I’ve used Audacity for a number of years, but would still consider myself a novice. Probably somewhat analogous to a guy who test drives a car but doesn’t know how to use all of the features, a number of & which he probably doesn’t even know exist.
I’ve thought about trying my hand at narrating audio books off & on for awhile, but more as a little hobby & not something to try and support my family on. Actually, my goal would be to make enough to pay for any equipment that I could then use for podcasting for my little business, narrating some books or other written material in a creative commons sort of way, just to make it available to others, Google Hangouts & just for playing around with. If I could make a little extra money on the side all the better.
Like most people, I’m trying to get the biggest bang for the buck with regards to equipment and don’t want to waste resources on purchasing items that are unnecessary overkill nor skimp so much such that I would soon be wanting to upgrade to something that would actually fit my needs with the original purchase(s) sitting unused.
Sorry for the lengthy background, but it might impact responses.
In looking at ACX file requirements I see that they ask for 16-bit, 44.1kHz files encoded to 192kbps.
I’ve browsed around the Recording Hacks website quite a bit and found a number of relatively inexpensive USB mics that seemed like they might suffice for this type of narration work. Actually, I had started out thinking of getting a cheap usb mic to record meetings since I had recently been “promoted” to secretarial duties for a couple of organizations & can’t write as fast as people speak. But then I started thinking that it might also be used for narration work, so why not kill two birds with one stone, but it looks like the quality needs to be a bit better for narrating audio books, so here I am.
In that audio interface shootout in particular, they were acting like a 16-bit interface was just too noisy for podcasting work and they didn’t really even look at anything other than 24-bit interfaces for that reason. I don’t know for sure, but I would have guessed that podcasting would typically have lower standards than audio book narration.
But ACX was only asking for 16-bit .mp3 files. In fact, they wouldn’t accept 24-bit files.
Does that mean that 24-bit mics or interfaces are overkill and a waste of money with respect to audio book narrations for ACX? Or would those types of interfaces & mics still give a cleaner signal for Audacity to work with and even when downsampled or whatever it’s called there would probably be a lower noise floor for that equipment even in the 16-bit format ACX is asking for?
Similarly, even the better interfaces with 24-bit processing maxed out at 96 kbps. Would that be an issue in converting it to the minimum 192 kbps required by ACX?
Thanks for any help, insights, recommendations, etc.
they were acting like a 16-bit interface was just too noisy for podcasting work
That’s misleading.
24-bit recording tends to be only available on higher quality microphones. Had they forced a 24-bit microphone to record in 16-bit, it would be fine. We have demonstrated several times that you can do perfectly OK at 16-bit, 44100, Mono.
Attached is an ACX-compliant, 44100, 16-bit, Mono clip I recorded with a rock band microphone, simple sound mixer and a MacBook Pro. I used minor level shifting and processing in Audacity to get there from the raw clip.
However, personal recording is not without its problems. The love affair with USB microphones is losing it’s glow because of several shortcomings. The worst one is “frying mosquitoes” noise in the background. This noise is burned into the USB system and it’s rough to remove in post production.
It can prevent you from achieving ACX conformance.
This is a technical problem, but there are certainly physical ones as well. We can’t take out the dog barking in the background. Sorry. Forget Noise Removal. It doesn’t remove noise. All it does is suppress certain background sounds that weren’t all that much of a problem in the first place.
We have posters going into production with the idea that Noise Removal is going to get them “Out of Jail.” It won’t. And these producers are pretty much doomed.
it looks like the quality needs to be a bit better for narrating audio books, so here I am.
If you can achieve graceful ACX compliance, you pretty much got many of the other applications nailed. ACX is simple but not at all easy. There’s no such thing as starting with a terrible recording and “cleaning it up” in Audacity. ACX final test is a human listening to the clip and if it sounds like you’re recording it from Mars through the vacuum of space, you’re dead, even if it does meet technical compliance.
This is where I give you the equipment list, right?
I can tell you how I and a couple of others did it. ACX is recommending a microphone digitizer and preamplifier I’ve never used before. The Blue Icicle. Jury’s out. I just don’t know. I have a cousin of that, the Shure X2U. I don’t use it much because it has low volume and there’s no good way to fix it.
I really need to touch up that picture so it looks less like I dragged it behind my car.
That’s on my list of things to try so I can list my observations.
