Does low frequency content lower your headroom? Why?

[barefoot]

Re: should i cut everything below 40hz?

if the lows below 35hz create a lot of distortion, should i cut it?

I hate to say it, but probably the best answer is Yes.

Brief transients of very deep bass even below 20Hz are present in a lot of material, particularly drums. Although it's fleeting, when accurately reproduced this very low frequency material can add an amazing lifelike and visceral quality to the music. Unfortunately very few speaker systems can come close to reproducing any, let alone any accurate information in this range. This frequency content does indeed usually just create distortion.

I would use a high quality low Q high-pass filter with a fairly shallow slope, no more than 12dB/octave, set at about 30Hz.

barefoot

http://barefootsound.com

 

[jugalo180]

okay barefoot

i know this may be a redundant topic for you, and i do appreciate your help and interest. so if i'm mixing down synth instruments or sampled instruments, vocals, or just plain any track it would be okay to include that low end roll off as a permanent step in my mixing? meaning i would benifit more by always including it as a step then to not right?

 

[barefoot]

Well, if you're willing to go the extra mile and add a low pass filter to individual tracks, then I would move it around and tailor it to the track. I personally try to use different corner frequencies and slopes in order to spread the "damage" around. For example, I might use 12dB filters with a cutoff frequency of 25Hz for the drums and 32 Hz for the bass. For vocals, lead synth, or guitar I might use a 6 dB filter set somewhere between 60Hz to 160Hz. It might be just my superstition, but by spreading things around I feel like there is less of a signature than putting one filter over the entire mix.

barefoot

 

[jugalo180]

okay got it

so you have preset filters that you add to each track individually instead of to the mix in it's entirety. if i'm understanding you right that's what i have been doing in my tracking software i have no presets that i'm comfortable with in my plugins yet, not until i learn them very well. i have to admit it's very tedious, but having low line equipment i figure that i would have to work harder to sound fairly descent, so i've been hanging in there. it just get's hard when i duplicate tracks and end up with 15-20 tracks that need to be mixed.

thanks barefoot

 

[barefoot]

Yeah, but I don't necessarily filter everything, just the tracks that have unneeded low frequency content. I try to work out those problems on individual tracks rather than the entire mix and I avoid using the same filter settings twice.

barefoot

 

[jugalo180]

originally posted by barefoot
I feel like there is less of a signature than putting one filter over the entire mix.

originally posted by barefoot
I avoid using the same filter settings twice.

so if i'm understanding correctly you use different settings so that the entire cd doesn't have the same feel to it? every instrument is recorded with different variables so they will require different adjustments? hmm, that's a really smart thing to do.

 

[kormaniac]

FRONT WINDSHIELDS!!! wow!!

Sheeesh c9-2001....I hope you were hiding in a bunker as you tested this dynamic weapon!!!
As for breaking the front windshield....I'm guessing that it broke from the actual car frame distorting under the low frequency loading....? Actually no...I take that back...its probably just the surface area of the windshield.... I'm guessing the windshield doesn't have a resonant mode anywhere near that frequency.

I'm just curious as to what that might do to your body.

 

[littledog]

so if i'm understanding correctly you use different settings so that the entire cd doesn't have the same feel to it? every instrument is recorded with different variables so they will require different adjustments? hmm, that's a really smart thing to do.

barefoot's not talking about different CD tracks as in different songs. He's talking about different tracks on a multi-track recording, like vox, guitar, kick, snare, bass, overheads, etc.

the idea is not to create contrast from song to song (you do that in other ways) but to make the filtering process "invisible" by using different slopes and cut-offs on different instruments. Perhaps making it less likely that someone could say "Oh, I can hear he used a 40Hz 12dB/octave filter on this song".

Although even Barefoot would admit that most of us probably wouldn't be able to tell either way.

 

[barefoot]

Although even Barefoot would admit that most of us probably wouldn't be able to tell either way.

Yeah, it's definitely not anything that even some self-proclaimed golden ear could point to. It's just one of those many little pinches of spice to throw in the mix and make it ever so slightly tastier.

barefoot

http://barefootsound.com

 

[jugalo180]

okay

got it

 

[ecs113]

Sonusman -I guess- where he said about separating into his and lows and compressing one or the other.

I think the general practice is not necessarily high vs. low, but in general to use an EQ in the sidechain to compress certain frequency ranges. If you are having a prob with bass eating up headroom, cut the higher frequencies on the EQ and leave the lower frequencies at 0dB. Only the frequencies passed thru the EQ in the sidechain of the compressor will be compressed. Vice versa, if you are having problems with sibilance, compress the offending high frequencies. But even these are only specific examples; you can compress any "shape" of frequencies you can dial into your EQ.

 

[Seanmorse79]

damn, I love this thread. I'm learning a lot. My guess, is that Curdura was interested more in application of a multi-band compression plugin, rather than EQ side-chaining a compressor (not that the results would be any different, and not that I'd challenge anything I've read here.)

barefoot, I didn't know you were a speaker designer. (LD let the cat out of the bag - now all the little ya-hoo's like me will be bugging you ). I realize you've moved from the topic, but I'm very curious - could you explain how distortion is measured?

