I recently purchased some equipment for recording audio books and think I may possibly have acted a bit too quickly with regards to the interface. I think that it is possible that the order might still be able to be cancelled, however.
For some reason I was thinking that a “line-in” jack meant that it would play either mono or stereo signals. I’m not sure how I missed it before, but just noticed that the Focurite Solo literature says:
Plug Any Instrument Straight In
Plug straight into a high quality DI that minimises the risk of distortion, or switch to line level to instantly connect your mono analogue synth or outboard gear to the rest of your studio.
I had hoped to have it connected to my computer and to my home stereo system, getting a secondary benefit of playing stereo from my computer through my home stereo system or recording stereo to Audacity from my stereo. That’s not humongous, I guess, but I wish I had worked through things a little bit longer.
But I’m also not sure that perhaps that’s how most of the interfaces work.
Other professional features include high-end AD/DA converters, built-in DIs, phantom power and true analog hardware monitoring of inputs for hassle-free zero-latency tracking, in either mono or stereo.
and
Blackjack also allows you to track with zero latency in either mono or stereo, so you can get the right vibe when recording stereo signals like acoustic guitars and keyboards.
Before trying to cancel the order on the Focusrite Solo & maybe trying for the Mackie Blackjack instead, I’d like to make sure that it could in fact record in stereo from my stereo or a mixer & send output in stereo as well. It sounds like it will do that, but don’t want to make a double mistake.
edit: It looks to me like the Focusrite 2i2 might also record & play in stereo. The back shows a left and a right line out, and the features page picture says “Stereo Recording” in the picture.
I guess I would have to get some adapter cables for either one of them, however, to go from a 1/4" jack to RCA & vice versa. My stereo only has RCA jacks and these interfaces only have XLR & 1/4" jacks it appears to me.
Funny… It was only a few days ago I dissuaded another forum member from the Focusrite “Solo” for that exact reasons you discovered.
Yes It does appear (based on reading the literature) that either the Mackie or the Focusrite 2i2 will do what you want. There are probably 5 or 6 other products at about that same price point that also have pretty much the same features.
That may have been me, but I was thinking that it was just a matter of having fewer channels. I was thinking that a 1/4" jack was in stereo… like a headphone jack. I figured I was no musician & probably wouldn’t need multiple lines in or multiple microphones. The extra $50 was the deciding factor for me given that misjudged scenario.
By the way, the specs on those two are a little confusing to me. Hopefully you wouldn’t mind enlightening me a bit.
For instance, the noise floor is apparently a pretty big deal for audio book recording. But the numbers in the specs seem to be somewhat variable and may not always be apples to apples comparisons?
Noise Characteristics
Equivalent Input Noise (EIN), mic input to USB record (A/D), 150 ohm source impedance, 22 Hz to 22 kHz: 60 dB (max) gain: –124.0 dBu
Equivalent Input Noise (EIN), mic input to USB record (A/D), 40 ohm source impedance, A-weighted: 60 dB (max) gain: –126.0 dBu
Direct Monitor Output Noise Monitor and To Mon levels unity/max: –95.0 dBu, 22 Hz to 22 kHz
USB Record (A/D) Noise Floor/Dynamic Range
(From mic input/min gain, 1 kHz –60 dBFS):
–112.0 dBFS Noise, A-weighted, –101 dBu equivalent mic input
noise at unity gain (11 dBu = 0 dBFS) –110.0 dB Dynamic range, A-weighted, (relative to –2 dBFS/+9 dBu)
USB Playback (D/A) Noise Floor/Dynamic Range (Monitor Output,
Monitor level unity/max, To Mon off/min; 1kHz –60 dBFS):
–97.0 dBu Noise, A-weighted, –107 dBFS equivalent digital noise, 106 dB Dynamic Range (relative to +9 dBu >
Why the different noise levels for different ohms? Is that the “Hi-Z” for instruments vs “Low-Z” for microphones?
What is the “Direct Monitor Noise” measuring? Noise going to monitors (as opposed to regular speakers)? or to Line Outs going to a mixer or amplifier or stereo system? ie. tape monitor out type outputs?
What is the difference between dBu vs dBFS?
edit: I see that the Focusrite 2i2 encodes up to 96 kHz vs 48 max for the Mackie.
