Does anybody in here use 24/96 recording?

MattinSTL

MattinSTL

Sound Modulator
I recently got my 671 back from Marantz and now it does super-clean 24/96 in a way that makes windows happy, but really I wonder how many people use 24/96?

I'm sure music studios do... and maybe a few guys who do a LOT of processing on the audio, but just how many people IS that anyway?

Even with the Microtrack 2496 I think we're at a point where 2496 is a marketing term more then a huge benefit? I realize that the bit depth and resolution increase allows for huge transparent moves in POST.

I guess when 16gig+ cards are dirt cheap this will be a non-issue... it's just funny that a 4gig card can be totally filled in a couple hours... and even when listening intently the sound is just a little bit better... noticeable, but not "wow".

I'm curious if most guys with 2496 ability are also using 1648 most of the time?
 
Hey Matt,

I record on the Nagra Area-BB+ recorders quite often, and YES to 24bit, but HELL NO to 96kHz.

The 24 bit should in theory give you drastically improved mapping of audio samples, and allow (I believe) you to achieve greater headroom and accuracy in the recording.

As far as 96kHz goes, the Nyquist limit states that the sample rate should be at least double what the maximum frequency the system will be called on to carry is. For us, thats usually well (well, WELL) below 20kHz, so therefore around 40Khz is appropriate (i.e. 44.1, 48kHz), but 96kHz is pointless.

i think the true magic of recording 24bit is in the mix... if you keep your project on 24 until the last possible step, your final dithered 16bit mix will have a much better dynamic range, and more accurate audio.
 
I don't think that 96 is worth it for dialogue, or really much of anything. I have done some stereo piano recordings at 24/96 and I find it EXTREMELY hard to tell the difference. I don't personally believe I could reliably pick it out in a blind A/B test. 96 takes something like 4x as much space and 4x the processing power. You will also most likely have to perform a sample rate conversion at the end to get your audio back to 16/48 anyway.

24 bit is crucial for mixing, and it is always a good choice to use 24 bit source files when you can. The extra headroom is needed by the mixer when you start adding together lots of tracks.
 
I use 24/48k when recording, or 24/44100 if doing not video.

Many tests done by sensible people have shown that the difference between 16/44100 (CD:s) and 24/96k can not be heard even with good equipment in blind tests. 1644 and 1648 are good enough for everything final.

Like previous posters say 24 bits gives more safety in setting the levels & mixing, so that is very sensible, but the product will be downsampled to 16 bits in the end.

Thus: 24 yes, 96 no.
 
I think its important to clarify the 24 bit thing:

Yes, it IS safety. But theres more to it than that. Even if your ear cant hear the difference between 16 and 24 bit, the difference in terms of accuracy when sampling is HUGE.

Now, cut to mixing this project when you have multiple layers of sound: fx, v/o. dialogue and ambi tracks - all those slight 16 bit imperfections start coming together, and at the end of the day, the dynamic range of the track suffers vastly considering even if you record at 16bit, almost no system ever ACTUALLY achieves 16bit resolution when mixing in 16bit.

BUT if the whole signal chain is in 24, when you mix down, you'll keep more dynamic range - and thats why its good.
 
16-bit can also suffer when running through FX chains like compression, reverb, etc. That's another reason music is often mastered in 24-bit, then dithered down to 16-bit as a last step.
 
tigeba said:
I have done some stereo piano recordings at 24/96 and I find it EXTREMELY hard to tell the difference. I don't personally believe I could reliably pick it out in a blind A/B test.
Click to expand...

I agree with most everything in the thread here, but not quite with this. I do about 100 classical piano recordings a year, and can say that indeed, there can be a great difference, especially with current condenser mics. It's very hard to get a recording that doesn't have a certain brittleness on the top end when recording classical grand performances, especially when you want to capture some of the hall ambience, and higher sample rates can actually make a significant difference.

However, this is one of the trickiest things to record well; material without the complexity and transients such as dialog, strings, etc. are unlikely to be very noticeably influenced.

And further, of course, the rest of the recording chain also has a great effect. Some mics don't emphasize the high end as much, and thus don't have that brittleness, ditto for preamps, converters. My solution in general is not to record at 88.2 or 98.6, but usually to choose more appropriate mics. Ribbons, e.g., can be great for capturing grands with a really smooth high end...

24 bit is crucial for mixing, and it is always a good choice to use 24 bit source files when you can. The extra headroom is needed by the mixer when you start adding together lots of tracks.
Click to expand...

I highly recommend 24-bit recording, and always use this sample size myself, but just for the record, on most DAWs and many mutitrackers the mix busses and internal representation have a lot more resolution than the sample size... ProTools, for example, handles everything internally at something like 48 bits. So headroom/resolution loss in mixing is not necessarily a major issue.
 
Ullanta,

I agree with you here, a ribbon mic is going to be far more effective than upping the sample rate to get smooth top end. In my own case (Yamaha C3 grand in an untreated living room -> Neumann KM84i -> Neve 1272 -> ProTools HD ) I don't find the higher sample rate to be all that compelling. That being said, I generally do it anyway because it's not a huge multitrack project, and I'm ultimately not going to put a bunch of plugins on it or anything. If I was recording piano as part of a multitrack project, I probably wouldn't bother with 24/96



ProTools TDM systems use double precision (48 bits) and many(most?) plugins maintain double precision across the entire recording chain. I am not sure about the specifics of ProTools LE or other host systems, but they are generally going to be 32 bit floating point, which is again, a lot of dynamic range.
 
Petrus said:
I use 24/48k when recording, or 24/44100 if doing not video.
Click to expand...

