"one-time" audio digitizing

What do you want the audio for? If it is an announcement, consider the environment in which it will be used.

Don't forget the rest of the audio chain, microphone and acoustic environment. It could very easily be your constraint.

Op Thu, 18 Aug 2011 07:30:59 +0200 schreef Don Y :

How are the samples stored?

Why does it need to be "best"?

Are you sure it's not good enough?

Why a sound card?

Simple Delta-Sigma ADC -> SPI input -> SD card / USB MSC

Have you considered using someone like Splashmedia to use their soundstage and recording equipment to do everything for you? Or are the sounds outdoors only/not transportable?

Yes. One of the applications is capturing speech. But, I will try to get access to an anechoic chamber for that. But, I don't want to have to deal with "whatever" mechanisms those folks happen to have to capture the signal (e.g., pro gear, tape, or some ad hoc collection of cruft that a professor may have assembled, etc.)

Other applications have the "audio" source available already in other formats. E.g., off tape or proprietary devices (where "audio" is used loosely to refer to the bandwidth of the signal, not its actual *nature*)

Some will be "recorded live"; others off "tape"; still others "dumped" from proprietary devices in an analog form (i.e., devices that weren't designed to export the data in any sort of digital format)

As opposed to *worst*? :> Note I use "best" as a relative term, here... not *absolute*. I.e., best among competing viable options vs. The Best.

I've never tried capturing audio and *looking* at the quality of the captured signal -- since any sort of quantitative evaluation of the capture mechanism/process would require knowledge of the "quality" of the source signal itself. (e.g., if you're digitizing ADPCM telecom audio you have a much lower bar than if you're trying to digitize seismic events! :> )

Something COTS that does this with the same sort of resolution/bandwidth?

See above.

They aren't all "sounds", strictly speaking. Rather, signals in the audio-ish part of the frequency spectrum.

I'm trying to see if it is *practical* to develop this capability for my own use of if I *must* rely on someone else to do it for me. If the latter case, then it means I have to plan more carefully how and when I undertake each activity.

E.g., if access to that capability is constrained (due to location, cost, competition with other customers, etc.) then you have to factor those constraints into your planned use/need. OTOH, if you have the capability yourself, you have more flexibility in how you utilize it.

(imagine having to share a PC with a group of people... it would undoubtedly impact how you arrange your workday and project schedule!)

It's pretty easy to build a "prosumer" grade recording studio in a closet or den. You just need a low noise environment and a decent mic and sound card (or sound USB thingie). If echoes are a problem you _don't_ need a fancy anechoic chamber -- just get a cloths drying rack and hang a blanket (or some wet cloths) on it.

If you're going to be doing empirical cut-and-try, then you may want to do all the cut and try with your own equipment, then when you know _exactly_ what you want go to the 'pro' sound people with your recordings, tell them what you want and what you don't want, and re- record as necessary.

No way. If a signal is manipulated, then there is a loss.

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

That is a big problem even on the country side.

The anechoic environment is a lesser problem. Just take a mike and cover yourself with a blanket.

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

Are these announcements or just sound samples. Music? Announcement tones? Or what? What is the playback environment/device?

IMHO unless you are feeding into some high quality audio gear, a PC sound card and capture will work. If you have analog sound recordings, sample and use something like Audicty to cleanup.

If what you want is voice capture look into a local radio station. A lot of radio stations now days are just remote control boxes but if you can find a local one with real people they will usually have a sound booth to do commercials, PSAs and such. We did some campaign ads. They recorded the ad with leadin/out music and sent us the mp3 file, but I could have gotten them in .wav format.

Call the smaller local station(s) and get to the sales folks. Tell them what you want to do. They will probably work with you. They need any sales revenue they can get and the booths sit empty much of the time.

Not, strictly speaking, "audio". Rather, low frequency signals (a few Hz to high end of audio range)

The reason for the "overkill" is to ensure ample "signal" to crunch numbers on it. E.g., not having to worry about clipping *or* falling into the noise floor.

