Wanting 192hz...what's best

Howdie...I checked out this debate last year. Thought I would check again now. I'm running Tascam 1884 with my Sonar 5. I want to move up to a 192hz rig. Thinking of
Onyx 400f......whats the consensus......is there something better? Any big problems with running it with Sonar.............Any feedback is appreciated.........

Why, are you recording a symphony or orchrestra. Its not needed and its a waste of disk space. 96khz is kinda overkill, unless im recording an acoustic guitar, but 48khz is my choice
Cj

ORIGINAL: CJaysMusic

Why, are you recording a symphony or orchrestra. Its not needed and its a waste of disk space. 96khz is kinda overkill, unless im recording an acoustic guitar, but 48khz is my choice
Cj

CJ is right. Actually I do record symphony orchestras, both for CD and FM radio broadcast - I've never been above 96Khz- no reason to be, and most of the time 48Khz has been perfectly fine.

I have an SACD organ quadraphonic release coming out in the spring - we did 24/96khz - captured on my Alesis HD24 all post done in Sonar - compared to my 48Khz stuff monitored on seven speaker systems - I doubt that most would hear any difference..

-D

Come on guys! The man asked a simple question! Don't be such besserwissers, I'm sure he's got his reasons for wanting 192 KHz. I mean, some people buy speaker cables for $10 000 a foot only so they can sleep well at night knowing that they have the Rolls Royce of speaker cables!

Sven

Uh guys, didnt he say 192hz - not 192 Khz? Maybe he is trying to get that downsampled sound.

ORIGINAL: time2go

can you even hear the benefit of the increased dynamic range on your speakers?

Phil

At the risk of being a besserwisser myself, sample rate does not describe dynamic range, bit depth does!

Seriously, I agree that all of this is sensible advice, but they are answers for a different question ("Is recording at 192 KHz going to be worthwhile for me?").

Sven

that brings up the interesting point of using semipro equipment at extremely high sample rates. odds are there is probably nothing to gain without super-ninja mics, pres & other equipment. then you need a pretty impressive amp & monitor system to hear if there even is a difference at all.

i would think there are a lot more downsides that upsides in that scenario because of increased disk usage and CPU load.

People need to understand the difference between dynamic range and frequency response.

The number of bits determines the dynamic range.

Sample rate determines frequency response. There is something called the Nyquist frequency .

Anyway, in a nutshell, the maximum frequency that can be reproduced from a digital recording (or any other sampled data) is half the samplinig frequency. Consider a square wave. If you sampled once per cycle, you would get DC. The greater number of times you sample, the more likely you are going to be able to detect the change from -5v to +5v or whatever the range of the square is. But if you sample 2x per cycle, you are really not going to be able to tell the difference between whether it is a square or a sine. That is why you need high sample rates to determine the harmonic content and accurately capture the shape of the wave.

Works the other way too. A 44.1k sample rate could reconstruct sine waves at 22.05k - pretty much at the edge of human hearing. You need analog filters to get rid of signals above that. They can induce phase shifts and ringing.

96k sample rate would go to 48k replay bandwidth. I am not sure of any speakers or power amps designed for that duty. I am pretty certain our hearing doesn't go that high. But for something with a really sharp edge (like a rimshot on a snare), a high sample rate should allow the phase relationships of the sine waves that make up that impulse be consistent enough that you dont have a distorted slope to the sound or a ring. I think at 96k, you dont need analog filters, you can use DSP to emulate a linear phase filter.

192k??? If you got rid of filters and the phase shifts they induced by going to 96k, what do you gain? I know that while some humans hear above 20khs, I dont think dogs hear at 96khz - which is the upper frequency of a 192k sample rate.

Ultimately, if you are printing to CD, you are going to be reduced to 44.1k and 16 bits. I say sample at 44.1 or 88.2 and use 24 bits, with those 8LSB being the detail that makes the difference when the signals are being mixed digitally so that what is reproduced when it is being converted to 44.1/16 is better.

What other posters said about mics, pres, speakers, mixers.... all that is true. If you want quality 192k, it is going to take some money.

