Before going any further, I just want to point this out: if you're looking for a sparring partner, you'll have to find someone else. I'm not here for flame-wars.
The crossover point isn't the problem here. I checked the frequency response of my sub and sent a chirp around that range to it and determined (subjectively) where the roll off was. I then sent different tones in that narrowed range to pinpoint the roll off. I estimate it was at about 90 Hz, so that's what I set it to, using a 90 Hz tone to set the crossover, since it's not easy to dial in. THX specifies 80 Hz. Even though my Audioengine A5s can go that low, I felt it was better to let the sub handle what it was built for.
Sounds like you got that bang-on - good job!
You said unity doesn't matter. Then you said it does for dialing in the speakers. It can't be both.
Re-read what I said. Finding exact unity gain on the crossover does not matter here, for three reasons:
- The crossover isn't meant to be your volume control/preamp, so its attenuation control is there primarily to cut down the incoming signal to prevent clipping (as needed).
- There is likely a sensitivity mismatch between the high and low sides for playback, so you're going to be adjusting those dials too (and may need to adjust the master attenuation control further to prevent clipping on the speakers or sub once you've done that).
- You aren't going to be listening at "unity" straight through the system (because it would be unbelievably loud).
You can't absolutely know that 12 o'clock on that dial "is unity gain" (you'd have to measure it to know; input sensitivity and impedance bridge vs the source will also impact this too).
However on the speakers and subwoofer's attenuation controls (the "volume knob" on those devices) you should look for whatever THEIR unity gain setting is, so that they aren't attenuating the signal that they're getting (because we're safely assuming it isn't clipping, thanks to what you've done up-stream on the crossover), because you already have control over that with the knobs on the crossover, and its probably (ergonomically) easier to adjust everything from the crossover, than to crawl behind your desk to get at whatever controls on the sub/speaker while trying to get everything set. Depending on how big the sensitivity mismatch is between the speakers and the subwoofer, you may have to use their onboard controls to help prevent clipping, but I wouldn't outright assume that's needed from the beginning - see what can be done with the controls on the crossover first.
There is no preamp or attenuation. I don't own a preamp.
To quote the great and storied busters of myths: well there's your problem.
I'm using the crossover control volume as a stopgap,
That will potentially work and potentially make more headaches. That control isn't really a "volume control" as much as it's an attenuator there to dial the crossover into the source (which is exactly why the manual tells you to rely on the clipping lamp (that lamp isn't indicating clipping for the speakers or their amps, its measuring clipping for the input)). Sure it can work in a pinch, but its less than ideal, is my point.
because otherwise I have to use Windows to control volume, and this causes a loss in SNR.
Using the soundcard's volume control may cause a slight loss (especially at the extreme lower end), but it is unlikely to be the case at the higher end and/or be audible (even at the extreme lower end), especially relative to all the losses you've already got with the source material, speakers, crossover, wiring, acoustics, etc...
Personally I'm a fan of stand-alone hardware (which, you guessed it, will probably be more of a "bottleneck" than most anything you can do in software) simply because I like hands-on control that's independent of the computer's will/state - if the machine crashes and wants to sound out a very loud signal (and yes, this happens, and no it isn't a "Windows problem" - I've watched OS X, BSD, and Linux boxes do it too) its nice to have a physical knob or button or *something* that you can grab and kill the signal, as opposed to frantically looking for power buttons or similar. Now, grand scheme of things, this is relatively rare scenario, but all it takes is this happening at 3 AM one time to scare you into never sleeping again to want such a setup.
I can afford a DAC/H-amp/preamp now, but I'm probably going to hold out until Christmas for that.
I mean, that'd work, but there's no real huge need for it (the soundcard is already giving you an analog line output and 6.35mm output for headphones). If I remember right, Schiit sells a basic passive volume controller for like $50 - that'd work too. There's also comparable hardware from various studio mfgrs, like Mackie, PreSonus, etc - look for "studio monitor controller" or similar (and if you want to keep the cost down, avoid the ones that are designed to run eleventy million different pairs of speakers, have DACs in them, etc). You might also be able to find a used 2ch component from "back in the day," like the Kenwood KC-105 as an example.
75dB is the rule of thumb, from my research, and 75 gives you peaks of 105dB, so 85dB should be even higher.
I have no idea what you've been reading (because I'm not there with you), but both of those claims are not quite right. However before we go on to that, I want to re-state what I said earlier:
- I am not here to argue.
