192 vs 96 quality dithering to 44.1?

Posted: Fri Jan 20, 2006 1:06 pm

Cucco,

I think it's not fair to say Tweakheadz is bullshitting. He actually said that there will be audible difference if you are recording acoustic instruments, orchestras etc... because quiet passages will benefit from the lowered noise floor of 24-bit sampling. And this makes perfect sense, I'm sure you'll agree. Whether or not everyone can hear the difference is another issue, although he did say that benefits will be less audible on most radio-ready music.

He has also said that choosing bit/sampling rates is a matter of opinion, and that his should not be treated as the gospel truth.

I did not find any misinformation in his guide, I thought it was quite well-written as well.

Posted: Fri Jan 20, 2006 2:19 pm

godchuanz wrote:

Cucco,

I think it's not fair to say Tweakheadz is bullshitting. He actually said that there will be audible difference if you are recording acoustic instruments, orchestras etc... because quiet passages will benefit from the lowered noise floor of 24-bit sampling. And this makes perfect sense, I'm sure you'll agree. Whether or not everyone can hear the difference is another issue, although he did say that benefits will be less audible on most radio-ready music.

He has also said that choosing bit/sampling rates is a matter of opinion, and that his should not be treated as the gospel truth.

I did not find any misinformation in his guide, I thought it was quite well-written as well.

Granted, I might be being a bit tough on Tweak, but there is misinformation in his article. Detailed below (Items in quote blocks are quoted directly from Tweak himself)

Quote:

Lets talk about sample rate and the Nyquist Theory. This theory is that the actual upper threshold of a piece of digital audio will top out at half the sample rate.

Okay, more of a nit-pick than anything else, but this is not the Nyquist Theory

Quote:

The theory is that audio energy, even though we don't hear it, exists as has an effect on the lower frequencies we do hear.

Again a nit-pick, but this makes it sound as though he's referring to the Nyquist Theory, which is not true. This is poor wording on his part.

Quote:

Back to the Nyquist theory, a 96khz sample rate will translate into potential audio output at 48khz

Nit picking - should be "<48kHz"

Quote:

Quiet passages will be less likely struggling to stay above the noise floor on your system.

Really?? -96dBFS noise floor - having trouble staying out of this? I know of maybe 3 locations on this continent that are capable of this kind of signal to noise ratio in an open environment - furthermore, mics with any kind of gain applied to them will automatically put you above this level. Again, nitpicking...

Quote:

The Bass will be tighter, and the vocals may sound "airier".

Vocals "airier" - I can buy that.
Bass tighter??? How? The fundamental and then next dozen or so harmonics aren't even in that frequency range. Those that might be in that frequency range are so minimal in amplitude that they should have no impact on the intensity or fullness of the fundamental.

Quote:

Recording at 24/96 yields greatly increased audio resolution-over 250 times that at 16/44.1

The choice of the word "Resolution" is poor. The word should be data. Period. The extra "resolution" does not make "cleaner" or "smoother" audio, it simply gives you more headroom (which Tweak does mention.)

In fairness, Tweak also mentions that talent is the key and that an experienced person with 16/44.1 would likely make a better recording than an inexperienced person with 24/96.

So, okay, I'll lay off of tweak, but there is a group of slanted statements here.

J.
[/quote]

Posted: Fri Jan 20, 2006 3:06 pm

[quote="Cucco"]

covenant66 wrote:

Rather than trusting Mr. Tweak, try chatting it up with Dan Lavry or Nika Aldrich or even Dan Weiss - these guys visit here often and are available at other forums too.

Well, guess what, I did a little bit of digging and guess what I found! I found a Nika Aldrich article lauding the UV22 as a process that cannot even be compared to the POW-R.

You can find it here:
http://www.cadenzarecording.com/Dither.html (Nika Aldrich's website)

This is your own quoted and trusted source that YOU gave.