Do all of your work in exported WAV and save your archive in WAV. Once you export an MP3, you’re stuck with the MP3 quality and it can only go downhill. ACX requires 192 quality up from Audacity default 128 so they can resample it to lower and lower qualities. Your original works need to be perfect, not less damaged.
You can produce in 24 bit or higher if you want, but unless you’re producing “Abbey Road,” I don’t think it gains you anything and it takes up drivespace for no reason.
If you’ve been “reading the mail,” you know that several people have been successful in using stand-alone sound recorders and leaving the computer turned off. Not an awful idea. I’m doing a sound project right now and I’m using a Zoom H4 (older unit) for the whole show. NPR used to do most of their field work on Tascam recorders. Those aren’t dreadful, either.
I know this smacks of ad copy, but I really do have my “recording studio” wrapped in a bandanna in my backpack. Pantone 293c Blue. The bandanna, not the recorder.
Nobody wants to hear this, but I recommend you start using Something and then figure out what the shortcomings are so you can buy The Real Thing later. The universal urge is to buy The Perfect Microphone right at the top and go straight to Fame and Fortune by Recording Audiobooks.
24-bit microphones are only USB mics, correct? From what you say below, it sounds like I should probably forget about those from the get-go. But what about 24-bit interfaces or mixers vs 16-bit models attached to a normal XLR microphone? I thought I had read somewhere that it is best to let Audacity work in 32-bit variable mode because it is so clean, and then let it downsize to 16-bit or whatever mode later. I’m wondering whether there would be less noise if one recorded in 24-bit mode with some interface, did whatever processing would be needed with Audacity if needed, and then export into 16-bit .mp3 format, vs recording it in 16-bit mode and then exporting it after any processing done to the raw recording.
Attached is an ACX-compliant, 44100, 16-bit, Mono clip I recorded with a rock band microphone, simple sound mixer and a MacBook Pro. I used minor level shifting and processing in Audacity to get there from the raw clip.
Cool. That shows that it can be done with an inexpensive microphone. Do you think that mic makes it more difficult to meet their compliance requirements, than one with a lower noise floor perhaps? Have you ever, or would you attempt, recording an audio book with that setup?
As a side note, how do you test a clip like that for compliance? (Sorry, I’m sure that’s answered in multiple places on here.) I imported it into Audacity & even though the recording line looked like it remained at zero, the sound meter was bouncing around, a bit, up above -45 at one point.
However, personal recording is not without its problems. The love affair with USB microphones is losing it’s glow because of several shortcomings. The worst one is “frying mosquitoes” noise in the background. This noise is burned into the USB system and it’s rough to remove in post production.
That was actually difficult for me to hear, though I’m not sure if that’s because of the relatively low quality stereo system or my old ears. I had to turn the volume up beyond half to get a fairly normal sound level for your voice as well. If I recall from googling it, the stereo is only 10W, though. I hear some hum even with the volume down & burning mosquitoes(?) the more I turn it up.
It can prevent you from achieving ACX conformance.
[quote=“JeffB”]
it looks like the quality needs to be a bit better for narrating audio books, so here I am.
[/quote]
If you can achieve graceful ACX compliance, you pretty much got many of the other applications nailed.
ACX is simple but not at all easy. There’s no such thing as starting with a terrible recording and “cleaning it up” in Audacity.
How much of role does the equipment play with respect to the noise floor issue would you say, vs room acoustics & environmental background noise? For instance, some of the mics I saw on Recording Hacks had 20 dB self noise but were relatively reasonably priced, and had a relatively flat frequency response & pretty good reviews. Others were a little more expensive but had self noise listed at 18 dB, 16 dB or 12 dB. More expensive yet, at $229 was the ACX recommended Rode NT1A which had self-noise of 5 dB. Are those differences significant? Would the lower numbers make it much easier to meet the ACX requirements with respect to the noise floor?
ACX final test is a human listening to the clip and if it sounds like you’re recording it from Mars through the vacuum of space, you’re dead, even if it does meet > technical > compliance.
Is that type of problem fairly common? If so, is it typically from poor equipment, operator error, or room acoustics? If that’s caused by being too far from the mic or something I suppose it’s relatively easy to fix, unless you’ve already recorded a 10 hour book that way. ;
This is where I give you the equipment list, right?
Yeah, now we’re getting to the good part.