I agree. Though as Exile once pointed out, if you're recording sound effects, do 96k or 192k. If you've got a sample at 192k, you can slow it down 4x and still have 48k. Or something :)
 
Exile is loaded with good info that you never thought about... that guy needs to condense some of his experiences into a "Big Thread" also.

Unfortunately for us the only way you can tap some of that knowledge is by asking the right question, at the right time, then the planets align and he spits out the answer you needed in a short... broken... sentence :-)
 
Well, if you record at 192k and slow it down to 25% or something, the clip is still at 192k at least in my audio studio softwear. Slowing down is not done by lowering the sample rate... If you think that by recording at 4x oversampling you can keep the highs you would need a microphone that can record to 80 kHz or something to have 20 kHz highs in the slow version.

Thus I do not buy Exiles views. Or what do I miss here?

Thus I see no reason to oversample, but there is a good reason to use more bit depth than needed in the final product .
 
Petrus said:
Well, if you record at 192k and slow it down to 25% or something, the clip is still at 192k at least in my audio studio softwear. Slowing down is not done by lowering the sample rate... If you think that by recording at 4x oversampling you can keep the highs you would need a microphone that can record to 80 kHz or something to have 20 kHz highs in the slow version.
Click to expand...

Your mics don't go to 80k? :p

You can certainly take a 192k audio file and futz with the headers (using something like SoundHack) and change it to a 48k file. Your audio software might resample it to 192k, but the fact of the matter are more samples there to deal with even if it does.

As far as having something up at 80khz, most of us probably aren't recording anything up there. On the other hand, that doesn't mean that there isn't anything being picked up above 20khz or that some of us don't have extended freq mics.

Recording at 192k with standard mics might be a bit overkill, but I see nothing wrong at recording at 96k for clips that are going to get seriously mangled anyway. I certainly wouldn't record an hour long concert at 192 though...

Edit - I just noticed you said slowing down isn't done my messing with the sampling rate. Fancy software might do some tricks, but you can certainly slow down/lower the pitch of a sound file by lowering its playback sampling rate from the sampling rate it was originally recorded at. Soundhack can do it. I'm sure other programs can too.
 
Yup Petrus, kidtexas and exile are totally right - havent you ever had a system switch from 48k to 44.1 by mistake and the women suddenly sound like men? Try it on a dat sometime... Only difference is that with resampling for speed, you keep the pitch, drop the speed.

And yup, lots of progs do it - Nuendo ladies and gentlemen.
 
Totally agree with timapter on the 24/48. The extra bit depth is huge in recording and post but the 96k sample is totally unnecessary for dialog recording.

In the year or so I have been using the 744t I have become spoiled on 24bit. I absolutely can't go back to 16bit now. Honestly not even sure if the 16bit feature works on my recorder. :) I did some A/B comparisons the day it arrived but haven't used it since.

Storage wise 24/48 only requires a bit more storage then 16/48. Especially when compared to 24/96 (roughly 350MB per hour for 16/48, 520MB per hour for 24/48, and 950GB per hour for 24/96). I have two 2g CF cards and that gets me through at least one, usually 2, days of shooting at 24/48 using 3 channels. With 2g cards coming in under a hundred bucks, you can't go wrong.

Cheers
 
I got a Kingston 4 GB card from Dell called an "Elite Pro" which is a 45X card... drum roll please... for $190

This was 3 months ago. Now THAT is luxury! Next spring I may get an 8 GB card just to be obnoxious, but as it is the 671 is like a hard disk recorder for me! I don't use the 671 EVERY time I shoot... just for music and when a wireless would be the only other option... because of that I've still got some of the very first recordings I made on this CF card! I still haven't had to clear space yet.
 
I was going to say the same thing..

96Khz and 192Khz aren't pointless at all. For us sound effects people, they are very useful.

Sometimes it seems people don't even understand what is going with sample rates. A sample rate is the number of times a second a sound is sampled. The more samples you have the better, in my book. Now, can anyone hear the difference between dialogue recorded at 96 and 192? I would bet not. But when it comes to sound design where I start stretching things apart and pitch shifting stuff down, having the higher sample rates are crucial.

-Dave
 
Dave_Fisk said:
I was going to say the same thing..

96Khz and 192Khz aren't pointless at all. For us sound effects people, they are very useful.

Sometimes it seems people don't even understand what is going with sample rates. A sample rate is the number of times a second a sound is sampled. The more samples you have the better, in my book. Now, can anyone hear the difference between dialogue recorded at 96 and 192? I would bet not. But when it comes to sound design where I start stretching things apart and pitch shifting stuff down, having the higher sample rates are crucial.

-Dave
Click to expand...

Hey! I've never thought about that before. I always record 24/44.1 because then I don't have to dither the sample rate to CD. But time stretching! There has to be a clear advantage there with more samples per second. Any other scenarios where the higher sample rate would be a definite, real world advantage?
 
timapter said:
Hey dave, for pitch correction in voices, I doubt 96k is worth it - or do you think it is?
Click to expand...

For pitch correction, it's not all that crucial, but for pitch shifting in terms of sound design, it is. There are times when I pitch shift something down A LOT, and the higher sample rates will actually record things that you normally can't here, but can once you've pitched them down. Bats are a prime example.



To answer you question gibbie, any session where you plan on doing a lot of weird stuff to your sound, I recommend using higher sample rates. When people record out in the field, unless it's production dialogue, I ALWAYS recommend at least 96Khz. You never know what somethng may get used for. One time I recorded a metal rod on a big flat piece of iron for something I was working on. It ended up not working for what I had intended, but I ended up pitching shifting and time expanding that thing a TON to get another effect for something else. If I had recorded at 44.1, then it wouldn't have worked.

-Dave
 
You must log in or register to reply here.
Back
Top