96k is better than 44.1k sample rate. why?

[Reggie]

I do know one rock solid fact why 96khz is *sometimes* better than 44.1khz.

If you pitch a sound far down, in 44.1, you will have less high end than in 96. The air of the human voice is in the highest audible register. Lets say you are recording a scream, there's gonna be stuff in the 10 to 20k range. Now pitch this a full octave down. The 10 to 20k will become 5 to 10k, and because 44.1 has no information above (give or take a khz) 24khz, you have nothing at all about 12khz or so. On the other hand, if you recorded at 96khz, and pitch that down a full octave, you will still have stuff in the 10 to 20 khz range.

Interesting concept, but I would be willing to bet that his information in the upper bands does not have much power to it; and even when using a high sample rate and pitching it down, it will probably be nearly inaudible and definitely not the "meat" of the sound. The "meat" of the sound will just be moved down an octave, and you may have some extra noise--possibly unwanted--that doesn't really match with the tone of the original unpitchshifted sound because it was above the freq response of your ears to begin with. I guess if all you are trying to do is make some weird noises/sound FX, then the point could be to bring the inaudible upper freqs down to a hearable octave in order to obtain a new noise.

 

[giraffe]

I used to think converting audio to digital was as simple as plotting voltages over time and you get a kind of line graph representation of your audio. But after reading some of Dan Lavry's papers I have found that there is other math involved to render and reconstruct the waveforms. Sinc functions and decimation and all kinds of goodies. Not so simple math at all...
http://www.lavryengineering.com/documents/Sampling_Theory.pdf

thanks, will read later.

 

[CoolCat]

Sorry to be overly technical. Somewhere back in another thread I had done listening and analysis on 48 vs 44.1. As a result, there is no question in my mind that 96 is better than 44.1, so is 48 and 88.1 for that matter, as I can hear the difference, and analysis verifies that there is indeed a difference.

I love the technical stuff posts, the visuals are greatly needed in my case greatly appreciated. Explanation with the visuals makes much more sense to me. Without 'em I'm screwed!
From what I gather your saying 48 is great, 96 is about as good as it gets technically? this is great info. to combat the hyper-mode marketing sunami we face.

yeah, G...thats why I laughed at the "as I stare at my "state of the art" obsolete dongle comment...
hell yeah!! vinyl, cassete, 8 track , minidisc, reel to reel, ADAT, Digital Tape,CD..... fhk!!.... I was a proud owner of a state of the art, 8track RECORDER!! and I'm not talking an 8track multitracker, but the bulky funky 8-track technology (probably the pig of old state of the art technologys)

But thats were I really appreciate the new technology, my records crackled and popped and warped, the tapes would get dull and lifeless over time, or the heads worn out...remember rewinding reel to reels?, it takes alot of work in comparison to the new HDrive stuff.
Its the user ease...like going from a outhouse to indoor toilets,imo, frkn great...you can have the old problems...I'm not that sentimental to my gear.
Now the Tubes do have that crazy harmonics stuff human ears like though, but technically its a long cry from precision-duplication of a signal, as I understand it.

info. like MSHilarious shows there's still some improvement room here, between 44.1 and 96....sounds like 192 is going to be a hard sell.

 

[mshilarious]

I love the technical stuff posts, the visuals are greatly needed in my case greatly appreciated. Explanation with the visuals makes much more sense to me. Without 'em I'm screwed!
From what I gather your saying 48 is great, 96 is about as good as it gets technically? this is great info. to combat the hyper-mode marketing sunami we face.

Anything north of 48 is as good as it gets, 64, 88.1, 96. Above those rates there is the possibility that a reduction in precision could occur. That is one of Lavry's points; I don't know how to test it or exactly what to look for. And I don't have 192

 

[SouthSIDE Glen]

I was a proud owner of a state of the art, 8track RECORDER!! and I'm not talking an 8track multitracker, but the bulky funky 8-track technology (probably the pig of old state of the art technologys)

Let me guess...did you have the good ol Panasonic 8-track recorder with the side-by-side green VU meters? That's what I and many of my friends grew up on when we were kids, and when I first learned the concept of signal I/O with the Rec In and Play Out RCA jacks . I must have been about 7 or 8 years old at the time...

like MSHilarious shows there's still some improvement room here, between 44.1 and 96.

As much as I appreciate all the work and all the graphs and whatnot that ms is putting into that point (and I sincerely do, ms! ), the actual ROI at 96 just remains a seeming waste of resources for all but the most golden-eared audiophile of applications; and even then it's questionable in my mind. To more than double the bandwidth of a recording in order to pull an extra 10% of spectral range out of the top end, especially when

a) that extra top end is in a range that only golden eared audiophles and golden retreivers can hear,

b) that extra top end is in a range that is usually not very accurately handled at most points on the analog side of the signal chain anyway, and

c) even if and when it is audible and accurately reproducable, is not guaranted in the least to be "pleasant sounding" to the human ear,

then I have to remain dubious that it's actually a worthwhile technology worth even bickering about. The jump from 8-track to cassette was HUGE for minimal cost, as was the jump from 78 to 33 1/3rd in vinyl. However, this is more like the jump from CD to SACD, which has minimal benefit, especially for the cost (both monetary and technical) involved. That minimal benefit is reflected in both cases by the extremely lethargic response by industry and consumer to the SACD, and by the small percentage of penetration 96k has made into the recording studio even at the pro level.

