Quote Richard Graham:

---

I've just read this entire thread, it's taken me all day, and I don't understand it AT ALL. I thought digital audio was perfect quality, like CDs?

---

Listening to 'Dueling Banjos' on my portable CD player whilst buggering a goat after drinking a 6 pack; THAT'S PERFECT.

Quote Music Manic:

---

There's a part I don't get:

---

Oh no...!

Quote:

---

aren't there other factors which contribute to the sound that appears at the I/O? What about the filters,clocks etc?

---

Yes, all these things can make a difference. But Nichols was seeking to find a way of explaning a specific phenomena he felt (subjectively) was a benefit to low frequencies when recording with 24 bits (as opposed to 16)

Quote:

---

Would like to know more about how Lo and Hi frequencies are dealt with in analogue and digital domains and what problems they cause.

---

Wouldn't we all!

hugh

--------------------
Technical Editor, Sound On Sound

Quote Crow:

---

To quote from the article, ‘As a percentage of the difference between adjacent samples, the quantisation error is much greater for a low-frequency signal than a high-frequency one’.

I’m really curious about this postulation, but I’m not sure that it’s at all a relevant comparison.

---

I just sat down with a notepad and pencil and worked out some examples that have satisfied my curiosity. Looking at it purely mathematically it does stand out and the numbers seem to shout at me, ‘This is significant’. But, a part of me says, is it really significant? Now that I’m happy with the maths, I’m still wondering whether the comparison has any bearing on how low and high frequencies actually sound at differing bit depths. I don’t have access to equipment to test it out subjectively myself, but I’m now wondering if it’s a common perception that bass frequencies sound ‘better’ when sampled at higher bit depths. And if so, is this improvement more noticeable than it is for higher frequencies?
I guess it comes down just as much to psychoacoustics in the end and I don’t mean by that the aural study of Jason Voorhees’s chainsaw in action.

Quote Crow:

---

Quote Richard Graham:

---

I've just read this entire thread, it's taken me all day, and I don't understand it AT ALL. I thought digital audio was perfect quality, like CDs?

---

Listening to 'Dueling Banjos' on my portable CD player whilst buggering a goat after drinking a 6 pack; THAT'S PERFECT.

---

Thank you for bringing us back to the essential point of this thread: WHAT ABOUT THAT BANJO?

We have a right to know!

We demand to know!!

We shall not rest until we hear The Banjo Worth Recording!!!

--------------------
Dynamite with a laser beam...

Quote UnderTow:

---

Well isn't that how we hear things? Isn't that why we talk about RMS values as they are closer representations of how we as humans hear things?

---

RMS measurements don't have an explicit time constant involved and tend to be closer to a measurement of the energy contained in a signal rather than peak amplitude... But yes, RMS levels do generally approximate better to perceived loudness in some reagrds than peak levels. Not sure howe RMS measurements can be related specifically to the concept you were suggesting earlier though.

Quote:

---

How so? We can only hear stuff after it has gone through the DAC with the relevant reconstruction filters so I don't see how looking at a plot that doesn't represent the final waveform explains what we do or don't hear.

---

He was trying to explain what is, fundamentally, a very simple point about the relative amplitude of quantisation errors during A-D conversion between 16 bit quantising and 24 bit quantising. That's what the graph is intended to show. It's not the clearest graph I grant you, but I think it more or less does what it was intended to do. The subsequent D-A conversion does nothing to change the principle that he is talking about, and would only cloud the explanation further...

Quote:

---

Maybe I am missunderstanding something here. Would you care to explain?

---

I have done, and I'm worn out!

Quote:

---

Yes but how does it affect the wave after reconstruction? Surely this is relevant?

---

Not to his argument, no. The relative amplitude of error is casued by the A-D process -- the quantisation accuracy. The D-A process plays no part in that whatsoever.

Quote:

---

I am trying to understand things. At the moment I don't quite understand why the reconstruction filters are irrelevant.

---

Because they don't have any effect on the quantisation levels. The filters simply remove the sampling artefacts. They play no role whatsoever in quantisation.

