What Format and Settings Do You Record and Edit At?

DavidNJ

DavidNJ

Carbonite Member
PCM at 33, 48, 88,96, 176, 192Khz? 16, 24, or 32 (floating) bits?

Or 1-bit pulse density modulation at 2.8 or 5.8Mhz?

Also, do you have a preferred tool.

Why am I asking? I have been recording 48khz/16-bit on the camera and editing in Adobe Audition (I really wish Cool Edit Pro was still independent).

Now, I can record at up to 192khz 24 and 1-bit 2.8Mhz. However, Audition doesn't support 1-bit so I converted to 32-bit floating 192.

Not sure if that was best or what is right...so I am asking greater minds then mine!

Thanks,

David
 
24bit/48khz is pretty standard. Anything higher is generally considered a waste of space, except perhaps for highly specialized applications...
 
XDCAM EX and HDV are standards also...neither any good. I've seen audio people saying always 24-bit...not sure on the sampling frequency. However if you are delivering 44khz or 48khz, it seems 96khz should be a minimum. 24-bit x 96khz is 2.3 Mhz, which is close to 1-bitx2.8Mhz,which is roughly equal to floating 32-bit 192khz in quality.
 
Bascially, Bit Depth gives us our dynamic range, and Sample Rate determines the highest frequency we can record.

So at 16-bit we have a dynamic range of 96dB. 24-bit : 144dB, and 32-bit: 196dB.
Our maximum frequency is simply half the sampling rate, eg 48kHz rate : 24kHz limit.

It is generally accepted that human hearing is roughly 20Hz - 20kHz, with a dynamic range somewhere around 140dB. Orchestral music has a dynamic range around 80dB.

The mics we use are generally rated to 16-20kHz, and so are the speakers our work is going to be played back on.

So just based on those numbers, 24bit/48kHz should be plenty.
 
With most mics, the benefits of 96kHz recording are subtle to the point of being nigh imperceptible. However, once you start pitch-shifting things downward that high-frequency detail comes in real handy. If nothing else, its a way to "future-proof" your audio assets as home theater formats progress towards high-resolution audio.

Hard-drive space is awfully cheap nowdays, so it makes sense to record at higher sample rates. When I started out 10 GB drives were the norm, but now I find myself shopping around for a new 1 TB drive. This progression has been even more dramatic for solid-state media like compact flash. Doubling the storage space by recording at a higher rate doesn't strike me as a prohibitive issue, anymore.

Just be absolutely certain that everyone in the post-production workflow is able to accomodate the recordings. Until rather recently many smaller post outfits balked at 24-bit 48kHz recordings, and there are still plenty of facilities that haven't made the leap to Protools HD.
 
It is my understanding that would be if sampling could capture everything in its range. Usually toward the ends of the range you get artifacts from frequencies above your nyquist limit.

Adjusting the signal (video, audio, accelerometer, pressure transducer, etc.) in post, if the capture is approximately the same signal delivered in the final result, won't that result contain more of the artifacts? If the noise being sampled had an 80db range, and was recording at a 96db range, it would require that the recorder had been calibrated rather closely to the signal and the signal didn't diverge from that range?

From your comment, 48khz is the minimum to capture or display the auditory range and 24-bit would be necessary to capture the full range in detail. Wouldn't that imply minimally 88 or 96khz at 24-bit?
 
Overlapping posts! 2:43 and 2:50...

Couldyou elaborate on "once you start pitch-shifting things downward that high-frequency detail comes in real handy."?

Note that the 24-bit@o6khz recording is roughly the same size as 1-bit@2.8Mhz, although the latter is equivalent to 32-bit@192khz.
 
Another point to make is that DAC's and ADC's are not perfect technologies. And they both generate less errors when working with 96kHz compared to 48k, which is especially good if you're on the cheaper end. Once you start spending upwards of $5-10k on a external DAC/ADC, this doesn't apply as much but for those of us not working at that level of gear, it's important to note.
 
Would that be more reflected in increased sample resolution (24 vs 16 bit) in contrast with sample frequency (48vs96vs192 kHz)?
 
DavidNJ said:
Couldyou elaborate on "once you start pitch-shifting things downward that high-frequency detail comes in real handy."?
Click to expand...
Sure thing. Let's say you were working on some pretty out-there sci-fi sound design, and were trying to come up with vocalizations for a giant mutant space-goat. Being a fan of early musique concrete, you decided to go for the classic slow-speed playback gag (ie. a 12 IPS recording played back at 6 IPS) with a library recording of a bleating goat (CD quality, 16-bit 44.1 kHz). This is pretty easy to accomplish in any DAW, and you simply down-pitch it the appropriate amount (one octave).