I can tell you how I and a couple of others did it. ACX is recommending a microphone digitizer and preamplifier I’ve never used before. The Blue Icicle. Jury’s out. I just don’t know. I have a cousin of that, the Shure X2U. I don’t use it much because it has low volume and there’s no good way to fix it.
For what it’s worth, here’s Recording Hack’s USB Audio Interface Shootout and Review that evaluates both the Shure X2U as well as the Blue Icicle. Shure came out the winner for 16-bit interfaces, but the Icicle is cheaper.
I really need to touch up that picture so it looks less like I dragged it behind my car.
Here’s their review of the Shure SM58. Some other articles about it are linked to on the right sidebar. Several of those articles mention it’s ruggedness and durability. Maybe you should leave the picture “as is” and see if you can’t get Shure to sponsor you.
One other question with regards to mics is that ACX & other voice over pages recommend large diaphragm condenser mics. A few sites that evaluate equipment & seem to have sound engineers doing the discussing seem to all favor dynamic mics, especially for beginners. One said something to the effect that he’d much rather have those at the lower end of the spectrum use dynamic mics. I think he said he thought they would do much better with a dynamic range than a condenser in that range. For me that would be in the high range, rather than the lower end, but I suppose the principle would still apply down to what I would consider the beginner level.
It sounds like the biggest advantage is that it picks up far less room noise and makes that end of things a lot easier. I guess the downside is that it may not pick up subtle nuances in the voice as much, though I don’t know if I could tell that very well.
Do all of your work in exported WAV and save your archive in WAV. Once you export an MP3, you’re stuck with the MP3 quality and it can only go downhill. ACX requires 192 quality up from Audacity default 128 so they can resample it to lower and lower qualities. Your > original > works need to be perfect, not less damaged.
I think I misspoke when I wrote: “Similarly, even the better interfaces with 24-bit processing maxed out at 96 kbps. Would that be an issue in converting it to the minimum 192 kbps required by ACX?”
I think those interfaces can do 16-bit to 24-bit with varying kHz, some up to 96 kHz rather than kbps. Sorry about that.
You can produce in 24 bit or higher if you want, but unless you’re producing “Abbey Road,” I don’t think it gains you anything and it takes up drivespace for no reason.
But would it have any effect on the noise floor or improve the quality in any other way once it was converted into 16-bit, 44.1 kHz, 192 kbps format for ACX? Is that format standard for other audio book companies, btw? It seems to me that the additional drive space consumption would only be until you get the final product in .mp3 format. I don’t know if people then get rid of the .WAV files, but I think I would at least export to .FLAC to save space and you could probably downsize to 16-bit at that point if you wanted, no?
If you’ve been “reading the mail,” you know that several people have been successful in using stand-alone sound recorders and leaving the computer turned off. Not an awful idea. I’m doing a sound project right now and I’m using a Zoom H4 (older unit) for the whole show. NPR used to do most of their field work on Tascam recorders. Those aren’t dreadful, either.
I saw that on one thread and was intrigued by it. It is relatively inexpensive and would be much handier for things like recordings at meetings etc. I’m wondering how much of a downside there might be, if any, with respect to getting audio book jobs. I’m sure the way a person narrates is probably far more important than subtle differences in microphones, but it seems to me there might be situations where it could cost someone a job. From what I’ve read it can be pretty competitive out there. I could foresee a case where someone is auditioning a bunch of people and a fairly subtle difference in the sound quality might be the straw that breaks the camel’s back, so to speak… or tips the balance away from you.
I know this smacks of ad copy, but I really do have my “recording studio” wrapped in a bandanna in my backpack. Pantone 293c Blue. The bandanna, not the recorder.
That type of portability sounds very attractive to me. I could foresee situations where I’d like to take it with me and work somewhere else, or interview someone, or record a talk etc.
Nobody wants to hear this, but I recommend you start using > Something > and then figure out what the shortcomings are so you can buy The Real Thing later. The universal urge is to buy The Perfect Microphone right at the top and go straight to Fame and Fortune by Recording Audiobooks.
I’ll stand over here and watch.
Koz
[/quote]
I know where you’re coming from, but I’m a cautious person by nature. I like to try and minimize my mistakes if I can. Learning that USB mics are probably not a good way to go may have saved me from buying something that I would have regretted later. I also might not get a chance to correct the mistake, or make the next one, if my wife gets a little too riled up.
