ORIGINAL: Scott Reams

ORIGINAL: jweldinger

Digital recording doesn't just filter out everything above ½ the sample rate. The audio wave of the original signal is altered. Instead of the complete waveform being recorded, a "sample" of the waveform is recorded at specific time intervals. It's not complete. That's why it sounds digital. Now if the sample were to be non-stop, it would be a different story. To make an analogy, it's like taking a 100 mile trip and filming at every mile marker as opposed to filming the trip non-stop.

Sure... but if you were to take a picture say... 100 times per second... would it seem "digital"? Of course not... even though it is still discrete frames. There is a point at which the resolution is high enough that we can't detect the seperation. One still shot every mile is -not- a fair comparison. We are on the border of what is perceivable here. The question is... at what point can no one tell anymore? That point -does- exist. The debate comes in when we start discussing -where- it exists.

-S

Good point. I was just making a general analogy. Your scenario is closer to reality.

ORIGINAL: rickgn

This whole debate about whether or not you can hear 192K or no it is getting kind of silly. NO! You cannot hear a SINGLE SINE WAVE at 192K. But consider this for a moment:

What makes music music? Is it a single sine wave?
What makes Tambre? Is it a single sine wave?
Does a 10 piece drum set send out a single 22Khz sine wave?
Does a guitar send out a single sine wave?

What happens when two instruments are not "clock locked" at 44.1 KHz and they are actually a few milliseconds off from each other? This is something you can only experience from listening to something live. 44.1 cannot capture the small nuances that occur when you start blending together multiple instruments.

When you blend together several signals you get harmonics. Not just upper harmonics, but lower ones as well. If you send a 33 KHz signal through an O-Scope or spectum analizer and then start adding other frequencies to it as well, you will eventually start seeing lower harmonics in the human audible range. Sure, they may be sitting at -30dB range but they are still there and they can sometimes make a difference in how something sounds. Even two identical signals slighly out of phase with each other will produce lower harmonics when blended together. 44.1 recording would simply filter this nuance out and it's lost whereas 96k or 192k can capture more of the natural sound and reproduce it more accurately.

So when you all are debating this, do a real test and listen to an entire performance recorded across multiple tracks at 192K. When the signals are combined it sounds more life like simply because there is more detail available. Try it for yourself and you be the judge. And no, your PC speakers or cheapo headphones will not help you hear the difference.

Well said. :)

ORIGINAL: 6stringsat100mph

Jweldinger,
While I would never claim to be as learned as you are and as most that have replied to this thread are, the sad fact is you are absolutely wrong. This argument may have been something to stand on 5 or 10 years ago but what is the trend IS the fact that Tape machines have run their coarse. It is sad I agree, but did you know there is ONE company left still making comercial 2" tape machines here in the states? The reason that everyone says "No, but we have Pro Tools" is because that is the industry standard. Tape recording is gasping on its last breath and mostly in wealthy pro musicians home studios. Not in the ones recording the big names for the record companies.
I repeat, I am no expert, but I read a bunch and have a subscription to recording mag and always read it cover to cover. This issue is brought up over and over again and the end result is the same. Get used to Digital recording because within 5 years you wont be able to purchase a 2" tape machine anywhere. Except used. And as I figure, very cheap.
Regards,
Mark

Mark,

Hey, I'm just a n00b when it comes to this digital stuff.

I thought Ampex was the last company to manufacture 2" tape, and that they had even stopped producing it. I do know that 2" tape is in high demand and can extract high prices, if you can even find it to by. It seems it is being horded. I wonder why that is...

I have 3 reels. Maybe I should auction it off on eBay.

I was completely turned off by digital when I heard my first MP3s, when that technology came out. A friend of mine turned me onto the technology and how great it was - the whole Napster to CD experience. I agreed with him until I played one of his 128kbps MP3 converted to CD Audio CDs in my home stereo (PA system). The bass was horribly distorted in the subs, and the highs were gone from the horns. The mids sounded canned. They might sound fine in a car CD player but not in a pro audio setup. Anyway, once I started registering copyrights with the LOC, I took a whole knew look at infringement.