Quote:

---

Well if the dithering process linearises the quantization steps, I don't see how the quantization error (when properly dithering) affects the low frequency waves in the way that Mr Nichols seems to be describing.

---

I think it fair to say that his argument has overlooked the effect of dither to linearise the quantisation process. Putting his argument another way, perhaps he is saying that the signal-noise ratio improvement when stepping up from 16 bit quantisation to 24 bit quantisation, is more apparent for low frequencies than high frequencies... and when you look at it that way, it does seem harder to justify....

Quote:

---

Do you see what I am driving at? If I am completely wrong, please explain me where my misunderstanding lies.

---

Yes, I do see what you are driving at... I quite understand the action of dither. The pertinent question is the what is the relationship between the dither noise and the signal frequency...

Quote:

---

I don't get this. If the steps are linearized, then his explanation is incorrect. Again, I might be misunderstanding things but I need convincing of that.

---

I think I am seeing the light. I have been trying to explain Nichols argument as presented... you are trying to relate it to a real-world converter that is properly dithered. We have been talking cross-purposes. I think I can now see where you are coming from.

The issue, perhaps, is one of how the signal frequency and dither signal interact, because the randomised error signal (dither noise) associated with each sample would be proportionately smaller for LF signals than HF signals in 24 bit quantising when compared to 16 bit quantising. My head is starting to hurt... I need to go away and think more about this...

Quote:

---

Hmmm ... I'm still not convinced. I think his argument is only valid in an un-dithered system.

---

Certainly his explanation has assumed that case, and yes, that is not a valid way to conceive a well engineered digital audio system...

But I'm not yet convinced that the intrinsic quantising error won't still have an effect along the lines of his proposal...

Hugh

--------------------
Technical Editor, Sound On Sound

Quote Hugh Robjohns:

---

I think I am seeing the light. I have been trying to explain Nichols argument as presented... you are trying to relate it to a real-world converter that is properly dithered. We have been talking cross-purposes. I think I can now see where you are coming from.

---

Hurray hurray!

Quote:

---

The issue, perhaps, is one of how the signal frequency and dither signal interact, because the randomised error signal (dither noise) associated with each sample would be proportionately smaller for LF signals than HF signals in 24 bit quantising when compared to 16 bit quantising. My head is starting to hurt... I need to go away and think more about this...

---

Ok, I have to think about this too. Please do share your findings when you have had some time to think about this.

Quote:

---

Hmmm ... I'm still not convinced. I think his argument is only valid in an un-dithered system.

---

Certainly his explanation has assumed that case, and yes, that is not a valid way to conceive a well engineered digital audio system...

---

Ok we agree on this.

Quote:

---

But I'm not yet convinced that the intrinsic quantising error won't still have an effect along the lines of his proposal...

Hugh

---

Fair enough. I am not convinced either way but I am tending away from Mr Nichols proposition.

UnderTow

I give up.

--------------------
Midipicks - the all new MIDI Guitar forum...

My "2p worth", I have been warming myself at this gentle blaze for a few days now, but the weather's getting much warmer and I could do without this extra heat.
Any sound in the lower frequencies (other than a sine wave) surely is going to contribute more upper partials that fall within the limits of our hearing range than a higher frequency sound. The more of these that can be portrayed by a higher bit rate, obviously the more accurate (fuller) the sound will be. Surely, don't the bass sounds benefit more from this accuracy simply because they have the advantage (even acoustically) of having more upper partials within the range to capture in the first place?

--------------------
Don't ya jus luvvit?

. . . you must surely all know the joke about the BANJO player, booked for a New Year's Eve gig.
When it was over, the hotel manager (who liked what he'd heard) gave him an advance booking for the next New Year.
"That's fantastic" said the banjo player "can I leave my gear here?"

--------------------
Don't ya jus luvvit?

Quote Richard Graham:

---

I thought digital audio was perfect quality, like CDs? It would help if everyone stopped going on about 'dithering', 'converters', 'bits', 'resolution' and all the rest of it. I mean wtf?? Doesn't anyone around here know ANYTHING about audio?

