Record at 96kHz/24bit 48kHz/24bit or 44.1kHz/16 ?

eco_bach

Well-known member
Hi
Just bought a zoomH1 and plan on doing some dual sync recording with my 7D. Will be using FCP as my NLE.
With a 4GB micro SD I get just under 2, 4 and 6.25 hrs recording time for the 3 different quality options in WAV format.

96kHz/24bit

48kHz/24bit or

44.1kHz/16bit

Is the extra quality of 96/24 worth the much reduced recording time, or should I stick with 44.1kHz/16?
Does it depend substantially on the actual audio being recorded ie spoken voice vs classical symphony vs nature sounds?

Any audiophiles care to chime in?
 
IMO, Audiophiles are those guys that spend $5,000 on a bucket of rocks to make their hi-fi sound better, but that's my own little side rant.

Sound for picture should almost always be 48kHz (or a multiple, like 96) at the greatest bit-depth you have available. Generally, you'll only really benefit from 96 or more KHz right now if you're doing FX recording. (Cue everyone else on here getting into an argument about sample rate in 3...2...)
 
Even for theatrical release, production dialogue at 48/24 is what's used. Going any higher is a waste of space. Now, if you are recording sound fx, then do the higher resolution.
 
Typically these small recorders only achieve about 15-17 bit dynamic range (even the very best field recorders get only 19-20 bits worth in reality), so 24 bits is more marketing than reality. 48 kHz is the standard video sample rate, so using that makes things simple. If, for some reason, you are short on card space, using 44.1/16 would be fine. Convert that to 48 kHz before dropping the audio into the editor (some editors convert automatically) and nobody will notice any difference.
 
As an audio post guy I really like 96k. When using NR and other processing I get significantly fewer artifacts at the higher sample rate.
 
With the settings you've just given for recording purposes,

1. 24/96 for SFX
2. 24/48 for Film Dialogue
3. 24/44 or 16/44 for Music-related

If you're running out of space and need more recording time and have no place to transfer, switch to MP3 mode. Don't kill me, but it'll save you. If your device has the capability to do 24/192, go for it.

Spoken voice for normal narration work, 24/48 or 24/44
Spoken voice for sound design purposes, 24/92, extreme conditions 24/192
Classical recording, 24/88 or 24/92. Demo purposes 24/44 or 24/48, all depending if it's going down to video or broadcast. Higher the sample rate, the less harmonic distortion on materials/sources with high transients. More than just harmonics.
SFX for library and sound design work, 24/192. Ambience, 24/92 or 24/192 if extreme/unique design situations.
 
Hi
Just bought a zoomH1 and plan on doing some dual sync recording with my 7D. Will be using FCP as my NLE.
With a 4GB micro SD I get just under 2, 4 and 6.25 hrs recording time for the 3 different quality options in WAV format.

96kHz/24bit

48kHz/24bit or

44.1kHz/16bit

Is the extra quality of 96/24 worth the much reduced recording time, or should I stick with 44.1kHz/16?
Does it depend substantially on the actual audio being recorded ie spoken voice vs classical symphony vs nature sounds?

Any audiophiles care to chime in?

If you are recording for video, do it at 24-bit / 48kHz.

Don't use 16-bit at all - 24-bit gives you the safety headroom and the ability to manipulate in a DAW.

I was at Abbey Road a couple of years ago with a multi-mic. recording of the LSO (all digital microphones) and the whole lot was recorded at 24-bit / 44.1 kHz (as it was for CD).

I record mostly classical for CD release and record at 24/96 or 24/88.2 - but foe video, stick with 24/48 and don't use 16-bit at all.
 
FX are more common at 96 now but it's still not the rule. Dialog is 48/24.

I'm a post guy and I can see how for some processing 96 would help but the only time I got a show at 96 I converted down to 48 because it was such a dog to deal with. That was a few generations of computers and drives ago so that probably isn't the case now. But I still wouldn't recommend 96 not enough up side for the increased workflow IMHO.
 
Since CD's are only 44.1/16, and most people are downloading MP3's anyway these days, I think recording music at any higher resolution is a waste of space, but that's just me.
 
If you are recording for video, do it at 24-bit / 48kHz.

Don't use 16-bit at all - 24-bit gives you the safety headroom and the ability to manipulate in a DAW.

Do you seriously believe that a $99 Zoom H1 makes 24 bit files with something musical in them in the 8 least significant bits?
 