Also you can check out Richard Ellen's article which states:

According to a paper by dCS Ltd, presented at
the 20th Tonmeistertagung in 1998, the most
commonly-noticed benefits of recording at 96kHz
sampling over 44.1kHz include: less ‘busy signal’
breakup — very good quality; better separation of
reverb and room acoustics from instrument output;
better balanced bass; better percussion (particularly
cymbals); and some stereo image formation.

it is available at: http://www.ambisonic.net/pdf/hiresaudio.pdf

Please also see this article

(http://www.dolphinmusic.co.uk/page/shop/news_story/a/news_id/e/120) which states that:

"Therefore 96KHz refers to 96000 slices of audio sampled each second."

It would seem logical that if you had more slices of audio, it would sound better across the board in all frequency ranges, not just add frequencies 22khz-48khz.

Why are you thinking that the Nyquist Theory is the only thing that applies when recording at 96khz. Do you have any sources?

Can you quote these sources or are these your own ideas? Quote that book directly, don't just tell me to go read it.

By the way, I think this is working out great. If I could have read this post a few years ago it would have seriously enlightened me and I think we are getting close to the bottom of this.

Posted: Fri Jan 20, 2006 3:31 pm

Crap, the quote on the last post didnt work so well. That is Cucco's quote at the top, not mine!!

Posted: Fri Jan 20, 2006 5:20 pm

You know the one thing most people here have not mentioned is that the most deleterious effects at 44.1kHz 16 bit recording, is the heavy brick wall filtering necessary. I believe that is the primary reason why most people don't like the marginal capabilities of 16-bit 44.1kHz recording.

One of the reasons the higher resolutions recordings do sound slightly better is the less needed brick wall filtering and the appropriately higher frequency at which it needs to happen. Nobody likes the sound of these filters but it is sort of like death and taxes, when dealing with PCM.

Few of our esoteric "exampliers" have expounded on the benefits of DSD whose sound appears to be our closest analogy to analog (of course in the listening tests they did not give us any examples of "saturation" which is still a nasty third harmonic clipped waveform). Such is life. Unfortunately as of this date, it's still too much money except for the very affluent. (hey why can't they double that 2.53MHz sampling frequency anyhow?) Just kidding.

Poor old
Ms. Remy Ann David

Posted: Fri Jan 20, 2006 6:32 pm

[quote="covenant66"]

Cucco wrote:

covenant66 wrote:

Rather than trusting Mr. Tweak, try chatting it up with Dan Lavry or Nika Aldrich or even Dan Weiss - these guys visit here often and are available at other forums too.

Well, guess what, I did a little bit of digging and guess what I found! I found a Nika Aldrich article lauding the UV22 as a process that cannot even be compared to the POW-R.

You can find it here:
http://www.cadenzarecording.com/Dither.html (Nika Aldrich's website)

This is your own quoted and trusted source that YOU gave.

Also you can check out Richard Ellen's article which states:

According to a paper by dCS Ltd, presented at
the 20th Tonmeistertagung in 1998, the most
commonly-noticed benefits of recording at 96kHz
sampling over 44.1kHz include: less ‘busy signal’
breakup — very good quality; better separation of
reverb and room acoustics from instrument output;
better balanced bass; better percussion (particularly
cymbals); and some stereo image formation.

it is available at: http://www.ambisonic.net/pdf/hiresaudio.pdf

Please also see this article

(http://www.dolphinmusic.co.uk/page/shop/news_story/a/news_id/e/120) which states that:

"Therefore 96KHz refers to 96000 slices of audio sampled each second."

It would seem logical that if you had more slices of audio, it would sound better across the board in all frequency ranges, not just add frequencies 22khz-48khz.

Why are you thinking that the Nyquist Theory is the only thing that applies when recording at 96khz. Do you have any sources?

Can you quote these sources or are these your own ideas? Quote that book directly, don't just tell me to go read it.

By the way, I think this is working out great. If I could have read this post a few years ago it would have seriously enlightened me and I think we are getting close to the bottom of this.