I know that from my commercial CD collection, the best sounding CDs are AAD. Albums like Deep Purple/Machine Head, Black Sabbath/Black Sabbath, Pink Floyd/Dark Side Of The Moon, etc. I know that these have been rendered down to 16-bit, 44.1kHz to get them onto CD, but the 24+ Track Masters and 2-Track Masters were on Tape. IMHO, either the engineers of today aren't as capable as the engineers from 30 years ago, or the process used today (DDD) is not up to par with the process from that time. This is not intended to be a slam on anyone here or elsewhere. All I know is what I hear. Even at that CD Audio in no way compares to vinyl. When CDs first came out I thought I'd never buy one because they sounded artificial. I think digital is more convenient and easier to produce for the masses, and that is why it is being successful - not because it is superior.

Now, In the last week I've actually done a couple of recordings at 96 and 192kHz and they sound life like. 44 and 48kHz are unacceptable but as far as recording at 192kHz, I'm a believer.

Now having said all of that when I return to the studio the project will be mastered AAD, until DDD either surpases AAD in what I can hear, or until analog recording is no longer possible.

A test to eliminate the speakers from the equation would be to sample the voltage and compare with various outputs from hardware. If the voltage is not showing ultrasonic frequencies, we can stop there.

Well that would be a factor of the quality of the converter and the filters not the sample rate. Jitter in---- Jitter out sounds like sh** even with the best clock on the DA.
I would make a bet that using quality AD / DA with an ultra low jitter clock with high quality filters at 44.1 would be indistingishable from 96khz sample......

ORIGINAL: sfbassplayer

ORIGINAL: jweldinger

professional recordings are done on tape.

Sorry, but I feel such a broad generalization doesn't really qualify for much ... having a great engineer, creative producer, and great outboard gear/mics have ~huge~ impact as well ... and can minimize whatever people perceive is being lost by not recording on tape.

Plus, since since the last remaining manufacturer of professional analog recording tape has shuttered it's doors, you'll need to come over to the dark side eventually.

Hehe. I'm slowly coming to that realization.

Also, as far as tape goes, the reason the big boys tend to still master to tape is that they are suspicious of the compatibility and logevity of digital formats, yet tape stands the test of time. Masters that are 50 years old still sound great. Important recordings still go to tape. The tape market has narrowed, and looks a bit uncertain with Quantegy, but BASF is still marketing in Europe, and IMO, there will always be a fringe-market available. And the important recordings will still go on tape.

ORIGINAL: SteveJL

Also, as far as tape goes, the reason the big boys tend to still master to tape is that they are suspicious of the compatibility and logevity of digital formats, yet tape stands the test of time. Masters that are 50 years old still sound great.

As long as they are stored properly, which can cost $$$ as a long term solution for irreplaceable masters, and even then, tape has a life span.

A friend of mine is constantly dealing with getting master tapes baked and restored so they can be converted to digital for use (like remixing or remastering) since tape has a shelf life. Many of the masters they'd like to use end up being unrecoverable ... in fact, one of them turned out to be the Average White Band's early albums which is only 40 years old at this point.

ORIGINAL: colo

Actually, I remember watching TV one night (PBS or CNN or whatever) and there was a special about the Library of Congress transferring all of their stuff to digital, like the first ever recording, the first ever this, the first EVER that, etc. The analog stuff was literaly disintegrating it was so old.

I believe I saw that program as well... the link and article I posted above is somewhat dated now.

ORIGINAL: jweldinger

Record a performance onto 2" Tape. Record the same performance, sampled at 16-bit/44.1kHz. You'll hear the difference.