The Mackie on the other hand has gain up to 60 dB vs 46 dB for the 2i2.
From the Audacity Wiki on sample rates, it looks like anything over 44.1 kHz is overkill for anything other than very specialized uses, given that human hearing only goes up to 20 kHz (for young people with good hearing) & the 44.1 setting will achieve that.
I’d be very happy to, though I wouldn’t be any sort of “expert witness”. But I suppose the observations, trials and travails of a newbie could still be of some value to others in some ways.
As you may recall from one of the past postings dB’s are always relative to some reference. dBs are a log scale and “10 dB” without any other reference is a ratio implying a 10x increase (or decrease) in power. Because power into a resistive load goes as the voltage squared, 10dB only requires a voltage increase of the square-root of 10 (~3.16), and a factor of 10 increase in voltage is 20dB. Your digital signals are basically digitized voltage so the same 20 dB = 10x relationship applies.
So “dBu” are relative the amount of voltage needed to generate 1 milliwatt in 600 ohms. Yes it is obscure and dates from the early days of telephone engineering. That works out to 0.775 Volts RMS. -124 dBu is 0.488 microVolts RMS (pretty damn small). “dBFS” means dBs reference to “Full Scale” ie the maximum code available in a digital system, that’s the units audacity reports as that’s the only reference that it has. When dB’s are used to refer to actual sound pressure levels (SPL) then the reference is a level that was thought to be the threshold of human hearing. These days SPL dBs are connected to the metric system by the relationship that 1 Pa (the metric unit of SPL) is about 94 dB. When microphones quote a “signal to noise ratio” it is normally reference to 1 Pa.
Unfortunately to relate this to the noise floor requirements for ACX you need more information. In particular you need to know the sensitivity of your microphone so you can compute it’s expected output with your voice and then compute the ratio between that output and the noise floor on the microphone preamp. However, I would not expect that unless you get a particularly insensitive microphone that the noise floor on this interface will be a limiting factor.
The noise changes relative to the source impedance because a significant part of the noise is the thermal noise inherent in any conductor.
I’m guessing that “Direct Monitor Noise” refers to the total noise getting to the Monitor outputs, which would be a combination of the noise in the inputs and the noise from the digital output interface.
Followup: Crunch the numbers for the venerable SM58… Shure datasheet lists the microphone’s sensitivity at -54dBV/Pa – dBV are yet another reference for the dB scale, this one a bit more rational at exactly 1V. 1 dBu = -2.2 dBV. So that -54dBV/Pa is -52 dBu/Pa. Take the difference between that and the -124 dBu spec’d noise floor on the BlackJack and you get 72 dB. Subtract that from 1Pa (94 dB SPL) and you get an equivalent SPL of the noise of 22 dB. Which should be more than good enough for ACX recording.
I also looked at the Audio Technica AT2035 – a relatively inexpensive large-diaphragm that I’ve noticed some on the forum using. It’s sensitivity is listed as -33 dBV/Pa – nearly 20 dB higher signal output than the Shure. But the AT being a condensor mic has a pre-pre-amp built into it so you should expect that. Using this mic you’ll be limited by the noise of the microphone itself which is spec’d at -82 Db relative to 1 Pa or 12 dB SPL (which sounds rather optimistic to me).
Note that you should take all of these specifications with a grain of salt. They are tested in “laboratory conditions” which you can bet the manufacturer has carefully chosen to get the best answers. Also the character of the noise can make a big difference. One would expect that the noise should be “random” sounding like a hiss, and that sort is far less objectionable than any noise with a constant tone (eg hum from the power lines, or the high-pitch whine that Koz likes to refer to as “frying mosquitoes”.)
Interesting. I had always pictured an increase in power as a change in wattage rather than voltage, but then maybe that’s almost the same thing since they’re all relative to each other.
Does that mean that speakers rated at 80 dB at 1 watt would put out 100 dB at 10 watts & 120 dB at 100 watts?
So “dBu” are relative the amount of voltage needed to generate 1 milliwatt in 600 ohms. Yes it is obscure and dates from the early days of telephone engineering. That works out to 0.775 Volts RMS.