I doubt you would find much advantage in recording at 24-bit, or at sample rates above 44.1 kHz. 16-bit data has an available dynamic range of 96dB. That gives you a pretty health margin over the ~65-70 dB needed to get past the ACX standards. The pros like to use 24-bit because it gives a lot of extra headroom with it’s 145 dB dynamic range. They tend to set the nominal peak level at -10 or even -20 dBFS instead of the -6 or so that most would use recording in 16-bit. The main advantage of going to higher sample rates (48, 96 or even 192 kHz) is that the anti-aliasing filtering that is necessary before the A/D conversion can be a much gentler filter, back in the early days of digital audio this was something of an advantage. These days most converters are fully integrated “delta-sigma” converters which are achieving the same end result.
Dynamic (moving coil) microphones have no noise. They are a coil of wire next to a magnet. Full Stop. Their only noise comes when the producer/announcer has to amplify the sound signal so it will be useful downstream. Raw microphone signals are molecular-level small. And that brings you to the Microphone Preamplifier or MicPre.
A MicPre’s job is to amplify the signal up to 1000 times. Nobody can do that without adding noise, and so they do. That’s the first place you get hiss (ffffffff). The lower hiss the better because you can’t actually take hiss out of a performance. This is why higher end sound mixers all print in big letters "Now Available With Patented Signal Spritzer Preamps!!! They are obliquely alluding to the quiet, high-quality of their MicPres. The performance will never get any better than that, so this is a good place to be good.
Dynamic mics are robust and can have very good quality. Electro-Voice used to show off their microphones by pounding nails with one and have it reach factory specifications when they were done. They are almost impossible to overload. This is the type of microphone to use on a difficult shoot because of high volume or uncontrolled environment…like a Dead concert. The SM-58 rock band microphone is a Dynamic. So is the Electro Voice Broadcast RE-20. Many radio shows have gone through that microphone (attached).
~~
Condenser microphones generally produce slightly better quality sound, but they’re a lot more complicated. They do it in such a way that the sound signal will not go down a cable at all. It is required that they build electronics inside the microphone itself to jam the signal down the cable to the MicPre. Your choices are battery inside the microphone (I own and have used several microphones like this), or you can send battery up from the mixer in the form of Phantom Power. Obviously, your MicPre has to “know” what phantom power is. My little sound mixer knows as does my X2U. That’s what “48 volt Phantom Power” in the ads means.
Most high quality condenser microphones will not work without that. Condenser microphones are the type where you can rate the performance by how much noise it makes before the signal ever leaves the microphone.
I don’t know of a USB microphone that’s not a condenser. Since there is no microphone cable, all the electronics can be jammed inside the microphone case along with the digitizer and USB electronics. My favorite one that I don’t own is the Samson GTrack.
I borrowed it from a guitarist in the next office for the sound tests. You can use it for overdubbing since it has low latency monitoring (headphones) in the base.
You can’t ever get more than one USB cable away from your noisy computer, so you should consider your environment very carefully with those. It’s not like XLR microphones which can be on the next floor from the MicPre. No, I’m not joking.
USB microphones are the ones that occasionally suffer from Frying Mosquitoes for which there is no known convenient cure.
Jason and I selected the audio interfaces for this evaluation based on two criteria:
We required a 24-bit ADC, because in a torture test of preamp gain and noise, no 16-bit device could compete. …
But that’s apparently an extreme situation. The author also said:
it is also notorious for one more thing: low output. This means it needs a lot of gain… and if you’re recording something especially quiet you’ll need a ton of gain. Inexpensive preamps tend to not have have much gain, or they do at the cost of more noise.
It sounds like that might be an issue for someone using a dynamic with very low sensitivity like the Shure SM7b or the Electro-Voice RE20. I imagine both of those are used rarely for audio books, especially by amateurs. I wonder if that advantage of less noise for that necessary high gain needed for those mics goes away if the 24-bit recording is then converted to 16-bit for ACX or other similar companies, assuming they all require submissions in 16-bit formats.
In one of the articles on Recording Hacks they were doing a shootout of $200 condenser podcasting mics, I believe, and the author slipped in the Electro Voice RE-20 for comparison. It sounded markedly better to me in the blind listening test, but then again it should, I suppose, given that it cost more than twice as much as the others and it’s pre-amp undoubtedly was as well.