The bottom line is that for a professional recording, I would settle for nothing less than a 24-track master on 2" tape. The master will be as true of a representation of the original performance as possible. I can then process that however I like, e.g. 2-track master on ADAT @ 96kHz, even down to 44.1kHz for CD Audio, but I will still have the 2" master for reference.

A comparison between analogue tape and uncompressed digtal is a comparison of apples and oranges. The digital recording will actually be more accurate in terms of what's heard by the human ear due to analogue recordings receiving an average 6dB of compression during the recording process. It's the same as with tube microphones--they sound great and really flatter the sounds recorded with them, but the sounds are being altered through technology. If a compressor is used during the digital tracking process, digital recordings can sound much like analogue tape.

It's not clear that the speakers can or cannot provide a sustained ultra- or infra- sonic output: do manufacturers bother with test equipment that's sensitive to ultrasonics?

Yes they do. That's why many show a clear rolloff above 20KHz in frequency response charts. It's a measurement, not a guess.

In fact, there are a few audio devices out there that are measured to have flat response well beyond 100KHz... although none of the devices that can do this are speakers.

-S

ORIGINAL: m11

jweldinger,

you are telling us, you're listening to music with PA speakers?
The PA speakers shown at your avatar?
PA speakers that are made to be loud but not to be precise?
With woofers that are too big and too hard to produce decent bass at lower volumes?
You're sitting on that chair? With your ears at the same level with that horns?
Horns that are made to amplify sound, but also bend the sound waves and distort them?
You want to tell us you hear the difference between 44.1kHz, 96kHz or 192Khz, or frequencies above 20kHz with that speakers?
ROTFLMAO!!!

Either you're listening at volumes where these speakers can't produce a half decent frequency spectrum, or you must be near to deaf.

And you tell us pro recordings are made with 2" analog tape, and in another post you tell us, that tape isn't avaliable anymore? ROFTL

And you are the one, who wants to make us believe that people can hear frequncies beyond 30kHz? ROFTL

And you have no answer to my question wether your wife hears a resonance that is lower to that 31kHz? How did you measure that 31kHz? Did you ever know that there are always resonances at 1/2, 1/3, 1/4 and so on frequencies?

And you tell us that recordings from the 70's sound better? You mean you like it better! Are you sure your personal taste is the best? Are you sure you can tell the difference between digital recording and overcompression?

I was very amused by your contentions that sadly too often where wrong.
I'm very sorry, but I really can't take you serious anymore!

Thank you for reading this.
Melf


btw. there are differences between the sample rates. But what you probably hear are aliases at lower than the half sampling frequency.

There are poles bewteen the subs and mains now. I'll have to update the avatar. The mains are above my head. The horns were blowing my ears out. It's a 300W/Channel Power Amp, and I run the system about 1/2 way up. It's crystal clear and it sounds crisp and solid. Everything is mixed externally and processed through outboard EQ, FX, Compressor, Limiter, Gate, and Sonic Maximizer. Passive Crossovers are in each main.

2" Tape is no longer in production, or at least only manufactured sparingly. Existing stock is being horded. Someone is using it, and it's not home or mom&pop studios.

I can hear above 15.5kHz. I can't hear the squeal from the monitor or TV. I have no way to verify that it's 31kHz, but that is the RF Stage frequency. Regardless if it's 31kHz or not, she's hearing something high pitched, beyond "normal" hearing range. No one else in the family can hear it.

My taste in music genres has no bearing on why I think that AAD CDs sound better than DDD CDs. You want to hear a killer mix? Listen to Steelheart/Steelheart. The music is nothing spectacular, except for maybe I'll Never Let You Go. But the recording and mastering are top notch...another AAD CD.

BTW, I know how signal frequencies are manipulated. I have a degree in Electronic Engineering and I wrote a 190 page thesis on Sound Synthesis in 1982, if I remember correctly. Maybe I can dig it up.

I have to be at a gig in 2 hours, so have fun...