- I did not say 75 dB was "wrong" or "a problem" - a lot of home theater message boards are full of people preferring to use that as their reference level as opposed to the 85 dB standard. This is done for a variety of reasons, ranging from personal preference, so-called "wife acceptance factor," or because the system is installed somewhere that will not tolerate "loud" noises at any time (e.g. in an apartment), or the amplifier/speakers can't do the 85/105 level, to name a few. But that doesn't change Dolby/THX's reference level specs (if you want to go read about it:
http://www.acousticfrontiers.com/2013314thx-reference-level/ and
http://www.thx.com/blog/questions/what-is-the-reference-level/)
I also think its worth explaining: this value is meant as a "reference level" for systems that can have their reference level set. What this means is the volume control shows values in dB of attenuation, and can have its scale shifted to some user-defined point. Specifically, the idea is that you calibrate the 0.0 dB point on said control to equate to 85 dB SPL at the listening position (which requires the speakers and amplifiers to be up to that task, as well as providing 20 dBSPL of headroom beyond it). It is not, however, saying you will listen with the volume control set at 0.0 dB (and even if you did, it'd require the source material to be constantly at - 20 dBFS to have constant 85 dBSPL - that's unlikely unless the source material is a test tone). I suspect most people do not, in fact, listen with the controls set at 0.0 dB. I'm also somewhat unclear on what you're actually doing with this "reference level" for your system.
To note something else: whatever you set the reference level at does not explicitly dictate what the peaks are (e.g. "75 dB gives you peaks of 105 dB"). In one of the links on THX Reference you see the -20 dB noise and its relationship to help figure out the 85/105 thing, but that assumes your amplifier and speakers can go to infinity and beyond (in real life they can't - they have a limit somewhere, you should size power/playback to whatever your peaks are, and frankly I'd say pad it a bit further than that, because clipping sounds nasty), and that the source material actually has that much dynamic range in it to begin with. For movies and some classical music, 20 dB of dynamic range isn't beyond the realm of possibility, but for a lot of modern music (especially pop and rock) there's a lot less dynamic range (this is primarily due to the "loudness wars") so if you set your volume at a fixed position using -20dBFS noise and shot at 85 dB SPL, you'd probably not be going significantly over 90-ish dBSPL. I also suspect a lot of broadcast/streamed content (including television shows and movies) to have more compression, in-line with more modern music. In your case it isn't entirely clear what you did with 75 dB vs the pink noise. And remember: this is a calibrated reference level, not a "this is how loud you have to listen" level.
Frankly I'd say this is well into the realm of "overthinking things" and probably not worry too much about this - you're unlikely to listen at such loud levels continuously (even 75 dB), and you aren't working from a system that has variable reference levels (sure you could put tape or other marks on the volume knob if you wanted), and you aren't building a home theater that's trying to go for Dolby/THX/etc certification. You just want to listen to music and have it sound good. So I'd say get the crossover point at a comfortable level (sounds like you've done this), get the levels between the speakers and sub close enough (so one isn't overpowering the other - this is "overcoming the sensitivity mismatch" point from above), set whatever balance controls you may need, and be happy. This should realistically take about 5 minutes. A volume control or preamp will make running this system easier too.
85 is the threshold for noise-induced hearing loss (over long, routine periods of exposure).
85 dB SPL is usually listed as the "bottom" on noise exposure charts (e.g. from OSHA and CCOHS), with a rated continuous daily exposure maximum of 8 hours (which is roughly a shift length). More than that (either more time or higher level output) and you're into potential for hearing damage due to noise exposure (but neither OSHA nor CCOHS give you an explicit "after X day of this" value that I'm aware of - just that being exposed to 85 dB for >8 hours a day is a bad thing). But nobody is saying you need to listen this loud (or for that long unabridged), its just a reference calibration point to correlate dBFS to dBSPL in a known way. I would be surprised if many people were actually listening that loud, especially to music (where there isn't much in the way of dynamic range (up
or down), so basically whatever the reference level is set to, is probably going to be the continuous output level too). I absolutely agree with "turn that ---- down" whenever possible and personally think 75 dB is probably on the "loud" side for regular listening. I'm not saying "crank things up to 85 dB and leave it there" - just that that's a reference standard for home theater. 75 dB seems to be what a lot of people in the community like for a reference point, which is just fine too (as I said), but even there I can't imagine running at that level continuously. If you want to see this "in practice" setup your TV or radio or whatever you usually listen to at a level you find comfortable, and then use your SPL meter to see what's that like. Maybe you do listen loud, in which case you might want to change that behavior, but I'm guessing that like most folks, you don't want a real-life rock concert in your living room every night...
Cheers.