~~
Condenser microphones generally produce slightly better quality sound, but they’re a lot more complicated. They do it in such a way that the sound signal will not go down a cable at all. > It is required > that they build electronics inside the microphone itself to jam the signal down the cable to the MicPre. Your choices are battery inside the microphone (I own and have used several microphones like this), or you can send battery up from the mixer in the form of Phantom Power. Obviously, your MicPre has to “know” what phantom power is. My little sound mixer knows as does my X2U. That’s what “48 volt Phantom Power” in the ads means.
If someone hit the “Phantom Power” button by accident when a dynamic mic that doesn’t need it is connected, can it damage the mic?
I don’t know of a USB microphone that’s not a condenser. Since there is no microphone cable, all the electronics can be jammed inside the microphone case along with the digitizer and USB electronics.
I thought I had seen one on Recording Hacks & did a search for USB Dynamics and they had 5 of them in their monster database. One USB ribbon mic also showed up in that search.
You can’t ever get more than one USB cable away from your noisy computer> , so you should consider your environment very carefully with those. It’s not like XLR microphones which can be on the next floor from the MicPre. No, I’m not joking.
Good point. Another drawback for the USB mics. I did see somewhere someone recommended putting your computer in another room & the monitor & keyboard in the room where you are recording to get rid of that fan noise. That wouldn’t work for a laptop, I guess, unless someone had a separate monitor, keyboard & mouse.
USB microphones are the ones that occasionally suffer from Frying Mosquitoes for which there is no known convenient cure.
Koz
I wonder if they’ll figure out how to get away from that someday.
Strictly speaking that’s not quite true. There is still the thermal noise of the equivalent resistance of the coil. However that noise will always be 3-5 db lower in level than the best possible pre-amp.
Likewise in a condenser mic, the noise generated by the microphone element itself is insignificant compared to the electronics (usually a j-fet amplifier) use to get the signal to a level that can be shipped down the cable.
According to that website the SM7B has a sensitivity of 1.12 mV/Pa, and the iconic SM-58 is at 1.85 mV/Pa. The difference is 4.4 dB, which while not to be ignored is hardly huge in audio terms. I suspect that as a class the 24-bit capable interfaces do have lower noise preamps, and lower electronics noise, and for that reason their choice of such devices for their microphone tests is reasonable.
If your task is recording a symphony orchestra who’s sound might range from the just 1’st cello playing pianissimo, to all 100+ players, including the horns & kettle drums going full blast, that 145 dB dynamic range of 24-bits is appealing. But for recording a single voice it’s overkill.
Unlikely, at least not at the price level that hobbyists will accept. There once was a big push from the CES to make optical SP/DIF the standard for digital audio between devices. The optical interface for the digital signal goes a long way towards isolating the inevitable noise that comes from computing devices. However, I haven’t seen much “pro-sumer” recording gear going this route. Sound cards with SP/DIF output (to connect to that stereo system with SP/DIF input) are fairly common. But SP/DIF on the capture side is pretty rare.
Using a Raspberry Pi for the record computer is tempting to test (there is a raspian compile of Audacity available). It’s small size and power level should have a lot fewer noise issues.
Changing phantom power with a non-phantom microphone connected can cause damage by having the power applied unevenly. Phantom is designed to be applied “balanced” so it appear on pin 2 and 3 at the same time and cancels unless you are a phantom microphone and expecting it. The RCA 44BX, as near as I can tell, isn’t entirely balanced as designed and the 48 volts gets applied directly across the ribbon through the output transformer. Dynamic (moving coil) microphones can survive that, but you can launch a ribbon by doing that.
Also, I wouldn’t blow into that or any microphone. Depending on the ribbon microphone, you only do that once.
It mostly comes down to price and size.
USB preamps tend to have better noise rejection on the power supply than USB mics - they generally have more physical space for building in supply regulation. USB mics tend to be fairly tight on space, which makes supply regulation more difficult and more expensive.
To be fair, it is not entirely the fault of the USB device. The power supplied by many computers is much more noisy than is desirable for high quality audio applications. Again that largely comes down to cost and size, but this time on the computer side. As USB is mostly used for printers / scanners / external drives and such like, low noise is not a high priority for most computer manufacturers, as long as it does not cause a problem for “normal” USB devices. High quality audio devices on the other hand, can be very susceptible to noise.
The high point of sound damage is “Microphone” and “USB” in the same sentence. Flynwill found this damage in a non mic-level UCA202, but he had to work at it—drag out the scopes and meters. I found my all too obvious damage in a mic-to-usb adapter and most of the other people with mosquitoes are on USB microphones. Combining a super sensitive audio amplifier with ratty USB power is deadly.
