Frequency Response: Which product is better?

[DT Chris]

The Frequency Response specification attempts to describe the range of frequencies or musical tones a speaker can reproduce, measured in Hertz. The range of human hearing is generally regarded as being from 20Hz, very low bass tones, through 20kHz (20,000Hz), the very highest treble. Presumably a speaker that could reproduce that range would sound lifelike. - from Understanding Speaker Frequency Response

I see Frequency Response ratings on everything from mics, to headphones and speakers, and many other things.

Shure SM57; 40 Hz to 15000 Hz
Shure SM58; 50 Hz to 15000 Hz
AT3035 Cardioid Condensor; 20 Hz to 20,000 Hz (pretty much lifelike reproduction)

My first observation is that the dynamic mics obviously don't have the same capability and range as the cardiod condensor. In fact the AT3035 is impressive, since it's frequency response pretty well matches the requirement of capturing 'realistic' sound.

Now what about speakers and monitor systems? Do the numbers still apply to the sound quality played monitored through the speakers?

Edirol MA-15D Digital Stereo Monitors; 70 Hz to 20,000 Hz (+?/-?dB)
This tells me these speakers will not reproduce the bottom end very well (they're missing 50Hz of range)

My Surround Sound System: RCA RT2250; 40 Hz to 20000 Hz (+3/-3dB)
Looks like a pretty good range to me, for monitoring/mixing purposes.

Logitech X-530(my computer surround system); 40hz to 20000 Hz (+?/-?dB)

Both of these surround systems have better fidelity than many of the so-called 'studio monitors'... so how does anyone know what speakers will reproduce the sound best, so I can accurately mix it?

Also when shopping around, how can I tell the quality of a speaker when reading the specs? I've already established that the range of 20 Hz to 20khz would reproduce 'realistic' sound. So am I right to apply this reasoning to determine which product may be better than another?

Can the same principle be applied to headphones as well?
If headphone A has a 7Hz to 7000 Hz, and headphone B has a 15hz to 17000 Hz, doesn't that mean headphone B will be superior in reproducing the sound played through them?

Frequency Response In Context: A big improvement would be a frequency response number that also includes the amplitude tolerance, expressed as "XHz-YkHz +/- 3dB." This tells you that the amplitude of the speaker's response relative to frequency does not deviate more than 3 Decibels from the center line. The "plus or minus 3dB" spec is regarded as a standard of sorts. The theory is that 3dB differences are "just perceptible," so a speaker whose response curve lies within that tolerance window is a reasonably accurate speaker. - from Understanding Speaker Frequency Response

So not only should I look at the Frequency Response Range, but also how much the speaker/mic wavers at certain frequencies. If they're around +3/-3 dB variance, then they're not too bad at reproducing the sound, correct?

My RCA Speaker system may be good enough for my purposes, though I don't have a graph to see the huge variance that may exist..

Just when I thought I had figured it out, and how to read the graphs....

Now that you know the importance (and limitations) of amplitude variations in frequency response graphs, you might ask: "does the frequency range tell us anything at all?" Yes, it does. As long as you know the amplitude tolerance (+/- 3dB), the frequency response range or width tells you how high or low the speaker goes. A speaker rated as 20Hz - 25kHz +/- 3dB will play lower bass and higher treble sounds than a speaker that measures 40Hz - 20kHz +/- 3dB. I wouldn't bet money that it would be the better, more enjoyable speaker, but at least I'd know something of value.

And now that you know how to interpret these numbers, you're ready to run right out and buy a speaker just by looking at the response curve, right? I wouldn't recommend it. Despite many advances in technology over the past 20 years, frequency response measurement is an imperfect science. The same speaker measured by two different labs may yield different response graphs. And some companies just plain cheat when they publish response curves. If it looks hand drawn, it probably was. - from Understanding Speaker Frequency Response



It's easy to see now which mic would be superior in reproducing the range of the human voice.

The SM58 has always seemed a little muddy to me, and lacking some lustre on the higher notes I would sing, now that I see the graph and somewhat understand it, I know why I feel that the mic is quite 'nasal' in nature. The reproduction ability drops off at the higher frequencies and is not reproduced as well as would be expected. Great for lower and mid rock vocals though.