Jumping Jesus on a Pogo Stick!!! Exclamation

Okay - first, Nika's opinion on UV22 is just that, an opinion. And, if you'll notice before, I stated I don't agree with much of Nika's subjective conclusions; however, he is nonetheless a smart dude who put together what some are referring to as the Bible of Digital!

Second - YES I am stating that this entirely has to do with Nyquist. Period. I don't care how many slices of audio you have - you only need two and only get two to draw an accurate picture of a wave, adding 95,998 extra pictures won't get you a more accurate picture of a 20Hz wave than 2 pictures will.

Here's some links to reading since you require them.
http://lavryengineering.com/forum_images/Sampling_Theory.pdf
In which Dan States:

Dan Lavry wrote:

You DO NOT need more dots. There is NO ADDITIONAL INFORMATION in higher sampling
rates. As pointed out by the VERY FUNDUMENTAL Nyquist theory, we need to sample at
above twice the audio bandwidth to contain ALL the information.

BTW...the paper is an interesting read, though there are those that again disagree with the subjective. However, I have yet to hear of ANYONE disagreeing with the objective portion.

Dan Lavry wrote:

The optimal sample rate should be largely based on the required signal bandwidth. Audio
industry salesman have been promoting faster than optimal rates. The promotion of such ideas
is based on the fallacy that faster rates yield more accuracy and/or more detail. Weather
motivated by profit or ignorance, the promoters, leading the industry in the wrong direction, are
stating the opposite of what is true.

Here's a good, sum it up quote from Dan which really puts it all into perspective.

Oh, and as for dithering, here are some folks that really like POWR - I've never heard of Michael Bishop or Bob Katz though...

A lot of people wrote:

"I've also just finished reviewing the POW-R. In My Opinion, all other WLR schemes now are mostly obsolete. The POW-R consortium has provided the music industry with tools and methods to preserve the quality of high resolution audio in 16bit renditions."
Glenn Meadows, Masterfonics, USA

"Absolute rave! Terrific. I just got an unsolicited letter of praise from a highly-critical client, whose CD I did this year using the identical equipment as last year (both live concert CDs) and the only difference was the POW-R dither."
Bob Katz, Digital Domain, USA

"We just listened to the Mahler 3 as it was remastered using the POW-R box and it sounds GREAT!!! Erica and I wanted to pass on to you our enthusiam over this processor. The sound is remarkably clear and presice. Things now sound as we remembered them at the session with full impact. What a difference."
Michael Bishop, Telarc Records, USA

I have tried your POW-R system for a week and I must say I am impressed. This must be the new standard for wordlenght reduction. [...] It is amazing how little it does and at the same time how essential it is. As one of my clients said " With this you add the Disney twinkling star to the sound ". Well put I should think.
Maarten de Boer, The Masters, Netherlands

You wrote:

Why are you thinking that the Nyquist Theory is the only thing that applies when recording at 96khz. Do you have any sources?

Uhhh, because it IS the only rule which applies (aside from those that don't have any bearing, such as "It will take a lot more hard drive space..."). Sources- sure. Anyone who knows anything about digital.

Nika wrote:

“Theorem 1: If a function f(t) contains no frequencies higher than W cps [cycles per second, or “Hz”], it is completely determined by giving its ordinates at a series of points spaced 1/2W seconds apart.”.........
The significance of this statement is often overlooked. Shannon tells us that the entire waveform, in proper amplitude, frequency, and phase can be recreated (as he says, “completely determined”) with only sampling points given at greater than half the highest frequency to be sampled.

Now, I've about beaten the crap out of this subject.

Look, don't take it personally, but it seems that every 6 months or so, someone new comes into the forum who acts as though they know how digital works, but in reality, all they're doing is spewing forth crap they read from someone else (or worse, off of apogee's website). We go through the same stuff everytime and it escalates into a pissing competition. If you don't believe me, search for "Digital" on this site or even better "44.1kHz" AND "96kHz." You'll see a crap load of topics where people come in (usually new) and argue the same crap you are and then have folks poop all over them and then we never hear from them again (or they move over to the Marsh where they don't even talk about audio anyway!)