Here’s a picture of the source of the “frying mosquitoes” in my case.
This is the USB power, measured inside my UCA-202 while in service. What is most interesting is not the burst of HF noise, but if you look close superimposed on it is a much lower frequency pulse about 100 uS long with a period of 1 mS (1 kHz). The computer supplying the power (via USB) is a small desk side with an Intel Motherboard that serves as my Home-Theater PC.
The UCA-202 actually uses pretty good design to isolate it’s input amplifiers from this noise. However, the issue in my case was that there were multiple ground paths. The UCA-202 is being used to digitize LPs from the stereo system. That stereo system is also connected via RCA cables to a Sony TV and to a CD/DVD player. The CD/DVD player is connected to the TV via HDMI as is the computer. So the issue is that those other ground paths between the UCA-202 and the computer are probably of lower impedance than the 24 gauge wire in the USB cable with the result that the noise ends up in the input. Not very much, somewhere around -85 dBFS, easily masked by the surface noise of a typical LP, but annoying none the less. I cured it by modifying the UCA-202 with an external 5V power supply.
OK, thanks. There seems to be a consensus that 24-bit is overkill for audio book narration. The reason I was wondering is that it sounds like meeting the ACX requirements, and presumably the requirements of other similar companies in the field, can be a struggle at times. One of those requirements is that Each uploaded file must have a noise floor no higher than -60dB RMS..
I don’t know if that is the most difficult spec to meet or if inputting 24-bit signal into Audacity would make any significant difference in the final product which is exported into 16-bit format for ACX. I did just find an article on the Audacity Wiki discussing bit depth.
Here are a few excerpts explaining why Audacity records in 32-bit floating point format that don’t seem to be directly on point, but may still be relevant:
The (theoretically audible) advantage of this is that 32-bit floating point format retains the original noise floor, and does not add noise. …
In many cases you will be exporting to a 16-bit format (for example if you are burning to a standard audio CD, that format is by definition 16-bit 44100 Hz). The advantage of using 32-bit float to work with holds even if you have to export to a 16-bit format. Using > Dither > on the Quality tab of Audacity Preferences will improve the sound quality of the exported file so there are only minimal (probably non-audible) effects of downsampling from 32-bit to 16-bit.
…
Which bit depth to use
If there will be no adjustment of gain after recording and no effects applied, the recording bit depth can be 1 bit more than the audio source bit depth without losing any quality.
If gain will be reduced after recording, recording with 1 bit more depth will avoid degradation.
When Gain Changes
If gain will be increased after recording, recording without degradation will occur if the number of bits equals the source bit depth plus 1 plus another one for each 6 dB or part of 6 dB of gain that will be added.
Where multiple operations are to be carried out on the recorded signal, each operation can be assessed in terms of how much gain it adds, allowing a margin of one bit for each 6 dB or part of 6 dB of gain, and another bit for each operation.
Using more bits than these figures will not give you more quality, but will enable more processing of the signal to occur before any degradation of quality occurs due to the number of bits in use.
using less bits than the above will cause some signal degradation. Whether it is noticeable will depend on the details of each case and the listener.
8-Bit
8-bit resolution is adequate for low quality audio, such as:
AM radio
78s
Microcassette
8 track
Telephone audio
Reel to reel at 3.75ips or slower
Pocket recorders at live events
Audacity does not itself support 8-bit recording. 16-bit is the nearest option. It is possible to export files in an 8-bit format, though Audacity defaults to exporting as 16-bit.
If medium quality sources are to be manipulated before saving the recording, it may be preferable to record in 16-bit to avoid any possible quality loss during application of effects. This does not apply to simple editing, and does not apply to low quality sources, whose resolution is below 8-bit.
16-Bit
16-bit matches audio CDs, and is thus suited where the better dynamic range and S/N ratio of CD quality audio is required. 16-bit is a good general purpose high quality setting. 16-bit recording is suitable for vinyl records.
24-Bit
24-bit recording may be used for signals that will be manipulated but still need to maintain the full 16-bit quality of CD audio. 24-bit is good for mastering.
If you’re merely listening to thousands of pounds of expertly chosen high end audio kit, and not doing large amounts of editing, there may be no real reason to exceed 24-bit depth.