In the studio, the engineer wanted me to use a Beta58 on a song, and I was like "that's a girlie mic!". It was on a rock ballad type song, where I went through quite the bit of ranges soft to wailing stuff. Now that I see it's range and graph, I can understand why it would probably good for some of my vocal stylings that would reproduce the whole range I was using... It lacks the bottom end though.. Before today I didn't know why. Now I do. Interesting stuff...

So it's true. Brandon says use different mics for different sounds. I'll keep that in mind when I start recording more vocals. The AT3035 or similar mics look very good and would be the best for my vocal range. Can't wait to be able to buy one and hear the difference

Anyone else got thoughts on this matter? Am I missing something?

[brandondrury]

My first observation is that the dynamic mics obviously don't have the same capability and range as the cardiod condensor.

This can be just as much of a blessing as it can be a hinderance. It's up to you to determine if a song benefits using X mic on X source. The science has almost nothing to do with a mic's effect on the song.

it's frequency response pretty well matches the requirement of capturing 'realistic' sound.

Who ever said recording music had ANYTHING to do with being realistic? You can watch a documentary on cops or you can watch the new Die Hard movie. Let's assume that an exciting song is like an exciting Die Hard movie (I didn't say "good", I said "exciting"!!). Obviously, everything in Die Hard is going to be fake. That's the point.

So not only should I look at the Frequency Response Range,

Be VERY VERY CAREFUL HERE! You said "look". The last time I looked at the bottom of the cd, I could see a damn thing.

Now, I can look at a frequency response chart and get some kind of idea about how that is going to sound. If you are a little newer or a little less experienced in this arena, you should ONLY go by what you hear. This means you shouldn't buy a mic unless you get a chance to try it out, if possible.

The SM58 has always seemed a little muddy to me, and lacking some lustre on the higher notes I would sing, now that I see the graph and somewhat understand it, I know why I feel that the mic is quite 'nasal' in nature. The reproduction ability drops off at the higher frequencies and is not reproduced as well as would be expected. Great for lower and mid rock vocals though.

Record a track with any mic, and boost 16Khz as far as the EQ will go with a very narrow Q. What happens? Well, you will hear some crappy high end get quite a bit louder, but that's just it. There are VERY few vocal performances that really need anything that high.

For what it's worth, Wagener coudn't even hear that high. Most 40-50 year old men can not hear that high. http://www.recordingreview.com/artic...-Ringtone.html

Many, many U2 recordings have been done with a 58 on vocals. Do they sound muddy? Not at all! Do they sound dull? Not at all! It's all in how you use it.

The most important info in the graph of the SM 58 is in the low end. They are demonstrating the enormous power of the proximity effect. The SM 58, when a person is 4-5" off of it (like they usually are with any condenser mic), is a very aggressive mic (as you can see with the upper midrange peaks). I've never noticed the lack of 18Khz, but I'm guessing I can't hear that high anyway.

As the chart shows, if you are 2 feet away from the mic, the mic is very thin.
So distances is EXTREMELY important when dealing with a cardiod mic with a tight pattern.

When we think of high end, this is usually found more at the 7-10Khz regions. Anything above 10Khz I call "air" and usually don't make a big deal about unless a certain song calls for that sort of thing (which is rare).

[DT Chris]

Cool, thanks for the feedback. I'll experiment with different ranges from the mic, and fiddle with the range on the EQ after to see the effect you are speaking of.

And, note taken, I should listen to speakers and not look at the specs I'm just concerned that I may not have the 'trained' ear to know which speakers sound better (unless they're beside each other).

[brandondrury]

And, note taken, I should listen to speakers and not look at the specs I'm just concerned that I may not have the 'trained' ear to know which speakers sound better (unless they're beside each other).

Well, this is something that we all fight. It just takes practice.

I've found that the only real way to learn this stuff is to do the best you can with the mix at hand and then take the cd out to the car and elsewhere. This is where the different timbre get highlighted in different ways. This is where you really hear the difference between X mics and find out if you made the right decision during your mix.