Most sound equipment specifications are in Watts RMS, though, aren’t they? Is there a simple formula for conversion to Watts in a typical American household with 110 V electrical systems? I see one site that lists watt = amp × volt but I don’t see how to reconcile/convert that to dBu.
-124 dBu is 0.488 microVolts RMS (pretty damn small). “dBFS” means dBs reference to “Full Scale” ie the maximum code available in a digital system, that’s the units audacity reports as that’s the only reference that it has…
When dB’s are used to refer to actual sound pressure levels (SPL) then the reference is a level that was thought to be the threshold of human hearing.
That was the only definition I had ever heard before coming here. I’m glad you explained that to me, and perhaps others. It’s pretty important to know that there are a couple of different reference points, otherwise it would be like reading about temperature changes in degrees with knowing that they were using a Celsius scale in one discussion & Fahrenheit in another.
These days SPL dBs are connected to the metric system by the relationship that 1 Pa (the metric unit of SPL) is about 93 dB. When microphones quote a “signal to noise ratio” it is normally reference to 1 Pa.
computing
Thanks. I was wondering what Pa stood for, as I had seen it in most of the specs listed for microphones. I did see someone had mentioned a simple formula in the comments/ratings section under some microphones for the “self noise”. He listed 94 as the reference point, which I assume is a rounding up rather than down from which the signal to noise ratio is subtracted. So that a mic with an S/N of 80 dB = self noise of 14 dB. That formula seemed to hold up for any mics I saw that li sted both of those specs. It certainly helped me to compare mics that listed only one or the other in that regard.
Unfortunately to relate this to the noise floor requirements for ACX you need more information. In particular you need to know the sensitivity of your microphone so you can compute it’s expected output with your voice and then compute the ratio between that output and the noise floor on the microphone preamp. However, I would not expect that unless you get a particularly insensitive microphone that the noise floor on this interface will be a limiting factor.
Well, I did finally break down and bought some equipment, which hasn’t been delivered yet, sight unseen. I ended up going way over my original budget of $200, though. I had been given two $100 gift cards by companies I deal with & thought I could buy a relatively inexpensive USB mic for audio book narration, but was dissuaded from trying tshat by some conversations on here. Then I was presented with a little side job of $205 & decided to go hog wild, figuring my very frugal accountant wife shouldn’t get too upset now. But I overshot even that budget a bit, figuring I did have an author who had agreed to let me try and convert one of his books to audio book format on a royalty split basis via ACX, so hopefully I could eke out at least enough to cover the $85 overshoot and appease my better half. Now I’m wrestling with whether or not I should add another $50 to the tab & go for an interface that can output in stereo & record a couple of “line ins” at the same time. The $99 single mic, single line in/out interface should actually probably work just fine for the audio book recordings, but would be more limited for hooking up to my stereo, or even to the Church’s mixer if they’re going to be using two different mics & mixer boards. Such is life, I guess.
But anyway, my mic on order is a RØDE NT1. The specs listed are: Cardioid (35 mV/Pa; 20 - 20,000 Hz) and Max SPL: 132 dB, Self-noise: 4.5 dB(A)
I’m guessing that the sound chain will be pretty quiet and not that big of a factor in trying to meet the ACX standards. After the order, though, I started hearing a lot more outside noises than I had expected. I thought we lived in a fairly quiet neighborhood, and I could just record at night after the kids went to bed with some clothes stacked up around the microphone and maybe a blanket or two draped over the top for a makeshift little sound booth that would isolate the mic pretty well from 3 sides at least.
But we’re less than a mile from an interstate as the crow flies per a check on Google Maps, and a 1/4 mile from another road with stop lights, but a 55 mph speed limit. When I listen at night in what I thought would be quiet time I can hear the rumble of trucks and an occasional plane flying nearby, not to mention a siren in the distance every so often. It doesn’t help that our house is wood frame studs covered by 1/4" to 1/2" blue foam sheets & vinyl siding on the outside & drywall on the inside. That’s probably not the best of sound barriers.
I don’t know if my clothes stacked high idea will work as well as I had hoped. I guess I’ll find out fairly soon.
The noise changes relative to the source impedance because a significant part of the noise is the thermal noise inherent in any conductor.