ORIGINAL: John Page

A test to eliminate the speakers from the equation would be to sample the voltage and compare with various outputs from hardware. If the voltage is not showing ultrasonic frequencies, we can stop there.

Well that would be a factor of the quality of the converter and the filters not the sample rate. Jitter in---- Jitter out sounds like sh** even with the best clock on the DA.
I would make a bet that using quality AD / DA with an ultra low jitter clock with high quality filters at 44.1 would be indistingishable from 96khz sample......

Re: converters: the closer you can get to the ideal reconstruction filter, the more accurate the sound (the closer you can get to "perfect reconstruction" from theory). There are many stages to look at to determine why and if higher sampling can sound better. Putting the voltage on a scope and using measuring equipment that can handle ultrasonics from the speakers would allow one to determine how the DAC's and speakers change behavior at various playback rates. These measurements combined with listener tests would help to understand what is going on at higher playback frequencies. A published test of this nature would be an interesting read...

The point was: is the rolloff a result of the speaker output falling off, the measurement equipment falling off, or a combination of the two (in the ultrasonic range)?

I sincerely doubt it is a matter of the measurement equipment rolling off. If it were... the results would be horribly inaccurate... and not viable for release to the public. There are standards that all of these companies must adhere to.

-S

ORIGINAL: Scott Reams

The point was: is the rolloff a result of the speaker output falling off, the measurement equipment falling off, or a combination of the two (in the ultrasonic range)?

I sincerely doubt it is a matter of the measurement equipment rolling off. If it were... the results would be horribly inaccurate... and not viable for release to the public. There are standards that all of these companies must adhere to.

-S

A good speaker is expected to have linear response from 20Hz-20kHz (industry standard). Speaker graphs typically stop at some point over 20kHz. I would want to see the frequency response characterstics of the measurement device over 20kHz to have confidence regarding ultrasonic output (which has not typically been important considering the common belief that it's not possible to hear in that range).

In the ultrasonic range, I would trust the measurements from a Laser-based device over a mic-based device. For example, Laser Doppler Vibrometry (LDV) [ http://www.compliantlab.sdsmt.edu/laser.html ]. Visteon (Ford) uses lasers in speaker design (not clear WRT testing): http://www.visteon.com/products/capabilities/audiosys_testing.shtml .

Hearing tests use simple sine waves

I have taken a standardized hearing test, and the test sounds were presented as simple sine wave tones. If this is where the 20-20kHz range comes from, the test is flawed. Clearly, if anyone can hear differences in blind tests with playback rates over 20kHz, there is more to hearing/perceiving sound than a simple sine wave test can show. If not already published, it would be to the hardware maker's advantage to perform independent blind tests showing that people can perceive improvements in quality at higher playback rates (using more complex sounds and music passages).

Speakers can produce ultrasonic energy
Timing+spatialization changes can be perceived

Given that some speakers can produce sound energy above 20kHz (not clear where the output energy goes to zero due to graph clipping), timing+spatialization improvements at higher rates (and perhaps other unknown effects), and that some listeners can hear the difference in playback quality at higher rates, producing music/sound products at higher rates is worthwhile if one desires the best quality possible. In the case of target media at 44.1/16 and/or MP3, it probably won't make much difference (as the listening audience may not notice any improvements. Such improvements would not be related to ultrasonics or timing issues, only final mix quality).

ORIGINAL: Scott Reams

I would want to see the frequency response characterstics of the measurement device over 20kHz to have confidence regarding ultrasonic output

If the test equipment had a rolloff, it would either be compensated for, or the results would not be posted. All of this stuff is considered when charting frequency response... and response charts -never- show characteristics of test equipment. Doing such a thing would be irresponsible at the very least.

-S

It would be a good idea to understand the characteristics of the test equipment as well as how far the DUT outputs ultrasonic energy given the controversy of the subject matter. I would test every possible aspect of the process and make no assumptions.