A combination of several different types of signals. Best to think of them as just "audio frequency". E.g., some are voice samples, seismic events, "biological" events, etc.

It's not a case of simply "changing formats" (analog -> digital) but, rather, being able to accurately analyze and manipulate those signals without worrying about other artifacts that weren't present in the original(s).

Ah, that's an interesting idea! Not the *broadcast* booth...

There are similar places in town (pro recording studios, university, etc.) so getting access to one isn't the problem. Rather, the problem is one of logistics -- "I need (just) this digitized, today. Can you squeeze me in for half an hour?" (plus the overhead of getting to and from the facility, etc.) It forces me to get *all* my ducks lined up ahead of time (like running punched cards on a twice daily schedule and spending the rest of the time grumbling because you forgot a JCL card and the pack was dumped -- leaving you twiddling your thumbs until the next run!)

Would have to be a closet as the rest of the house is far too "open" to be able to control/hide extraneous sound sources.

I fear the (internal) sound card would place the PC in that same environment (fan noise, etc.). As well as worrying about all the electrical noise *in* the PC.

This was why I was thinking of looking at the Extigy (external, hangs off USB bus) in comparison with the Audigy2 (internal, hangs off PCI bus). I.e., presumably I could locate the PC "outside" and just snake a long-ish USB cable into the "booth".

Is this "enough"? Or, just "better than nothing"? (seems like that closet is going to get awfully crowded! :>

For speech this would be a problem. I.e., you can't get the *same* speech at a later date. So, ideally, when you get *the* speech samples that you want, you want them to be as clean as possible (no do-overs)

Yes. Even in our "quiet" neighborhood, you can't safeguard against a firetruck driving by (I have *no* idea why a firetruck is dispatched along with each ambulance -- is that in case the patient is ON FIRE??!) or a warthog flying overhead, etc.

And, you still have the acoustic and electrical noise of the PC to contend with...

Hmmm, let's see... sitting in a closet with a blanket over your head holding a phallic symbol... um, I think I'll pass on that, thankyouverymuch!

;-)

My dad used to be the chairman of the board of directors of the Boring Volunteer Fire Department, in Boring Oregon (as well as being the assistant chief -- but how he managed to both report to and be the boss of the chief is a different story). There, ambulances were owned and operated by private companies, and were responsible for patient transportation and upkeep starting from roughly when the patient was put onto the gurney. The fire department was responsible for getting the patient stabilized, extracting the patient from whatever hole or crashed car or tree or whatever they were stuck in, etc.

The fire department can do this either by having a separate rig (called the "rescue rig" at the Boring VFD), which is dispatched any time that there may be injured people, or by putting all the rescue stuff onto a fire truck, which is dispatched any time there may be a fire _or_ injured people.

Since the fire truck often needs to get dispatched _anyway_, many fire departments have taken to reducing the number of truck drivers (which are more expensive to have, even in a volunteer group) by going the "everything on one truck" route.

It's up to you. I've successfully made voice recordings in a corner room of a downtown building, with only closed windows and foam rubber on the walls for sound deadening, and a sign on the door that says "quiet" to reduce external noise. You're going to have to take multiple passes at it anyway -- the recordings with the occasional honk or siren in them get put into the same bin as the ones with the coughs or cussing. Or you end up taking the front half of one clip and splice it into the back half of another.

No. *Two* vehicles are dispatched for every "first aid" call. Both are dispatched from the Fire Department (though I can't attest to whether both happen to come from the same *station*). I.e., you will see a truck *and* an ambulance driving down the street literally seconds apart!