I guess there are tradeoffs to increasing or reducing that impedence to reduce that thermal noise, or it’s relatively insignificant. Or maybe there are some inherent limits to doing so, otherwise they wouldn’t have the big(?) difference between the higher impedance of guitar inputs vs microphone ones.
I’m guessing that “Direct Monitor Noise” refers to the total noise getting to the Monitor outputs, which would be a combination of the noise in the inputs and the noise from the digital output interface.
[/quote]
Thanks I guess that means that the headphones will pick up a little less noise since they’re monitoring at the analogue level before it is converted to digital.
Is there some formula that would convert mV/Pa to dBu or dBV? The Recording Hack’s page for the SM58 gives these sensitivity(?) specs: Cardioid (1.85 mV/Pa; 50 - 15,000 Hz)
What number(s) would be borderline or the boundary for “good enough for ACX recording”? Or is that a variable based upon how loudly the narrator talks, room background noise &/or other factors?
I also looked at the Audio Technica AT2035 – a relatively inexpensive large-diaphragm that I’ve noticed some on the forum using.
Yeah, I think that’s a newer, upgraded version of their AT 2020 with lower self-noise.
It’s sensitivity is listed as -33 dBV/Pa – nearly 20 dB higher signal output than the Shure. But the AT being a condensor mic has a pre-pre-amp built into it so you should expect that. Using this mic you’ll be limited by the noise of the microphone itself which is spec’d at -82 Db relative to 1 Pa or 12 dB SPL (which sounds rather optimistic to me).
It does seem that comparing dynamics vs condensers is sort of an apples to oranges comparison in that regard. As I understand it, the dynamics have close to zero noise. The SM58 page linked to above posts N/A in the SPL/noise section near the bottom. In that section, though, they usually give a Max SPL and self-noise for the condenser mics, but I think all of the dynamics are listed as N/A. Somewhere else they said that the Max SPL, which I think is defined as the SPL where the mic itself will clip, is so high on most dynamics that it’s virtually irrelevant. I think they said it would be at roughly the level of sticking the mic in the end of a roaring jet engine or at a volcanic eruption, and no one could record at those levels anyway.
But is there a mathematical formula that would bring the dynamics into some sort of equivalence with condenser mics with respect to meeting ACX guidelines, or hopefully exceeding them?
Note that you should take all of these specifications with a grain of salt. They are tested in “laboratory conditions” which you can bet the manufacturer has carefully chosen to get the best answers.
Yeah, but hopefully the laboratory conditions from lab to lab are pretty close to each other. If so, at least the various components can be compared reasonably well against each other. I guess we’d have to add a margin of error with respect to meeting some absolute guidelines, such as those posted by ACX.
I’ve also wondered if those “microphone shootouts” or “USB interface shootouts” are always fair. I could see some manufacturers just pulling some unit off of the shelves to send to the testers, while others might cherry pick a much better than average unit to box up and send to them. Having the testers buy them off of the shelf like Consumer Reports does, would probably be a more accurate way of doing those tests, but who wants to do that. I’m sure the guys running those websites don’t have nearly the budget Consumer Reports does.
Also the character of the noise can make a big difference. One would expect that the noise should be “random” sounding like a hiss, and that sort is far less objectionable than any noise with a constant tone (eg hum from the power lines, or the high-pitch whine that Koz likes to refer to as “frying mosquitoes”.)
Is a constant tone easier to filter out than random noise? It seems to me like perhaps it would be, but I think Koz said that those frying mosquito sounds are all but impossible to filter out without damaging the actual recording.
No, you missed it. 10x power (watts) change is 10 dB. That is the definition. 10x Voltage change into a constant load resistance is 20 dB because a 10x voltage change results in a 100x power change.
There’s two simple electrical rules to memorize. First is “Ohm’s law” which says that V = I * R, where V is the voltage in Volts, I is the current in Amp(pere)s and R is the resistance in Ohms. Second is the power equation you list above P = I * V where P is the power in watts, and I and V are current and voltage as before. From these you can derive the rules P = (V^2)/R and P = I^2*R. The former is why 10x voltage is 20 dB.