---

This has to be a joke, right?

--------------------
www.ehsound.co.uk - Live Sound Hire & Services

Quote Michael Harrison:

---

Quote Richard Graham:

---

I thought digital audio was perfect quality, like CDs? It would help if everyone stopped going on about 'dithering', 'converters', 'bits', 'resolution' and all the rest of it. I mean wtf?? Doesn't anyone around here know ANYTHING about audio?

---

This has to be a joke, right?

---

What? The whole thread? I thought a couple of bits made sense.

--------------------
Dynamite with a laser beam...

Quote Michael Harrison:

---

Quote Richard Graham:

---

I thought digital audio was perfect quality, like CDs? It would help if everyone stopped going on about 'dithering', 'converters', 'bits', 'resolution' and all the rest of it. I mean wtf?? Doesn't anyone around here know ANYTHING about audio?

---

This has to be a joke, right?

---

See my other thread, about next month's SOS article:

http://www.soundonsound.com/forum/showflat.php?Cat=&Number=288006&page=0&v iew=collapsed&sb=5&o=&fpart=1#288006

--------------------
Battle flags are flown at the feet of a garden gnome.

Quote Steve Hill:

---

.. I thought a couple of bits made sense.

---

So Steve, you advocate a 2-bit recording system over a 16- or 24- bit one. Interesting!

I have my own methods for spicing up old digital multi-track tapes for CD release. Step back Mr Nichols, please….

Stick your old 3M digital tape reels in the washing machine with a big dollop of Aurial Diabolical. There is of course a lot of debate as to which conditioner to use; I prefer to use Bit-Shitter, as it really gets rid of those NASTY stains.
When it comes to line drying the tapes, it’s been scientologically proven that an East wind will add 3 dB of OOMPH to the underlying BAD ASSNESS of the recordings. A West wind will add 10 chart placings to any recording without a banjo. Recordings with a banjo will of course immediately be returned to the Ku Klux Klan and automatically released on the next volume of Now That’s What I Call Incest Vol X.
A Northern wind will subtly mix in a choir of strangled pygmies at the choruses. The pygmy chorus will be phase-shifted and flanged as per the EEC guidelines on these matters.
A Southern wind will virally mutate your recordings and produce a line full of extended mix tapes, at least one of which will induce epilepsy in guinea pigs.
A full moon is best avoided as it maliciously overdubs a bagpipes solo at the bridge.

Edited by Crow (29/04/06 10:55 AM)

Quote Steve Hill:

---

We shall not rest until we hear The Banjo Worth Recording!!!

---

remind me to drop it round..... I have several recordings to prove that it's possible.... some of them are even musical....

Max

--------------------
if you don't know who i am, i aint gonna tell you.

some of Stockhausen's followers have written a collaborative piece called.


"Sonata for car crusher and Gross (144) Banjos"



--------------------
if you don't know who i am, i aint gonna tell you.

Quote Max The Mac:

---

Quote Steve Hill:

---

We shall not rest until we hear The Banjo Worth Recording!!!

---

remind me to drop it round..... I have several recordings to prove that it's possible.... some of them are even musical....
Max

---

I once heard John Mclaughlin playing a banjo in his inimitable machine gun stylee. This wasn’t busking in Alabama or Finsbury Park, but at a gig with Jack Bruce, Bilham Cobly and Stu Goldberg. Whether that was worth recording is down to your personal peccadilloes.

in actuality, i have recorded and mixed several albums of "traditional" (mostly Irish) music, complete with Banjos, and while I find the Banjo as humorous as the next guy, i'd have to admit that these albums were worth recording.... including the Banjo parts....

make of that what you will.

--------------------
if you don't know who i am, i aint gonna tell you.

'Steely Dan Play Acoustic Live on Banjo' engineered by the very capable producer, yet somewhat (dither)ing writer, Roger Nichols.

I'd pay good money for that.

What a wonderful world it would be.