Exactly. This is the (big) downside with trying to do it at a "professional establishment": get the samples home and start analyzing them... then discover they aren't useful -- back to the studio, etc. You spend all your time on (silly) "overhead" (i.e., driving back and forth)

If it makes sense (i.e., if good results are COMMON with this sort of "soundcard" approach), then I'm willing to "formally" (though temporarily) convert one of the walk-in closets to a "mini-studio" for the duration. (I've found that carpeting the walls has a dramatic deadening effect on a small-ish enclosure)

I just don't want to discover that the (electrical/acoustic) noise floor sucks, etc. (this is why the external sound "box" seems like the safer thing to try -- all else being equal)

It's mainly for extra muscle, in case it's needed. Also, the fire truck carries a bunch of heavy duty rescue gear that the ambulance usually does not (same thing, really).

And why not? If the fire truck isn't otherwise occupied, it's just the cost of the drive to site and back - and if it is occupied, they'll just send the ambulance. And the fire truck often leaves the scene well before the ambulance does, if they're not needed.

Besides, it's an excuse to get on the road and turn on the lights and sirens.

Sure, if you have resources to spare, "why not"? But, firefighters here take 9 furlough days annually (budget shortfalls) so why place extra demands on their time? (as well as wear and tear on the vehicles)

If they are responding to a call re: "chest pains", it's doubtful the ambulance driver is going to need "the jaws of life", a few hundred feet of fire hose *or* 50 ft of ladder!

(we see these pairs of vehicles pretty regularly in the neighborhood so it makes you wonder. E.g., neighbor was frequent consumer of these resources (since passed away) so it was not uncommon to find our driveway blocked with emergency vehicles)

But that has to impact their ability to respond to a *real* need, right? Otherwise, wouldn't they locate the firestation *in* this residential neighborhood? (all of the fire stations here are located on main roads -- not "back roads" peppered with "speed bumps", etc.)

For example, a truck dispatched here would take several minutes to get back out of the neighborhood to the "main road" about 0.5 mi due south. They'd have to circle the neighborhood to get turned around (since even a K turn would require driving to the nearest intersection and jockeying the vehicle back and forth). Cross at least three speed bumps (they slow to a *crawl* when they cross these) and *then* they can get to the intersection where they can pull into traffic (I'm not sure that the intersection has one of those "emergency vehicle override" traffic signals that would change the signal to their favor)

The fire station is less than 2 miles from here. I suspect they could drive to this subdivision from the station in less time then they could get back to the main road *from* here.

There *must* be a formal policy about this stuff. And, hopefully it was created by "smart minds" (and not folks with a financial stake in the matter). E.g., why do they drive the trucks to the grocery store? Can't "Joe" drive to the store in a regular "city vehicle" *and*, in the off chance that a "call" comes in, drop everything and drive (in that same city vehicle) to *meet* his fellow firefighters "at the scene"? Why should three or four firemen make the trip to the store *with* the equipment?

(Note that I'm not complaining here but, rather, truly wondering what the "science" behind any such "policy" must be)

So, instead of putting your energy into bitching about it here, put it into running for your local fire board, or city council, or whatever group it is that makes these decisions.

Having seen exactly how much unpaid time my dad put into participating in running a fire district, with little reward from the citizenry other than just the sort of pointless complaining I'm seeing from you, I have little patience reading it.

You think you can do a better job? Do it. Run for the fire department board of directors, or the town council, or whatever it is that runs your local fire department. You don't want to devote two or three evenings every week away from your family, with the time either mostly or completely unpaid? Then don't bitch -- someone, probably several someones, are making informed decisions on your behalf. And unless you're living in a _big_ city, they ain't getting rich at it. You, on the other hand, are not devoting a minute of your time to anything productive when you bitch about it in a forum that isn't even going to be seen by anyone local.

If you're not smart enough to go over the recordings there, and if they're neither smart enough nor responsible enough to make sure you sit down and listen before you leave, then you both deserve what you get.

As I said, I've had good results, in environments that "shouldn't" be good.

You may well discover that, the first time around. You always have to mess around a bit if you want quality work.