Moving a microphone two inches releative to it's source will make a far, far larger difference in quality of the final production than doubling the sample rate ever will.

G.

 

[CoolCat]

but sometimes, thats what true tech-gearheads do is upgrade for tiny minimal improvements....overall they add up.
like a NASCAR Engine for an analogy.
to have the best technical piece money can buy, that 10% improvement here and there adds up, throughout the entire engine...piece by piece, the absolute best shit they can get their hands on...leading edge.
precison tuned. 96Khz...

from a financial standpoint, yeah, probably crazy...the customers don't know its there, the MP3 gang doesn't hear it thru their EarwaxBuds....especially if its a Home Recording room in a basemnt shared with a washer and dryer.

and it all goes down the drain if the race cardriver doesn't have any skills and runs your masterpiece car into a wall!!

i think I'll quite posting on this...I'm getting tired of myself...
I'm discussing "which is better"...the concept of "better"....as I can't keep up with the 44.1 and 96K stuff.

I'm still confused, but thats normal.

 

[bennychico11]

and what's the frequency range of the SM57 microphone?

 

[SouthSIDE Glen]

like a NASCAR Engine for an analogy.

True enough. When I look at the NASCAR analogy, this is how it fits to the sample rate debate in my mind.

For 99.9% of the people who would be even remotely interested in this board, being concerned about going to 96k is like being concerned about adjusting the angle on the real spoiler by half a turn, when the car one is driving has a stock 3.4 liter V6 in it and the most high-speed experience the driver has is driving down the interstate to visit Aunt Rose. Does the spoiler make a difference? OK sure, it changes the aerodynamics. Is it a difference that actually *matters* in the case of Team Rose's performance? No. Are there a million things that Team Aunt Rose needs to concern themselves with and spend their time considering long before they worry about geeky technical details like wing adjustment? Absolutely.

G.

 

[Reggie]

and what's the frequency range of the SM57 microphone?

Flat out to about 40K or so.....haha yeah right. I think it drops at around 18K (or maybe it was 16K?).

Anything north of 48 is as good as it gets, 64, 88.1, 96. Above those rates there is the possibility that a reduction in precision could occur.

Yeah dude, preach on. Once you get past the effects of the anti-aliasing filter (plenty of space to filter when at 96K), then the only areas to really improve a PCM digital convertor are with the clocking, the analog circuit, and other mechanical type bits. Although I suspect that at some point the notion of reduction in precision at the higher sample rates will become irrelevant as the smart dudes find ways to throw electrons around faster and faster. I guess audio recording marketing weasels gotta feed their families too! <sarcasm>I can't wait to hear the first album recorded at 384K; it is gonna be sweet! I think Korn had one of the first albums done entirely at 192K. Yeah, it was at least twice as good-sounding as their one before that. </sarcasm>

 

[bennychico11]

Flat out to about 40K or so.....haha yeah right. I think it drops at around 18K (or maybe it was 16K?).

by the time it reaches 20kHz it's already -10dB down.
one of the most popular mics, combined with the filter slope....is there any positive with recording at a higher sample rate? Do we only reserve the highest sample rates with the flatest condenser mics?

 

[SouthSIDE Glen]

by the time it reaches 20kHz it's already -10dB down.

Actually according to that chart it's even more severe than that; that response line never even reaches 20k, and is already 10dB down by the time it ends at around 16-17k.

G.

 

[Reggie]

by the time it reaches 20kHz it's already -10dB down.
one of the most popular mics, combined with the filter slope....is there any positive with recording at a higher sample rate? Do we only reserve the highest sample rates with the flatest condenser mics?

If EVERYTHING is recorded with a 57, then no I would have a hard time seeing any benefit recording at a higher sample rate. There is a little benefit to recording at higher sample rates if you record some of your tracks with mics that don't roll off the high end so much. Of course you don't want to set a new sample rate for each track you record depending on the mic being used, so pick one and stick with it. And as I have said elsewhere, until people start buying 96K CDs and 96K mp3s, I don't think there is a great big reason to worry too much about sample rates above 44.1K.

 

[TragikRemix]

I'm all for it, and that'll be the day I start tracking everything at 96K.

I'll third the notion....

that, or i start being able to make some decent recordings, or i win a give away powermac quad with a terabyte of HD space, and 4 gigs of memory.

