96 - bollocks to PC! Out with it man!

[Tapeworm]

Originally Posted by Online-Mastering

That's so true. BUT: For anybody who is mixing at home under NOT so good listening conditions and with small (multimedia) speakers or little speaker that roll of at 60 or 75 Hz already, the use of an analyser could indeed be helpful...

This is exactly what I use it for. Though I don't have "multi-media" speakers, I like to make sure that there is nothing strange going on in the 20-45 range which is below what my monitors will tell me. If everything is OK, then away it goes. I could be wrong, I don't know but it works for me.