Caught a bit of Eagles in concert, 1973-vintage, on BBC4 on Friday night. Jeez, that boy could play a banjo...

--------------------
Dynamite with a laser beam...

Apologies for the time it has taken me to get back to this thread. Illness in the family...

Anyway, have given this much more thought now, and talked to a few boffins and eggheads about it. The general concensus is -- as Under Tow was suggesting -- that Roger Nichols' arguments don't apply to a correctly dithered system. The quantisation errors he is comparing between 16 and 24 bit systems simply don't exist when the system is dithered correctly.

A correctly dithered system is inherently linear and quantisation errors translate into a constant noise floor -- which is lower or smaller for 24 bit systems than 16 bit systems.

Therefore 24 bits does sound better than 16 bits because the noise floor is lower. Some may perceive the improvement to be greater at low frequencies than high frequencies.... but Nichols' quantisation arguments don't provide the explanation I'm afraid.

Thanks to UnderTow for his patience and perserverence in arguing his point.... after a confused and frustrating start I think we got there in the end

hugh


--------------------
Technical Editor, Sound On Sound

I'll just add as a post-mortem (as an engineer and problem-solver)...this whole thread illustrates the keen importance of:

1. Clearly defining your assumptions and givens.

2. NOT applying conclusions made under one set of assumptions to another set of assumptions.

In this case, a conclusion based upon an undithered system has been applied to a dithered system.

You'd be awfully surprised how many "problems" are solved by forcing someone to go back and state all of their assumptions and givens to their argument.

Anway, I'll get off of my teacher/soap-box now.

P-X

--------------------
My Home Studio Build Thread

Quote PrinceXizor:

---

...this whole thread illustrates the keen importance of:

1. Clearly defining your assumptions and givens.

2. NOT applying conclusions made under one set of assumptions to another set of assumptions.

---

Agreed 100%! Erroneous logic like you describe is applied with alarming abandon to discussions everywhere on the Internet - at the least sowing mass confusion, and at the worst causing serious misinformation overload with potentially dire consequences for someone down the line.

I spent the last 10 hours reading this entire thread .... with 15-minute breaks after every 3 minutes of reading. It wasn't the maths and scientific lingo ...... honestly ....... but rather it was that banjo sound invading my head!

Tylenol-3 isn't helping much. Does anyone know where I can get morphine from the black-market? Disassociation of mind and body seems to be the only solution.




OH MY GOD ...... thoughts of banjo mariachi bandits invading my head again!
HELP!!!!!

--------------------
Adopt kids, not styles!!!!!

Quote OmarHash:

---

OH MY GOD ...... thoughts of banjo mariachi bandits invading my head again!HELP!!!!!

---

If you plan on imbibing ayahuasca any time soon, please let me know so that I can arrange for a phalanx of banjo strumming skateboarding dwarves to enliven your experience

P.S.
No offence intended to skateboarders or banjo players…. or people of restricted growth.

Quote OmarHash:

---

I spent the last 10 hours reading this entire thread ....

---

Sit down and fekin learn, the rest of us have to! Have some Peyote, I find it helps keep me awake.

The Banjo is all part of learning... They pump it in here... subliminal like, they are engineers, they have techniques.

Shame really that Undertow has alredy issued a rather splendid apology for his obviously correct argument and has therefore morally abdicated the right to bask in the glow of having been correct all along.

Incidentally, I was just going to support you there mate

The only thing I don't get is................oh stuff it, I'm going on holiday tomorrow so I don't care.



A.

--------------------
Andi, www.thedustbowl.net Mixing, Mastering, Audio Editing at The Dustbowl Audio

Wow. I've just read this thread. Phew!

I read the the RN article a month or so ago and its 'logic' has been rattling round my head.

I subsequently started two threads questioning RN's assertions - one on this board - and got a reply from Hugh Robjohns saying 'Let's not start the fight all over again' - and eventually got pointed to this thread. Well, I didn't know this topic had closure - I haven't read anything in SOS (magazine) about the accuracy (or otherwise) of that article.