Just my ¥2.3

WHOAAAA muddy waters here..........thanks everyone. Well i was thinking of it because I do record a lot of acoustic instruments. Most my music is wrapped around a 60 year old Martin O-17 guitar....(very sweet sound)......and lots of vocals.........sort of a country/rock/folk/Irish/ from the planet Pluto sound..........I use synths but lightly................I'm always looking for more depth and presence in the sound.........by the way my monitors are Beheringer Truth's..........the computer is an AMD 64 duel core 4600

ORIGINAL: space_cowboy

People need to understand the difference between dynamic range and frequency response.

The number of bits determines the dynamic range.

Sample rate determines frequency response. There is something called the Nyquist frequency .

Anyway, in a nutshell, the maximum frequency that can be reproduced from a digital recording (or any other sampled data) is half the samplinig frequency.

Indeed.

Consider a square wave. If you sampled once per cycle, you would get DC. The greater number of times you sample, the more likely you are going to be able to detect the change from -5v to +5v or whatever the range of the square is.

This is a bit misleading for two reasons: 1) there are no real square waves in nature 2) You only need to capture the harmonics that humans can actually hear. That is, below 20Khz or so. Harmonics of square waves are sine waves so it is much easier to just look at sine waves when considering sampling.

But if you sample 2x per cycle, you are really not going to be able to tell the difference between whether it is a square or a sine.

You don't have to. The harmonics of a square wave are at higher frequencies. So if you have a 18Khz square wave, only the fundamental is in the audible range and that is the only thing you need to consider. If you take a lower frequency square wave, then some of the harmonics will be in the audible range but they will also be in the nyquist range of any converters running at 44.1Khz or above.

So in short, look at a square wave as the summing of sine waves at different frequencies and all becomes clear. :)

That is why you need high sample rates to determine the harmonic content and accurately capture the shape of the wave.

Not true. You only need to cover the human audible range. Anything above is not worth capturing.

Works the other way too. A 44.1k sample rate could reconstruct sine waves at 22.05k - pretty much at the edge of human hearing. You need analog filters to get rid of signals above that. They can induce phase shifts and ringing.

Note that in all modern converters, the sampling happens at 64 or 128 times the base rate so the analogue filters only need to to remove frequencies at 0.5 * 64 Fs or 0.5 * 128 Fs. For 44.1Khz that is 1.4 Mhz and 2.8 Mhz respectively. It is not very hard to create analogue filters for those frequencies that have no audible artefacts.

So the issue becomes the digital filters. Some converter manufacturers do a better job than others... But 96Khz is already overkill for this purpose. 192Khz (and above) is just a marketing scam.

But for something with a really sharp edge (like a rimshot on a snare), a high sample rate should allow the phase relationships of the sine waves that make up that impulse be consistent enough that you dont have a distorted slope to the sound or a ring.

This can also be achieved at 44.1Khz for audible frequencies. It is harder and not all manufacturers manage (for whatever reason) but it is possible.

I think at 96k, you dont need analog filters, you can use DSP to emulate a linear phase filter.

You always need analogue filters really.

Ultimately, if you are printing to CD, you are going to be reduced to 44.1k and 16 bits. I say sample at 44.1 or 88.2 and use 24 bits,

Agreed.


UnderTow

I agree with everything undertow says......I tought him everything he knowns by the way ......Naaah, I wish . It's always educational with him, take care dude!

Lots of advice here, ranging from very good to flat out wrong. I won't compound your confusion by adding my own dissertation on the subject and will instead keep my advice simple:

Given that you're doing folk-ish acoustic stuff, I'd recommend that you put your money into microphones and don't record higher than 88.2KHz.

That's a 16-bit 44.1KHz wave file you hear when the SONAR installation disk starts up.

If it sounds good to you, that's because it was well-recorded, irrespective of sample rates. Somebody at CW would have to give the definitive answer, but I'd be very surprised if it was recorded at 192KHz.

My advice to a color timer turned sound recordist:

The only way to capture the subtle emotion of acoustic music is to track in 24P. :-)

Leave the 60i to the radio stars.

best,
mike

well, I'll have to admit I'm more partial to 30P. But, That's cause we never shoot video for film out. We shoot film for film out.
30P is a great way to go for both Broadcast and DVD distribution. I also like working with Varicams and creative "crank" sensibilities... but we conform to 30P in post if it's for TV.

but now I'm off topic.

Oh yeah +1 for good mics, preamps, and technique.

mike


edit for spelling and clarity