So, again, don't take it personally, but please, if you're going to come here and argue with a scientist (me) and a seasoned engineer (Remy) have something other than the literature from the Apogee website.

J.

Posted: Fri Jan 20, 2006 6:39 pm

RemyRAD wrote:

You know the one thing most people here have not mentioned is that the most deleterious effects at 44.1kHz 16 bit recording, is the heavy brick wall filtering necessary. I believe that is the primary reason why most people don't like the marginal capabilities of 16-bit 44.1kHz recording.

One of the reasons the higher resolutions recordings do sound slightly better is the less needed brick wall filtering and the appropriately higher frequency at which it needs to happen. Nobody likes the sound of these filters but it is sort of like death and taxes, when dealing with PCM.

Few of our esoteric "exampliers" have expounded on the benefits of DSD whose sound appears to be our closest analogy to analog (of course in the listening tests they did not give us any examples of "saturation" which is still a nasty third harmonic clipped waveform). Such is life. Unfortunately as of this date, it's still too much money except for the very affluent. (hey why can't they double that 2.53MHz sampling frequency anyhow?) Just kidding.

Poor old
Ms. Remy Ann David

As with a lot of things Remy, you're spot on!

Although, I did make reference to the filtering in one of my previous posts.

Oh, and they are doing a double rate DSD now!! Holy crap on a cracker! How much space will that take on a hard drive??

Back to the filters - a filter of the magnitude necessary to work with 16 bit audio at 44.1kHz so we wouldn't hear it needs to work at roughly 120dB cut per 1/10th of an octave. YIKES. That causes a major phase ripple, not to mention image smearing due to the resonance!

If it were possible to have a standard PCM delivery beyond RedBook 16/44.1, this would be the single most positive factor. The frequencies beyond 20K would be a factor, but minimal at best.

Stay cool!

J.

Posted: Fri Jan 20, 2006 10:23 pm

I'm cool daddy........Can I have the keys to the car?

Remy Ann David

Posted: Sat Jan 21, 2006 11:05 am

I stopped reading after this part, just to jump in:

covenant66 wrote:

Plus, who in the world wants to master a 16/44 recording?

Joe Lambert and John Golden both mastered stuff we did at the studio here in Harrisburg, PA. We were using a Mackie d8b system at 16/44. Until last month, that's what everything was recorded at. We just switched to a Nuendo 3/RME Fireface system at 24/44.

Posted: Mon Jan 23, 2006 12:09 am

Most people seem to notice a slight difference (usually percieved as "clarity") between 16-bit & 24-bit, but I haven't heard too many people claim to hear much of a difference between 44.1 khz and 96, let alone 192.

In answer to the original post, I personally think it's a good idea to record and mix at 24-bit/32 FP if possible, since one never knows what the future may bring...a mastering house may prefer it that way for example. I do think the audible differences are small and it may be some time yet before there are commercially available CD's at greater than 16/44, but it would be nice to be able to re-mix/re-master some things at a higher bit rate if it ever happened.

I'm not an expert on digital audio, that's a view simply based on listening & hearing the opinions of others.

This discussion is centered on what must be the biggest can-o-worms facing audio engineering today. Good debate guys.

Posted: Fri Jan 27, 2006 4:08 pm

ok, i fall into the "inexperienced" category, and i've to compensate for my lack of funds through VSTs - most of these are freewares...i'm a poor student home studio person, and this is my first post, so dont bash my ass in!!