My observations: most people believed what they read in SOS - and then defended it like it was their religion. And like a religion, they took it at its word, rather than questioning the accuracy of the science behind it.

UnderTow stuck to his guns - and there was a lot of resistance - but he was right, and that seems to be generally accepted now, and maybe even understood. Has anyone acknowledged that? (I've just seen that Hugh did - kudos)

And where has RN been during all of this? Mocking up plug-in UI's in MS Paint?

The best that could come of this is that people allow their accepted beliefs to be challenged - and defend their positions with logic rather than the prejudice they've doemonstrated here.

For some, SOS magazine is a reference, and as such there's a greater responsiblility to the facts and readership at large, rather than to the staff covering each other's asses.

After all, it's not New Labour . . .

can someone please condense all the good bits for me to read...

in particular banjo jokes and jokes about funny plugin names...

thanks dudes...

Quote gerard:

---

can someone please condense all the good bits for me to read...

in particular banjo jokes and jokes about funny plugin names...

thanks dudes...

---

It's a load of old bollocks isnt it? What we have here is a bloke, an American bloke no less. He made a shitload of money designing nuclear bombs [theyre bombs when something goes wrong]. He then made a load more out of working with some of best artists in the world. In fact he made the benchmark digital recording. The one people probably a/b with more than any other... Possibly still unsurpassed...

Here are the credits Fagin's 'The Nightfly'...

Recorded and mixed entirely on 3M digital 32 track and 4 track machines at Soundworks Digital Audio/Video Recording Studios, N.Y., Village Recorders, L.A., and Automated Sound, N.Y.

Chief Engineer: Roger Nichols

There are few people on this forum who wouldnt piss-spunk if he asked them to make his tea for an album. The bloke virtually invented digital recording!

He has now upset a few people by procuring a plug-in company, putting groovy facias on the apps and quadrupaling the price. Something that we will all do given a resurgance in the third party recording studio market, and the cash to buy a cockey facility somewhere.

Will these accusors be better recordists because they have an oak door on their enormous and infinitely well equipped studios? no they won't. But they will quadrupal their prices given the market to support it!

So what we have is a load of jealousy, hypocricy, back-biting, politics and big knobishness. It stinks!!!

Personally, as a long time reader of SoS, i'm extremely pleased that someone of this calibre is writing a column. I think it's a real coup. I like his style, its friendly, a little brash perhaps, but down to earth enough for me to understand.

"Nichols article on 24 bit recording WRONG!" Don't make me laugh... how many global number one benchmark albums have you engineered?

Edited by Hugh Robjohns (08/06/06 11:14 AM)

Quote ow:

---

There are few people on this forum who wouldnt piss-spunk if he asked them to make his tea for an album.

---

You do realize you can only speak for yourself?

Quote ow:

---

"Nichols article on 24 bit recording WRONG!" Don't make me laugh... how many global number one benchmark albums have you engineered?

---

He is right because of a having had a 'global number one', rather than the accuracy of what he says?

Enjoy your church, brother.

Let's look at those list of credits on that album again:

Quote:

---

Chief Engineer: Roger Nichols
Overdub engineer: Daniel Lazerus
Tracking and mixdown: Elliot Scheiner
Sequencing, percussion and special effecs: Roger Nichols and WENDEL II
Digital maintenance: Wayne Yurgelun, Mike Morongell (Soundworks); Bill Roach and Jiri Donovsky (3M)
Assistant engineers: Wayne Yurgelun, Mike Morongell; also Cheryl Smith and Robin Lane
Original Mastering: Bob Ludwig at Masterdisk, N.Y.C.

---

So what does a chief engineer do when Elliot Scheiner is the tracking and mixdown engineer, and Daniel Lazerus is the Overdub engineer? There are also 4 assitant engineers listed! Wow!

It is pretty hard to go wrong when you have these 2 guys in your corner, and then mastered by Bob Ludwig to boot! There is no way this album could have sounded bad with that technical crew and then with the musicianship that Donald assembled as well. Once again... "WOW!"