 

[TragikRemix]

But all the distortion will be at frequencies above 22kHz. I can run a test to prove that. I can also crank the resolution of those graphs up as much as you like

can we see 192 vs 44.1? pretty please!

what do those pro studios do then?? do they record at 192k?? i friggin doubt it, they probably use 44.1 16bit and they just dont tell anyone!!

well, ok so they probably do 24 bit but thats a WHOLE separate issue, which i think everyone agrees there is a large difference between.

why do they bother with 192 interfaces, at 2x the price of 96k? this on a pro level, i mean, like the crazy hollywood/tokyo/european whatever studios.

at our level (most of us) 44.1 is sufficient... if you have more availible, go ahead, you can use it by all means, but its not absolutley necessary.

 

[TragikRemix]

and im jokin, dont worry about the graph MSH..

 

[Reggie]

why do they bother with 192 interfaces, at 2x the price of 96k? this on a pro level, i mean, like the crazy hollywood/tokyo/european whatever studios.

Because not every crazy pro studio has an expert in digital sampling theory. All they see is 192K is bigger than 96K and therefore better. They feel the need to stay on top of the curve, and so they buy this newer better thing whether they really need it or not. Interface makers make them in the first place to make money; to keep people buying things. About the only manufacturer as far as I know that hasn't pushed the hype is Lavry. And I think Apogee mentioned something one time about their convertors sounding the same at 96K and 192K or something like that, but don't quote me on it.

 

[AnotherJazzMan]

I see it both ways...

I had direct experience with digital to analog comparison at a local pro audio shop. Amplifier, electro-statics and connections were kept the same. The same recording was rendered on both CD (AudioQuest) and LP (Rega turntable).

The areas where I heard the biggest differences between the digital audio and analog audio was in percussive and other impact input (rim shots, cymbal crashes, etc.). I strongly suspect this is because this input has more energy in the higher frequencies (and the harmonics present above 20KHz).

Most musicians have been introduced to the concept of timbre....which is notably affected by harmonic content, both below and above the audible limit of 20KHz.

The other factor influencing what we hear is the relative power contributed by each harmonic. As frequency increases, the amount of power contributed by a particular frequency tends to decrease.

To me, this means that it is not at all clear to what extent harmonics above 20KHz, even if represented in the audio signal, will contribute to what I actually hear.

I expect 96KHz sampling to represent percussive inputs better than 44KHz. Why? Of musical sounds, percussive inputs are the closest thing I can think of to 'impulses'. An impulse is a wave that looks like a spike - very high amplitude, very short duration - which means that it is rich in high frequency content.

At the same time, I expect to hear little difference between 44KHz and 96KHz when the input is the human voice, a stringed instrument or a wind instrument. That would tend to support much of what others have posted here...not much difference noted.

For me, I have the disk space and processing horsepower to manage 24/96KHz. I do so because:
1) it is the industry standard for archiving and high-def audio
2) I can always cut it down if I need to

The problem with recording in 16/44KHz is that you can't create the wider bit range (reducing error) or the higher sample rate IF you decide you might care.

On the other hand, you spend a more processing time and disk space for a result that in many cases, yields little if any appreciable difference.

Depends on what you value, IMHO

AJ

 

[SouthSIDE Glen]

I expect 96KHz sampling to represent percussive inputs better than 44KHz. Why? Of musical sounds, percussive inputs are the closest thing I can think of to 'impulses'.

A nice concept, but look at the actual math. Percussive strikes may be the closest thing we care about to instant impulses, but they do not appear so "impulsive" at the resolution of digital sampling. At 44.1k, an impulse would have to be shorter than 1/22nd of one millisecond for it not to be accurately reproduced mathematically (well, maybe 1/20th or 1/19th of one millisecond after anti-aliasing LP filtering). That's an extremely short period of time in the macro physical world, and is more than short enough of a sample gap to handle any percussive attack slope I can think of.

It's been said here before, but since this is a three year old thread that's been resurrected, it's worth saying again: if one hears something sound noticeably better at 88.2k, or 96k, instead of 44.1k, it is almost certainly because that particular make and model of converter sounds better operating at that frequency for circuit design reasons other than the effect of the actual sample rate itself.

If one has a converter that not only sounds better at a higher sample rate, but contributes more to the sound than the sonic cost of running through SRC to bring it back to 44.1 for distribution, and they have the computer capacity and horsepower to deal with it, then by all means use the higher sample rate.

But that does not necessarily extrapolate to generally saying that higher sample rates are better. It only means that particular device operates better that way, because it's the device that's making the difference, not the sample rate itself.

A Ferarri actually drives better at 100mph than it does in downtown city traffic. That doesn't mean that 100mph is a better speed. And it certainly doesn't ring as true for a Prius as it does for a Ferarri.

G.

 

[recordist87]

44.1 is all I've used...sounds fine to me