The general formulae for dB are 20log(V/Vr) or 10log(P/Pr) where “log” is the logarithm base 10. So for a given voltage Vx, dBu = 20*log(Vx/0.7746).
Things get more complicated when the voltage is not a constant DC but rather a collection of collection of alternating frequencies. Folks start talking about “Impedance” instead of Resistance. That’s because the current is not necessarily in phase with the voltage. The formulas listed above still work but you have to consider the phase of the signals, so we engineers start using complex numbers (remember those from high-school algebra?) to represent the values. It’s probably too complicated a subject for these pages.
Pa is short for Pascal. The 93 was my typo, which I corrected but not before you managed to quote it.
Indeed that looks like a pretty nice mic. I doubt you will be able to get your recording environment anywhere close to 5 dB. Blankets are good for damping echos which is important for avoiding the “talking the shower” effect, but they will do little to suppress extraneous noises. For that the first thing is to close off any doors and windows. There’s a long thread that just resurfaced in the Audiobook forum (because the poster finally finished his first project) working in an apartment near La Brea and Olympic here in LA.
With the windows open at night I occasionally hear a train. The nearest tracks are in the LA harbor ~ 6 miles (and a lot of hills) away as the crow flies. Your exterior walls are probably not that bad. I’m guessing there is also a layer of plywood between the foam and the studs as well. The windows and doors are the bigger issue. Double pane is a lot better than single pane. Airtight weatherstripping helps a lot also. (If you live someplace where it snows then you my have all of that. Here in LA we tend to ignore the weatherstripping because it really doesn’t matter much.)
see my last post. 1 mV = 0.001V in case you hadn’t figured that out.
It depends on how loud you talk and how close you are to the microphone. A reasonable target is probably a background SPL below 25 dB.
A dynamic mic is a completely passive device so it’s noise level is the same as the thermal noise of a resistor of the same resistance. Johnson–Nyquist noise - Wikipedia Your preamp will always have an input noise that is greater than this lower limit, so there’s no need for a specification.
That’s basically what I did in my example. Given the microphone sensitivity and the noise spec for the preamp, you can calculate the equivalent SPL of the noise. If you have a condenser mic you can do the same calculation, but you also have the noise from the pre-pre-amp in the microphone which is nearly always greater.
Yes it is easier to filter out. The problem is when it overlaps the frequency spectrum of the program. For power-line hum if it is mostly 60 & 120 Hz you can notch those frequencies out of most voices without much damage. However if you try to notch out the 1kHz mosquitoes there are odd side-effects because your voice is now missing those frequencies as well.
OK, so for a given resistance watts always = volts squared. That’s pretty simple, even for me, though the equations would be a little easier to memorize if watts was represented by “W” rather than “P” and amps was “A” rathera
It wasn’t intuitive for me, though, since my main exposure to these is via household electricity where we always speak of voltage as being constant within a circuit, either 110v or 220v depending upon how it’s wired in the circuit breaker box, and the wattage is different depending upon what appliance or light bulb is plugged into any given circuit. I guess I did realize that there was some mathematical relationship between watts & amps with amps being the much larger measure. The circuit breakers were all marked in amps and ones marked with 20 could presumably handle twice as many watts flowing through all of the appliances plugged into that circuit as a circuit with a 10 amp breaker.
The general formulae for dB are 20log(V/Vr) or 10log(P/Pr) where “log” is the logarithm base 10. So for a given voltage Vx, dBu = 20*log(Vx/0.7746).
Things get more complicated when the voltage is not a constant DC but rather a collection of collection of alternating frequencies.
Uh oh, by my standards things are already getting pretty complicated when trying to remember variable names in multiple equations involving logarithms. At least it was in base 10, though.
Folks start talking about “Impedance” instead of Resistance. That’s because the current is not necessarily in phase with the voltage. The formulas listed above still work but you have to consider the phase of the signals, so we engineers start using complex numbers (remember those from high-school algebra?) to represent the values. It’s probably too complicated a subject for these pages.
And here I thought we were already dealing in complex numbers, so I googled it. Yeah, I remember touching on imaginary numbers in some math classes many decades ago, but if I recall, I was thinking “What possible use could ‘imaginary numbers’ have in the real world?” It looks like yet another of those times where I was thinking wrongly.