This is not meant to be a condemnation, but a question. What does a chief engineer do in a situation like this??? Obviously something pretty special, otherwise Donald would not have listed him as such.

--------------------
James
"Never interrupt your enemy when he is making a mistake" ~Napoleon Bonaparte~

Quote Doublehelix:

---

So what does a chief engineer do when Elliot Scheiner is the tracking and mixdown engineer, and Daniel Lazerus is the Overdub engineer? There are also 4 assitant engineers listed! Wow!

---

I can think of two things off the top of my head: Having the money to buy the studio in the first place and/or being good friends with the band and thus arranging to have the credits. This is speculation of course.

Anyway, the guy is still wrong in his article.

UnderTow

"I can think of two things off the top of my head: Having the money to buy the studio in the first place and/or being good friends with the band and thus arranging to have the credits. "

now that's what I call fighting talk! You've gone from pointing out a technical error (possibly) to inferring that the guy has no skills at all! I don't think so. We simply must have donald or walter in to settle this

Quote tipex:

---

now that's what I call fighting talk! You've gone from pointing out a technical error (possibly)

---

Make that certainly.

Quote:

---

to inferring that the guy has no skills at all!

---

Not at all. It is just idle speculation. There are many other possible reasons for him to have those credits. I just gave two possibilities. Make of that what ever you whish.

I was not in the sessions nor do I know anyone that was so I have absolutely no idea what Mr Nichols did or didn't do. Still, if someone else did the actual recording AND the mix it does make me wonder, as others, what the title of chief engineer means.

UnderTow

I also got pointed to this forum and thread as the 'answer' to the question of this article having 'technical inaccuracies'.

I've read the article, and this thread, and other forums/comments on this. (I'm glad Undertow kind of got things communicated eventually, but his terminology is far from standard, and kudos to Hugh for his efforts). I'm referring specifically to the section titled 'The Bits'.

I don't really see a good explaination. I also don't see any comments in the June SOS (US/International version at least). All I see is effectively "what the article says is irrelevent with proper dither", but I also say that what the article says is on shaky ground from the start and right through.

I don't see any responses to some of the biggest technical problems I see in it, although one or two people hinted at them, wondering, but no responses.

My first comment, he is doing error as a percentage of the 'step size', since when has this been a valid measure of error for audibility? Error signals/noise are normally in absolute terms or relative to the signal level, not the instantaneous rate of change of 'voltage' level.
This may give you bigger numbers in some places with a high frequency signal, but remember the peaks and troughs of the waves have very small differences between samples. The errors he talks about, are basically at the same signal level irrespective of frequency.

Also, his whole premise that the noise level is 'less', and hence less audible, for higher frequencies is, er, inaccurate.

Let's ignore dithering for now, and just focus on the quantization noise aspect, as the article does (rightly or wrongly).

Standard signal analysis methods look at 3 main signals:
1. the error(noise)-free perfect desired signal (analog input signal for this situation)
2. the quantized signal (A/D output - that inherently defines signal between samples)
3. the error (noise) signal.
Then:
A/D output = perfect signal + error signal
i.e. A/D output is the same as adding (mixing) a noise signal to your 'prisine' signal.
This error signal is effectively a real signal that you can hear. If you regard 24-bit as perfect (or effectively perfect), you can hear this error signal for 16-bit quantisation by simply having a full 24-bit signal, truncating it to 16-bit then taking the difference between this and the original.
Math does the same.
You can then analyse the error signal.
To be least offensive on the ear, this noise signal should be truly random, with no tones, or other strange noises. This is why you use dither, to make this error signal random noise (flat frequency content).
With straight truncation the error is not audibly nice, and has artifacts that are dependent on the original signal.