32-Bit
If you want or need the highest standards (for example, operate a recording studio), expect to do a large amount of manipulation of the data before export, and have audio source equipment with an extremely low noise floor, 32-bit recording (which is the default setting in Audacity) will give the best possible quality and avoid the bit depth having any effect on the sound even after heavy manipulation of the audio.
Finding audio sources capable of providing signals with better dynamic range than 24-bit resolution is a demanding task. A 32-bit data stream records 65,000 times the dynamic range of 16-bit CD audio. In real world applications, a lot of those bits will be normally recording nothing but very low level background noise.
Maybe you guys could get manufacturers to send you a bunch of their equipment to play around with and you could run tests on them like they do over at Recording Hacks. Not that you have a lot of extra time, but it might be kind of fun.
It’s not the noise. Audacity uses 32-bit floating so it doesn’t overload. If you apply a filter or special effect that causes your sound to increase in volume, 16-bit sound would get destroyed. In 32-floating, you just bring the volume back down with any of the volume tools and keep right on working. It does eventually overload, too. But not likely because of any production tools you used.
Since the noise is determined mostly by the analog portion of a microphone system, 24-bit sound will allow you to have much better, more accurate and higher quality noise.
Each uploaded file must have a noise floor no higher than -60dB RMS…
It’s not difficult at all to meet that with good quality, not necessarily expensive equipment in a quiet room as I and others proved recently (rock band microphone, tiny sound mixer, MacBook Pro and Starbucks). What you can’t do is meet all three ACX specifications with a cheap microphone system connected to a crappy computer in a noisy environment. Almost all of our failing posters violate one or more of those conditions.
The rest of them have “magic” problems. “No, I’m sorry. You are doing everything right, but I have no idea why your computer is doing that.” Those are disturbing on many levels since we can’t see their computer. I have one of those. I have a hum problem that’s not put down to electrical connections or even audible…??
There is a really, really tiny subset of posters that have a perfectly fine recording and they’re measuring it wrong.
Oh, indeed meeting that noise floor requirement can be quite difficult. Starting with the issue of acoustic noise present in your “studio”. Your normal speaking voice (assuming the microphone is a few inches away) will have a peak SPL somewhere in the 80-90 dB range. ACX allows you 57 dB between the peak program (which they say should be at -3 dbFS) and your RMS noise (less than 60 dbFS). That means that the background noise in your studio needs to be 23 dB SPL or better – which is very quiet. Obviously you can make it better by getting closer to the mic and by speaking louder. It’s all relative as they say.
Your second problem is going to be electronic noise from various sources, including the microphone itself (good ones have an equivalent noise level below 20 dB SPL) the microphone preamplifier, and noise introduced by whatever you use to digitize. Any of these will completely swamp the quantization noise of encoding in 16 bits (-96 dbFS) – if we assume you have a loud voice which is peaking at 90 dB SPL, and you set up to give yourself 6 dB of headroom above that (ie 96 db SPL will clip) then that quantization noise is the equivalent of 0 dB SPL room noise.
Yes, in simpler terms, signal processing (meaning all of the various manipulations we might subject our sound to) involves arithmetic, and arithmetic with integer formats (eg 16-bit) leads to accumulating round-off errors, which will manifest themselves as additional noise. Worse the noise may be correlated with the actual signal in nasty ways. You can actually do quite a lot with 16-bit integer processing before this noise rises up to anything close to that -60 dBFS ACX requirement. But Audacity does do all of it’s processing in floating point to avoid the issue.
I was thinking of exporting the original raw version to .FLAC and then just working with it and saving it in Audacity until finished, and then at that point exporting it to .mp3. Maybe saving it in Audacity as a couple of different versions if there are a lot of changes, especially if it is over several days or so. WAV, Audacity & FLAC are all lossless, correct?
You can totally do that, but FLAC isn’t the universal file format that WAV is. Audacity picked WAV as the default format because all three major computer platforms can effortlessly open them. Never do production in MP3.
Warner Records called. They want WAVs of your latest book . How soon can you post one of your performances?
In Recording Hacks reviews of mics, such as those listed in their Entry Level $200 Condenser Microphones and their USB Dynamic Microphone Shootout and their Ultimate Podcast Microphone Shootout not to mention their reviews of interfaces, such as SM7B Budget Audio Interface Shootoutthey listed things like the “self-noise”, bit rate and so on.