From what i understand, if u record in 24/96, and just mix it down raw, it's pointless IMO - the added overhead is REALLY useful if you use tons of plugins - like i do - since with the sue of every plugin, there is a small drop in quality. if like me you use tons of plugins, recording at 16/44.1 is actually going to give u pretty shit sound, theoretically, while with 24/96, u'll still have the kind of overhead where when u take a mixdown, your mix will sound a whole lot better than the same stuff recorded in 16/44.1

can someone PLEASE explain dithering to me - i just mix in 24/96 and then take a mixdown in 16/44.1... i'm very bereft of purist pro audio gear

AD/DA - ESI ESP 1010 wamirack - yes yes, very shoddy, but got to make do with what i can in my extremely tight budget!
DAW - Nuendo 3 - cough cough, lets not question the origins of this, purely used for learning purposes...
guitar tracking - line 6 podxtpro - now this is a TRUE gem - sure it has it's shortcomings, but in its price and range of tones, it's unbeatable! A tube amp would beat it, but then i'd have to beg on the streets for a while!

Note: am a student, some help would be greatly appreciated!

Posted: Sat Jan 28, 2006 12:11 am

Good work folks..Lets keep it going . Thanx to all for the great information being shared here. It is a boon to those, like me, who have a few decades of experience but little in the digital world. I'm always learning. Methinks a sticky is at hand here.

Posted: Sun Feb 26, 2006 4:51 pm

Ya'lls boys n'goils take some time to read the following technical papers, that way you can separate the subjective from the market hype and commercialism that pervades the industry and realize what is happening with actual physics of sound. Your lives may never be the same. I was born again and again and again, and now, Valhalla is a wonderful place -

http://www.lavryengineering.com/documents/Sampling_Theory.pdf

http://www.lavryengineering.com/white_papers/sample.pdf

http://www.lavryengineering.com/white_papers/jitter.pdf

http://recforums.prosoundweb.com/

http://www.lavryengineering.com/lavry_forum/

Posted: Fri May 09, 2008 2:44 am

Quote:

Cucco - Second - YES I am stating that this entirely has to do with Nyquist. Period. I don't care how many slices of audio you have - you only need two and only get two to draw an accurate picture of a wave, adding 95,998 extra pictures won't get you a more accurate picture of a 20Hz wave than 2 pictures will.

I'm no expert in this but I'm perplexed by the comment above.

I would reckon you need at least 4 slices to represent a wave. More samples to accurately represent a complex wave. An infinite number of slices to get a perfect representation.

An interesting thought crossed my mind regarding "hearing a difference". There are people who doesn't mind listening to audio on VHS vs DVD. They say they can't tell the difference. They don't care.

My question is, "Do these people have bad ears or good ears"?

I think they have good ears, so good that they can clearly hear the poor quality audio, and discern speech/sounds on "technically noisy" mediums like VHS, VCD, mp3 etc.

So I'm saying Audio Engineers must have very bad ears to make really good recordings. Noisy recordings irk them as they can't hear the content clearly so they have to make cleaner (better) recordings.

It is also known that prolong exposure to sound causes deterioration to hearing. Ageing includes the activity of prolong exposure to sound. An experienced AE will have higher degree of bad ears than a young Newbie.

Since newbies have better ears, Convenant66 might just be right when he said he was able to hear the difference between a 16/44 and 24/96.

Cucco might just be "hard of hearing" and lamblasting based on theory.

And I must admit that I'm in the camp of "hard of hearing" as I get irritated by noisy mediums and gears. Thanks to speakers like NHT2.5i, I get to enjoy the music and performance of a good recording.

Posted: Fri May 09, 2008 3:22 am

Since you've dug up a 2-year old topic, I may as well point out the source of your misunderstanding.

The trouble comes in the word "draw" in the post you quoted. Indeed, if you want to draw an arbitrary waveform accurately on paper, you need as many points as are distinct within the thickness of your pencil. But the waveforms we are talking about are band limited by an anti-aliasing filter when originally sampled and put through a reconstruction filter when converted back to analogue. The act of band limiting and reconstruction constrains the possible shapes of the waveform.

The Nyquist theorem postulates that just more than two points per cycle of the highest representable frequency are sufficient to determine both the amplitude and phase of that frequency, and hence is more than sufficient to describe all frequencies lower than that.

The reconstruction filter effectively fills in the missing sections of waveform between the dots. It is the system designer's task to ensure that this filling-in is unique, in which case it must represent a band-limited version of the original signal.