Anyway, getting back how this relates to what the article says, for 16-bit quantization compared to 24-bit, the error signal, which is the same as adding a real/desired signal to the original, itself has a max peak-to-peak amplitude of 256 24-bit LSBs (i.e. 256 x the least significant bit voltage value for 24-bit).
The point is that this signal amplitude can be the same across all frequencies.
What the article is suggesting is that when you add this noise to a high frequency signal you hear it less than with a low frequency signal. If you have a regualr quiet mid-range sound that you mix with just a louder low frequency signal, then mix it just with a louder high frequency signal
does that mid-range sound sound quieter with the high freq added, than with the low freq added? From the article's logic it would.

Now back to dither a little bit, and A/D at 16-bit (i.e. old stuff), if your input signal has a better noise floor than the 16-bit converter, then without 'analog dithering', the A/D is going to be dominated by equantization noise, as the accuracy of the input signal is effectively being truncated (just like truncating 24-bit). If you don't purposefully add noise to the incoming signal to a simple, 16-bit A/D, you'll get the signal dependent quantization noise (i.e. bad noise). Plus older A/D had low end problems anyway. But I'm starting to gert of my area of knowledge now - I'm not a device historian, I can't say for sure what old 16-bit A/D did, but it's far better today.

Which kind of ends up with the 'answer' given before: what the article says on this is irrelevent in any system that has dither implemented correctly. But I'll add, please don't apply the reasoning to anything else, as it is fundamentally flawed anyway.

After starting with, what I consider to be a rather shaky premise and dubious error calculation method, the article does make several numerical (can only have multiples of 256 'steps' between 16-bit values), arithmetic (those number do not make 700%) and diagramatic errors (straight lines between samples, the 'low frequency' diagram zooms in to a higher frequency component).

I have a lot of respect for fighter pilots, but I'm not going to give much weight to their explanations of the nuts and bolts of aerodynamics. Roger Nichols is an ace fighter pilot.

Bring on the flames, abuse, or even plain ostracism. I'm unlikely to mind. Although intelligent criticism would be nicer.

G

...anyone got anymore banjo jokes?


by the way, nicely worded post GJordan... are you new to the forum? don't worry we will soon have you talk'n trash and crap like the rest of us...

har har...

Quote gerard:

---

...don't worry we will soon have you talk'n trash and crap like the rest of us...

---

Yes, I'm running overnight tutorials. This course is designed to be available when 'you' need it. Here at the ow college of trash & crap we structure our courses around you!. We make it easy to study! Click here for more information, course availability and rates. Book Now! places are limited! Success guaranteed!

Quote gerard:

---

...anyone got anymore banjo jokes?

---

Q. What's the difference between a 4-string and a 5-string banjo?

A. A 5-string burns for longer

Q. What's the definition of an optimist?

A. A banjo player with a pager

Q. What's the definition of perfect pitch?

A. Throwing a banjo into a skip/dumpster without touching the sides.

(The old ones are the best)

--------------------
Dynamite with a laser beam...

Quote GJordan:

---

My first comment, he is doing error as a percentage of the 'step size', since when has this been a valid measure of error for audibility?

---

I think Roger was simply using his (bizarrely calculated) 'percentages' as a way of highlighting the relative sizes of the proportional errors.

Quote:

---

Error signals/noise are normally in absolute terms or relative to the signal level, not the instantaneous rate of change of 'voltage' level.

---

The voltage increment denoted by the least significant bit in a digital quantiser is, by definition, an absolute, fixed amount, with a clearly defined relationship to the peak signal level. Roger wasn't talking about instantaneous changes as such, just the possible level differences between adjacent samples.

As I undertsand it, essentially he was comparing the difference in error signal amplitudes between 16 and 24 bit quantisation... although as we know, in a properly dithered system there is no error signal, only random noise...

Quote:

---

but remember the peaks and troughs of the waves have very small differences between samples. The errors he talks about, are basically at the same signal level irrespective of frequency.

---

Sorry, don't follow the first sentence at all. If you compare the area where an analogue signal passes from the negative half cycle to the positive half cycle (the zero-crossing region) then, when sampled, adjacent samples of a low frequency signal in this are are likely to have relatively similar quantisation values -- becaue the rate of change of the signal is reltively low. Adjacent samples of an HF signal are likely to have significantly different quantisation values because the rate of change of the signal is relatively high.