 

[ecs113]

My guess, is that Curdura was interested more in application of a multi-band compression plugin, rather than EQ side-chaining a compressor (not that the results would be any different, and not that I'd challenge anything I've read here.)

Aaahhhh - multi band compression. I thought he was talking about chaining multiple compressors... lol

Yeah, the theory would be the same, compress the offending frequencies...

 

[Flo' Dolo]

off-topic, but still a kissing cousin (Ewwww!!!!)

I often write/mix stuff with lots of low-end stuff going on from hip-hop to breakbeats to electronica (like yellow magic orchestra and other early electronic artists - i dig listening to sounds made with a room full of arps and stuff), etc. I often have to mix a (what I consider, at least) to be really low levels to avoid distortion. I use Mackie 624s, and am seriously considering adding a subwoofer (this means I NEED a sub). Of course, I'm aware or the mackie sub, though I've always been told to check out M&K and you mentioned Velodyne (I never heard of them).

A lot of this music is geared towards clubs (and the subwoofer set), so the rumble factor is waaaayyy important. Feeling it is as important as hearing it. I can't just get rid off the freq. that other systems can't reproduce, sometimes it's the queasiness that comes from that sound that puts thetrack over the edge in a club/live situation... Do the same suggestions re: eq & compression still apply in this case?

<edited this part out, need to try some things, Thomas...>

Oh, and what do you think about incorporating a sub? I chose not to get the 824s thinking the 624s w/ a sub would be quite cool. What should I look for? What exactly am I listening for?

What about a barefootsound sub?

Thanks for taking the time to read this. Any guidance would be hella appreciated...

Flo' Dolo

 

[barefoot]

Sean, Flo'Dolo,

Sorry for the long delay in my response. I've been using up my vacation days that I never got to take in 2002. Now, unfortunately, I'm back to the grind.

Distortion Measurement

There are various techniques and a lot of specific issues but in its most basic form you drive the speaker with a sine wave and record the output with a high precision microphone. When you put that recorded signal through a spectrum analyzer you get a large peak at the frequency of the input sine wave and several smaller peaks resulting from distortion induced by the speaker. The heights of these extra peaks are measured relative to the height of the main peak and given as percentages. This process is done for many different sine wave frequencies so the percentage distortion values can be plotted a function of frequency. Often you will see the primary 2nd and 3rd harmonic distortion values plotted. Otherwise all the peaks are summed together and the total distortion as a function of frequency is plotted.

Intermodulation distortion is measured similarly except two sine waves with closely separated frequencies are used instead of a single sine wave. The harmonics of each of these sine waves are ignored while the sum and difference frequency peaks are measured to determine the percentage of intermodulation distortion.

I've attached a distortion plot for probably the best 6.5" midbass woofer available today. You can see THD is around 0.3% in the midrange. As you go down in frequency the distortion starts shooting up dramatically. It's about 5% at 60Hz. Now remember that this distortion is the harmonics of 60Hz. In other words, false signals at 120Hz, 180Hz and higher which are clearly audible at -26dB. And also remember that this is for the very best 6.5" midbass driver available. This driver has much better performance than the midbass drivers in a Mackie HR624 or even a Genelec 1030A, so you can imagine how bad their curves look.

Deep Live Bass

For live mixes you probably do want to weight the mix more heavily towards the low end. But despite the fact that you may have large subs, pro sound systems often do not reach much lower than 40Hz. For live situations personally I remix my music specifically for the system I will be playing through. The place I perform most often is my friend's 4000s sqft graphic design studio which doubles as an unofficial dance club about once a month. There I built them a custom sound system with subs that begin rolling off shallowly at 32Hz. Their closed boxes and shallow slopes allow them to still produce significant amounts of well controlled energy down to 20Hz. So when I'm performing there I don't use any hi-pass filtering on my mixes. When I occasionally play through other systems that typically roll off at 40 Hz using ported speakers it's important to cut out the very low content. Ported speakers are extremely uncontrolled below their lower cutoff frequencies and basically require hi-pass filtering. So check out the specs on the subs you're using and roll off the lows appropriately.

Subs

Subs are a great idea. Not only do they give you extra bass extension, but they also free the monitors from all that deep bass allowing the upper frequencies to run more linearly. Yeah, you should definitely go with 624s plus subs over the 824s. The 624s have better midrange anyway.

I do build subs. The design I most typically build is an 18-inch super linear, extremely long throw driver in a rock solid 4 cubic foot closed box. Like I mentioned above, the lower f3 is 32Hz and the shallow roll off helps produce very accurate very low bass. I always recommend a pair for stereo. I hand build the electronic crossover myself, tailor the curves to the particular monitors, and can even incorporate active bass compensation circuitry to bring the f3 even lower.