It's never easy, but over time it does get easier.

The frequency charts are necessary for learning, but they are kind of the cheat sheet in the back of the book kind of thing. You'll really learn from doing sweating it out in the trenches.

Brandon

[liquidair]

Chris,

I think you may be more or less on the right track, but you are missing a few things. A dynamic mic is just a woofer, so it's response is dictated by what can get it moving. A condenser mic is two charged plates that detect slight changes in pressure. So based on design, a condenser should have better low and high end response, because you don't have to get a coil inside a magnet moving. Bass waves are too big to accurately hit a 1" circle, and ultra highs can't move it. However, bass's pressure can be picked up by a 1" circle (the diaphram, I can't seem to figure out how to spell that word today), same with ultra highs.

Looking at any response, you need to look at amplitude as well. Looking at the sm58 you see a 5db presence peak starting at 2kHz on up to about 7.5kHz, then it jumps back up around +5 centered right at 9.5Khz. This is why the 58 is the "standard" live mic for vocals, it's boosting the presence range of vocals. I too have found the 58 to be muddy, but that's due to the lack of high frequency response above 10K. Cutting highs is like boost low mids, that's where the mud lies.

When I look at the AT3035 response, I see a pretty flat response until the highs. Strictly speaking, this mic will be the most accurate of the lot listed.

Analyzing and then knowing what different mics do, and using them in different applications like Brandon does, is a smart use of a mic. It's a built in EQ to the source. A 57 is a standard guitar mic because a). A guitar amp doesn't have much coming out in the extreme lows, and there's very little going on at the extreme highs. b.) the presence boost at 5kHz is a perfect fit for the guitar.

Here's the big problem with just looking at specs. Luckily the shures have pretty smooth response curves, however some mics aren't this nice. Remember that boosting or cutting one frequency has the effect of cutting or boosting another in a different place. So unless you know exactly what those peaks and dips effect, then you may be surprised when you actually hear the mic.

Getting back to accuracy brings us to Monitors. When you see a rating such as the Edirols, from 70-20kHz, and they don't list a +/-db range, you have to be skeptical. This is for two reasons. One, using your RCA system as an example, is that the RCA system will produce much less and higher than it's rated 40- 20k, it's just that above and below this it's not within + or - 3 db. SO when Edirol doesn't give it's tolerance, that's prolly bad bad news. We could have -10db at 70Hz, which means our ears will hear 70Hz half as well as it would at a frequency rated at 0dB. The second thing is, if you have a +10db boost somewhere, then to our ears that is doubled over the frequencies at 0dB. When you then add the entire range on the Edirols, we may have frequencies 4 times louder than others, try mixing on those. Whoosh!

The problem with the Logitech and RCA systems is again two fold, and the specs won't tell you this no matter how much you look. The first is that you are using a subwoofer on both. In an anechoic chamber (where they measure the responses)a sub is fine. But in the real world, where there are walls and standing waves, That sub is not extending the low end +/-3db. It's probable that without acoustic treatment, the low end responses may be just as bad as the Edirols. The other thing is, Studio monitors are designed to be accurate, Stereo speakers, to sound good. There are certain frequencies that sound good and bad to our ears. 50Hz sounds good, 200 sounds bad. There's alot of these, 1Khz is typically bad, 5Khz is typically good. Stereo speakers are designed to slightly cut the bad, boost the good. This will make things sound better to a user, allowing a cheaper less precise speaker to be made. So when your stereo speakers sound better then most monitors, it's due to the sweetener designed into them, precisely NOT what you want in a monitor.

My KRK V8's, which I love and translate extremely well, are 42-20Khz +/-1.5db. That is extremely accurate. I ordered a pair of E8's from them that are 40-32kHz +-1dB.

One other thing about specs is you gotta think about somethings that might not be so obvious. THe standards for generating specs are as follows. The speaker is set up in an anechoic chamber, a mic is placed 1 meter and a 100db signal is sent to the monitor and recorded. What is not so obvious or standardized is what signal is sent. Sometimes its a sweep where only one frequency plays at one time, the other (and best) is pink noise where all frequencies are played at the same time. This can give drastically different results, you get flatter if you sweep. Cheaper companies do this.