As I see it, this was the fundamental assertion on which the rest of Roger's thoughts were based... and this bit, at least, is correct.

Quote:

---

Also, his whole premise that the noise level is 'less', and hence less audible, for higher frequencies is, er, inaccurate.

---

Quite

Quote:

---

Let's ignore dithering for now, and just focus on the quantization noise aspect, as the article does (rightly or wrongly).

---

This is treading the same dodgy ground as Mr Nichols... and strictly, we should be talking about quantising error rather than noise, to avoid further confusion. By definition, noise is random whereas quantisation error is statistically related to the input signal. A correctly dithered system exhibits a noise floor. An incorrectly (or absent) dithered system exhibits no noise floor at all, but signal errors (distortions) which are related to the source sound.

Quote:

---

A/D output is the same as adding (mixing) a noise signal to your 'prisine' signal.

---

This is getting messy. The output of an A-D is a sequence of binary values. However, in a properly dithered system, the signal encoded in those binary values is, essentially, the source signal plus a very small amount of random noise.

In an undithered system, the encoded signal is essentially the source plus a raft of source-related distortions. The amplitude of those distortions remains fixed (and related to the size of the LSB) -- and so take on a greater significance and audibility as the source signal decreases in amplitude.

Quote:

---

If you regard 24-bit as perfect (or effectively perfect), you can hear this error signal for 16-bit quantisation by simply having a full 24-bit signal, truncating it to 16-bit then taking the difference between this and the original.

---

First, in a perfect world, an undithered 24 bit system exhibits the same problems as an undithered 16 bit system...low level signals will still produce quntisation errors and the same distortions. The only difference is the absolute amplitude of those distortions.

However, it's not a perfect world and the inherent gaussian noise generated by the interface electronics will automatically act to dither a 24 bit system into near perfection anyway... making the comparison rather difficult.

Of course, truncating a 24 bit signal to 16 bits strips off any dithering at the 24 bit level, leaving you with an undithered (and hence distorted) 16 bit signal. Interesting to play with, perhaps, but not really relevant to the
practical difference between recording with a properly dithered 24 bit A-D versus a properly dithered 16 bit A-D -- which was the premise of Roger's article.

Quote:

---

With straight truncation the error is not audibly nice, and has artifacts that are dependent on the original signal.

---

Agreed...

Quote:

---

the error signal itself has a max peak-to-peak amplitude of 256 24-bit LSBs (i.e. 256 x the least significant bit voltage value for 24-bit). The point is that this signal amplitude can be the same across all frequencies.

---

You're saying the maximum error is potentially one LSB, which is correct, and if comparing a 16 bit system with a 24 bit one, the maximum 16 bit error amplitude equates to 256 LSBs in a 24 bit system... yes, with you so far...

Quote:

---

What the article is suggesting is that when you add this noise to a high frequency signal you hear it less than with a low frequency signal.

---

ROger was suggesting that if you count up the number of quantising levels crossed (in terms of 24 bit LSBs), the relative proportion of maximum potential error compared to the overall difference in quantising level for two adjacent samples is far greater for LF signals than for HF signals. The pure logic of what he said seems right to me, but the concept itself is deeply flawed and not applicable to practice.

Quote:

---

If you have a regular quiet mid-range sound that you mix with just a louder low frequency signal, then mix it just with a louder high frequency signal
does that mid-range sound sound quieter with the high freq added, than with the low freq added? From the article's logic it would.

---

Well, we might have to take noise masking into account here... but really, you are talking about mixing dissimilar, unrelated sounds. He was talking about distortion products generated by and related to the specific source... but we are getting bogged down debating irrelevant points.

Quote:

---

I have a lot of respect for fighter pilots, but I'm not going to give much weight to their explanations of the nuts and bolts of aerodynamics. Roger Nichols is an ace fighter pilot.

---

Amusing analogy! I like it...

hugh

--------------------
Technical Editor, Sound On Sound