In order reproduce very low bass you need to move a lot of air. The laws of physics dictate that in order to this you need a large speaker in a large box. The only good way around this is to use a smaller speaker with an extremely long cone travel in a smaller box and active feedback control like the F and HGS series subs made by Velodyne http://www.velodyne.com/.

barefoot

 

[DeadPoet]

I just found this very interesting thread... Barefoot, you mean what's happening below 60Hz isn't reliable on a Gen1031 too or is there a big diff with the 1030s ? (not that my room is up to par with the speakers, and I'm just starting to learn working with the 1031s)

I should stop reading these kind of posts, they're depressing..


Herwig

 

[bleyrad]

i did some informal (ie with an ECM8000 in my crappy room) distortion analysis of my YSM1p's, the results were quite interesting. there was very little distortion above 300Hz, but apparently at around 110Hz - where it was the worst - the distortion was only 10db quieter than the fudamental. it gets better when you turn down the overall volume of the speakers.

i find that odd, because i always regarded these speakers to have a really tight, clear low-end. in music, i don't notice it at all. but when i did sine wave tests, they had to be super quiet when i was testing below 300Hz in order not to hear any distortion.



also, i'm curious: why do you recommend 2 subs for stereo sound?

 

[barefoot]

Barefoot, you mean what's happening below 60Hz isn't reliable on a Gen1031 too or is there a big diff with the 1030s ?

Luckily there are some mitigating factors. Ported two way speakers like the 1031s have an advantage that cone excursion is limited in the lower cutoff region. As more and more energy is radiated from the port, the woofer cone moves less and less. So the rise in distortion isn't really so dramatic above the tuned port frequency as the raw driver data might indicate. Below the port frequency, however, all hell breaks loose.

Assuming the same input power, I would guess the low frequency part of the raw distortion graph for the 1031 8" driver looks similar to the attached graph. This 6.5" Seas Excel driver is higher performance, but the larger 8" Genelec driver doesn't need to work as hard to move the same amount of air so it probably all equals out. The 1031's midrange would certainly not look this clean though - especially the 3rd harmonic distortion curve.

also, i'm curious: why do you recommend 2 subs for stereo sound?

One, because I think all the drivers, including the subs, should be closely time aligned. In my installations I place the subs as close as possible to the monitors. This is the best way to get the crossover right. A mono sub placed equidistant between the monitors can suffice, but it's a compromise. Secondly, because producing very deep accurate bass is a big job. I think a pair of 12" subs is absolutely minimal for a small to average size control room.

barefoot

http://barefootsound.com

 

[DeadPoet]

Luckily there are some mitigating factors. Ported two way speakers like the 1031s have an advantage that cone excursion is limited in the lower cutoff region. As more and more energy is radiated from the port, the woofer cone moves less and less. So the rise in distortion isn't really so dramatic above the tuned port frequency as the raw driver data might indicate. Below the port frequency, however, all hell breaks loose.

I'm guessing the port tuning is at about 50Hz, at least, that's what I find out of their spec.sheet . The harmonic distortion at 90dB SPL @1m on axis would be < 1% between 50...100Hz.

Hmmm.. reading this spec is interesting They actually have a filter 18dB/octave going from 45Hz down, so they know the speaker isn't performing too good below 50Hz because of the harm.dist.... sorry for the rambling, I'm thinking and concluding out loud here

The 1031's midrange would certainly not look this clean though - especially the 3rd harmonic distortion curve.

Every answer includes a new question... attached is the freq.response... the mountain at about 520Hz would probably be what you're saying? Before that mark the response looks clean, after that, it looks like a Belgian higway , but I guess it's all in perspective, these are dips and bumps I don't hear, right ?


thanks,
Herwig

 

[barefoot]

... the mountain at about 520Hz would probably be what you're saying? Before that mark the response looks clean, after that, it looks like a Belgian higway , but I guess it's all in perspective, these are dips and bumps I don't hear, right ?

No, you definitely hear those bumps. They make up the characteristics that still clearly differentiate the sound of different speakers all having relatively flat reponses. The bumps result from a variety of factors including dispersion (directivity) variations, distortion and resonances. There may also be some measurement artifacts in there as well, especially close to 20kHz.

The change at 500 Hz might have to do with better controlled dispersion and reduced woofer cone resonances below that frequency, but it's hard to tell. Convoluted in this graph is also likely a difference in measurement techniques between the high and low frequency regions. Usually frequencies above 200Hz are measure in the normal way you would expect, with a test microphone placed 1 meter in front of the speaker. Low frequencies are more difficult because you would need an extremely large and extremely expensive anechoic chamber to avoid room effects in this region. So typically the measurements are done in the true near field with the microphone placed about 1 cm or less from the cone. The difference in amplitude is corrected for and this near field plot is spliced to the normal high frequency plot. Finally there's the likelihood with an Audio Precision measurement that there are more data points per octave in the high frequencies compared to the low. So there are lots of things to consider when trying to interpret even a seemingly simple frequency response plot.

Thomas

http://barefootsound.com