Indeed that looks like a pretty nice mic. I doubt you will be able to get your recording environment anywhere close to 5 dB.
Yeah, I imagine background noise will probably drown out the recording chain noise for the most part. They are additive, though, correct? For instance if a 25 dB SPL noise floor was the limit to meet ACX guidelines and environmental background noise was 21 dB, would that push it to 26 dB total, thereby exceeding the ACX guidelines? Or does that go back to the logarithmic scale again and they’re not really additive?
Blankets are good for damping echos which is important for avoiding the “talking the shower” effect, but they will do little to suppress extraneous noises. For that the first thing is to close off any doors and windows.
I did see one one site that there are two separate issues to deal with… noise isolation, and noise dampening, or absorption. I guess I’ll have to try and figure out a way to try and do both. It’s too bad I spent all my allowance on gear, a little like the 16 year old who spend all of his savings on that hot new car and now doesn’t have enough money for gas, much less licensing, taxes and insurance.
There’s a long thread that just resurfaced in the Audiobook forum (because the poster finally finished his first project) working in an apartment near La Brea and Olympic here in LA.
I bookmarked that thread to go back to. I read a bit and there’s probably a lot I can learn from it, but it is some 39 pages long, so it may take awhile.
Your exterior walls are probably not that bad. I’m guessing there is also a layer of plywood between the foam and the studs as well.
Unfortunately, not. We had a big hailstorm some years ago and we, along with a lot of other neighbors, had to have our siding replaced. No plywood. I think a burglar could get into the homes with a good knife or box cutter by cutting through the vinyl siding and punching through the foam & then the drywall.
I guess it keeps the obese burglars out, though. They would still have to shimmy between the 2X4s.
The windows and doors are the bigger issue. Double pane is a lot better than single pane.
At least we’re ok there. We only have blinds on the windows though, along with some of that metallic reflective material in my home office that I’d like to double as a “recording studio”. Maybe my wife wouldn’t mind if I hung up some sound deadening curtains there. She’s a little fanatical about not putting holes in the walls or ceiling, though.
Airtight weatherstripping helps a lot also. (If you live someplace where it snows then you my have all of that. Here in LA we tend to ignore the weatherstripping because it really doesn’t matter much.)
It can keep out the bugs too, though. Actually, I need to redo the weatherstripping on our front door, and adjust it a bit as well. I think they probably never had it hung perfectly straight or it got a little out of kilter before we bought it. They expand and contract quite a bit from the rather extreme temperature changes here in the Midwest and it can scrape a bit at the top under some conditions and can have little gaps here and there at times. The threshold can be adjusted up or down with a screwdriver as things expand and contract and I’ll probably have to be a bit more diligent in adjusting it as the conditions change.
Well they do at, but not quite as you might expect. If you add two equal noise sources (eg two sources at -10 dBFS) you will get noise that is 3 dB higher in amplitude (so in the example the resulting noise will be at -7 dBFS). If you have the “wave stats” plugin in Audacity you can prove this by example. Generate two tracks of “white” noise with amplitude of 0.545. Wave stats will tell you that each of these tracks has an RMS level of -10 dBFS. Now select Tracks->Mix and Render. The result will be a single track that is the sum of the previous two which wave stats will tell you is at -7 dBFS RMS. This makes sense if you think of the power being added, 3 dB corresponds to doubling the power. So now consider two noise sources where one is 3 dB lower in magnitude than the first, the resulting noise is about 1.8 dB greater than higher source before. For your example where the difference was 4 dB the result is 1.4 dB higher or pretty close to the 26 you stated. If the difference is more than 6 dB you might as well just take the higher number.
But I overshot even that budget a bit, figuring I did have an author who had agreed to let me try and convert one of his books to audio book format on a royalty split basis via ACX, so hopefully I could eke out at least enough to cover the $85 overshoot and appease my better half. Now I’m wrestling with whether or not I should add another $50 to the tab & go for an interface that can output in stereo & record a couple of “line ins” at the same time. The $99 single mic, single line in/out interface should actually probably work just fine for the audio book recordings, but would be more limited for hooking up to my stereo, or even to the Church’s mixer if they’re going to be using two different mics & mixer boards. Such is life, I guess.