The other thing is ports and radiators. A front firing port will get picked up by the mic (and your ears) and thus show up on the specs. Mackies and their copies use rear firing radiators. A mic in a chamber is not gonna pick up this at all, but in your room, it will be a factor. Mackie claims 40-20kHz +/- 1.5dB for their Hr824s...wrong. That's just wrong. The only way to get that response out of a pair of Mackies is to completely stuff your back walls with acoustical absorbers such that nothing can bounce back into the room. That might need 10-12 inches.

You are on the right track and it's good to see you doing some research and thinking about this stuff. I hope that I helped out a little and let me know if you have any questions or need clarification to anything I said here.

[brandondrury]

Thunderous post, dude!!!

I want to add one more thing that you didn't touch on. Transient response. Mics with similar frequency response curves often sound WAY different with impulses.

You also hear this kind of talk with preamps as well. A "fast" mic or a "fast" preamp is usually get at illustrating the detail in transients. To oversimply things, a fast mic will pick up more high end when close micing a snare drum, for example.

Obviously, percussive instruments are going to be effected by this, but vocals and acoustic guitar are too.

Brandon

[DT Chris]

Originally Posted by liquidair

I think you may be more or less on the right track, but you are missing a few things. ...
You are on the right track and it's good to see you doing some research and thinking about this stuff. I hope that I helped out a little and let me know if you have any questions or need clarification to anything I said here.

Awesome, awesome explanation, my friend. Thanks for taking the time to post such an informative post. This is what I like -> detail. You addressed my concerns very well (as you see I took time to make a detailed question with examples). I appreciate the response, I'll be reading it a few times to absorb it all, and research some monitors using this info

[liquidair]

No prob Chris. I have studied and practiced my ass off since I was a kid, and it does me no good to just keep it all in to myself. What a waste that would be. Anywho, I noticed one error I made, when I was talking about stereo speakers. I said (still haven't figured out how to quote yet):

This will make things sound better to a user, allowing a cheaper less precise speaker to be made.

It should read cheaper, better sounding speaker. Precise was the wrong word, there is nothing precise about equalization used to sound better.

[brandondrury]

This is what I like -> detail.

You think you like detail. This is an objective overview and doesn't really apply to creatively finding the best tone for a recording.

Example: If someone asked me if I wanted a "detailed" snare drum, I'd probably say "Sure! Why not?". If they asked me what kind of mic I'd use, it would probably end up being a 57. I'd probably think it was pretty detailed.

I just read a huge article on the Nuemann U47 a few weeks ago. It's basically considered THE vocal mic of the 20th and 21rst centuries. The originals go for more than the price of a Kia car on Ebay. I read that they've been used on more #1 songs than any other vocal microphone...or something like that.

What was the most suprising trait of the Neumann U47's sound?
Lack of detail! Seriously!

Most attribute the MAGIC sound of the U47s for the Neumann to be able to sort of "ignore" weird mouth noise, sibilance, and things of that sort. By ignoring these things, there is a certain thickness to the sound that seams to sit very well in the mix.

I'm not saying detail is a bad thing. I like my share of it on certain acoustic guitar songs, but details is not what it's cracked up to be. Again, listen to any U2 song and tell me if the vocals sound "undetailed". I have had acoustic guitars that were a bit too clicky sounding. Changing the pick is step #1, but sometimes switching the mic to something "less detailed" was the right choice.

After recording for 6 years pretty much full time and having a pretting strong understanding of frequency response and all that crap. I can't listen to a vocal and tell what kind of mic was used (dynamic vs condenser vs ribbon) unless it's an A/B/C test. (I have been able to pick out the upper mid boost of the 58 on a couple of local recordings, however).

So learn about this stuff a lot. I certainly have. However, in the end, you'll find that pretty much all this theory is just that. It has little to do with the reality of making great recordings.

Brandon