Unfortunately, the resulting sound lacks detail and sounds artificial (space-goats notwithstanding). When you halve the playback speed, the resulting sound has an effective sample rate of half the original rate. Therefore, the new goat bleat tops out around 22.05 kHz, with a nyquist frequency of 11.025 kHz. All those harmonic details that our ears are looking for are absent in the new effect.

If, on the other hand, you were using a newer 96 kHz recording, the resulting pitch-shifted effect would have a nyquist frequency of 24 kHz. Of course, only a handful of mics can accurately pick up sounds in excess of 20 kHz, but most mics will give you a fairly gentle high-frequency roll-off above their effective frequency range. Even still, that extra bit of detail makes a big difference.

If you're lucky enough to have access to an unusually well-stocked mic cabinet (and a goat), you may even try recording new goat vocals with a mic that has a flat response up to 80 kHz. Roll those tracks at 192 kHz, and you could easily pitch them down two whole octaves before things start sounding wrong.
 
Hard drive space aside there are no practical reasons to record 96 "normally". SMuir's point is valid to a point with FX recording. I don't think it's so much the high detail that makes the difference since A) most adults have a pretty good roll off once you start to get above 10-11K so we are not going to miss much up there and B) mics don't pick up much up there to move down. What you do gain is a smoothness to the (slowed down) sound that is a plus. People get all wrapped up in the supposed technical reasons it "should" be better but they mostly just don't happen in the real world. There is the great "home theatre argument", except 1/2 the mix houses (for ad's) still insist on making sure your mix works in mono on 2" speakers because of the cheap TV crowd. I have seen pretty much no "home theatre" that was set up and balanced correctly that wasn't owned by a sound person. And finally final delivery is so far down the scale from even "good" quality 44.1/16 that it makes it all sort of a wank. Maybe when everyone is on BluRay? Most people are watching highly compressed network/ cable feeds that are essentially giving them normalized MP3 files. Dolby Digital is at 44.1/16 and in an AC3 (pretty much the same as an MP3) compressed stream. SDDS (in theaters) plays off a CD. So thats eight tracks of feature length plus the TC track on a CD. Do the math it's another highly compressed format. Prologic is 4.o matrixed into two channels. The point is that for certain post reasons it might be useful to go higher than 48 but only as acquisition not as a working rate and it makes no sense to record production higher than 48.
 
Certainly the release has to be in a distribution friendly format. However, applying numerous speech enhancements and noise removal is common. The higher frequency and resolution oversampling allows that to be done cleanly.

We are so used to video cameras being directly sampled or under sampled we might have missed the advantages. Oversampling moves all sorts of artifacts out of the range of your final product.
 
DavidNJ said:
Certainly the release has to be in a distribution friendly format. However, applying numerous speech enhancements and noise removal is common. The higher frequency and resolution oversampling allows that to be done cleanly.

We are so used to video cameras being directly sampled or under sampled we might have missed the advantages. Oversampling moves all sorts of artifacts out of the range of your final product.
Click to expand...

Well I wasn't talking about video I was talking about 35mm feature films. There is even less reason if your shooting video.
 
Video can have higher contrast and detail then film...and is often viewed in a quieter environment with better acoustics.
 
DavidNJ said:
Video can have higher contrast and detail then film...and is often viewed in a quieter environment with better acoustics.
Click to expand...

In what universe? Networks compress the hell out of the audio just because the home environment is usually so noisy. Theaters, teenagers on cel phones not withstanding, are a much more controlled and isolated listening environment. They have better sound systems that have better maintenance and calibration. Video has to get up to something like 8k before it matches the resolution of 35mm film. Now a lot of films are going through 4K digital intermediaries these days so what's on the screen may only be 4K. BUT that is one heck of a lot better resolution than any home system is getting. Good 4k digital cinema is streaming 149 gigs per minute, so you have to change out those BlueRay disks at 3 a minute to keep up with that. Or to put it another way it would take 270 BlueRay disks for your nominal 90 minute epic! And that is about 1/2 of 35mm resolution. RED is probably the only camera that even comes close to films contrast ratio, but that also depends on what film stock your using. What ever video is capable in acquisition will get trimmed back to "safe" levels if it's going to a TV, so any really high contrast video is only going to be seen in a theatre, not off a DVD or BR disk. And we haven't even talked about 72mm...
As a point of interest, before relatively light and portable camcorders TV news used to be shot on film. Supper eight was used mostly, because it was a higher resolution format than broadcast video.
 
Most of the movie theater I've been too in NJ and NY are trash. The rooms are long and narrow. The projection equipment makes the 24" TV in my kitchen look good. The sound his boomy with no clarity.

Video easily beats film because of losses in transfer. Not to debate the technical issues myself, John Galt of Panavison has presented this: How Many Pixels In Lawrence of Arabia?


I've used that in trying to talking about 5DM2 vs EX1/3 images.
 
You must log in or register to reply here.
Back
Top