Do you seriously believe that a $99 Zoom H1 makes 24 bit files with something musical in them in the 8 least significant bits?

It's not even about the last 8-bits, but the refinement of the gain step from absolute zero to the highest peak of the signal. Even the most expensive piece of audio gear won't do precisely 24-bit.
 
It's not even about the last 8-bits, but the refinement of the gain step from absolute zero to the highest peak of the signal. Even the most expensive piece of audio gear won't do precisely 24-bit.

Those "refinements in gain steps" are in the least significant bits. When they are random noise, there is no refinement. In that case those bits are useless HD or flash card space eaters.

The best recorders get about 120 dB dynamic range, which is equal to 20 bits. As it is more practical to use 24 bit sample depth (3 bytes per sample) than 20 bits (2.5 bytes per sample), 24 bits is used. So with the best recorders like Sound Devices, Nagra, Deva etc only the last 4 bits are trash. With cheap recorders with only 16 bit real resolution, or less, churning out 24 bit samples is just marketing with no practical value at all.
 
Since CD's are only 44.1/16, and most people are downloading MP3's anyway these days, I think recording music at any higher resolution is a waste of space, but that's just me.

NO - NO - NO - NO - NO !

Always record at least 24-bits - it gives you the safety headroom and the ability to manipulate in a DAW without degrading the final CD quality.

44.1 kHz is fine for CD (or 48kHz for video) - higher sample rates only extend the top end frequency response. What recording at a high sample rate (88.2 or 96 kHz, for example) does is to allow a gentle filter rather than a brick-wall filter at 20kHz which you have to have when recording at 44.1 or 48 kHz sample frequency.

MP3 is cr*p and I cannot listen to it - it's just horrible.

But less and less people are downloading MP3 as storage gets cheaper and lossless formats are becoming more common. Some record companies now have 24-bit / 96 kHz downloads in lossless FLAC format - this is vastly better than CD quality.

So recording low quality for MP3 is a dead-end and will not future-proof the recording at all.
 
I wouldn't go quite that far but I agree with the thought. I'm not a big fan of high sample rates but at a strong advocate of 24 bit.
People love to argue the math but even on cheap recorders I can hear the difference of 24 bit. If it's all trash it's good sounding trash, or maybe it's just getting you a real 16 bits instead of the 10 or so that most 16 bit recordings actually had. I don't really care it sounds better to me so I do it.

High sample rate for most things is probably a waste. I certainly think it is for dialog. But on things that you are going to process a lot it gives the computer more to chew on. So I can see it for music and I use it for some FX.

There is though kind of two arguments. The first is what to record at and the second is what to work at. I always work at 48/24 even with 16 bit recordings. When you bring in the files there is nothing in those last bits but the minute you work on them there is. Reverb is smoother at 24, EQ should be smoother, volume moves are smoother etc. But I'm doing mostly FX so I'm "growing" the sounds. Our sessions are at 48 and when we record FX at 96 they are worked on at 96 and converted down to put in the timeline.

My problem with working at 96 all the time is that I have almost no FX at that rate and never get production elements at that rate. So it seems a big waste of space with no real upside.
Music production though you could be 96/24 the whole time and with the amount of processing in a lot of music that could be a big plus.

From a technical stand point you never want to be in the position of having to up sample for delivery. So you are much better off working at a higher quality than your delivery format.

It used to be a problem because conversion took a LONG time and the conversion often had some quality loss. Today conversions are very quick and the algorithms are very good so there really isn't a reason to lock your self into working at the delivery spec. And as John points out delivery spec's are likely to going up.

NO - NO - NO - NO - NO !

Always record at least 24-bits - it gives you the safety headroom and the ability to manipulate in a DAW without degrading the final CD quality.

44.1 kHz is fine for CD (or 48kHz for video) - higher sample rates only extend the top end frequency response. What recording at a high sample rate (88.2 or 96 kHz, for example) does is to allow a gentle filter rather than a brick-wall filter at 20kHz which you have to have when recording at 44.1 or 48 kHz sample frequency.

MP3 is cr*p and I cannot listen to it - it's just horrible.

But less and less people are downloading MP3 as storage gets cheaper and lossless formats are becoming more common. Some record companies now have 24-bit / 96 kHz downloads in lossless FLAC format - this is vastly better than CD quality.

So recording low quality for MP3 is a dead-end and will not future-proof the recording at all.
 
Back
Top