Please be aware that the article is not correct. Nichols does not understand digital audio and should not be writing technical articles for SOS.

If you want to learn about digital audio, I suggest you have a look a these articles instead:

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

http://www.users.qwest.net/~volt42/cadenzarecording/DitherExplained.pdf

UnderTow

Hmmm...well, that was not very useful (other than the links). Do you have a specific quibble? I'm sure the technogeeks around here would appreciate your opinion.

P-X

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

I've not read the article in question but I can spot faulty reasoning a mile off.

You have so far failed to provide:

1) Specific parts of the article with which you disagree; all you have said is "the article is not correct"
2) Evidence to support your assertion that "the article [or part of it] is not correct"
3) Evidence to support your assertion that "Nichols does not understand digital audio"
4) Reasons why Nichols "should not be writing technical articles for SOS"
5) Your real name and credentials

If you wish to start a sensible dialogue on this article (or anything for that matter) you first need to learn how to construct an argument - simply referencing other people's writing on the general topic of digital audio gives no credibility to your unsupported claims.

I agree with last comment.

UnderTow, you may be right but you should cite where 'Nichols' (what, not even worthy enough of full name?) is wrong to justify your argument.

You should really try to present a more balanced tone to gain support for your statement.

Regards to all,

The Producer.

Undertow knows hardly anything. He saw a book once and he heard his teacher say something about Nyquist theorem. Oh! He is only 12.

Edited by tex (26/04/06 09:13 PM)

I was shocked by the article and responded as such. Dinner got in the way of a proper explanation. Here it is:

Quote:

---

With most digital audio now being recorded at 24-bit, nobody thinks about the ultimate 16-bit destination and what that conversion does to the sound quality.

---

This comment is not supported by any data and is a bit of an insult to all the people that do think about it. He seems to be saying that he is the only person that has thought about this.

Quote:

---

The first impression is that you are just changing the noise floor from -144dB to -96dB, which isn't going to hurt the punchy high-level mix that you spent so much time on. But there is much more to this conversion than meets the ear.

---

This is incomplete to say the least. You could have a recording with a noise floor way above -96 dB FS. It won't suddenly go down to -144 dB FS just by converting to 24 bits.

Quote:

---

Just to refresh your memory, 24-bit audio gives you 256 times more resolution in the position of each sample on the waveform.

---

This isn't correct. It only gives more resolution to low-level signals. This can be easily demonstrated by converting a 16 bit file to a 24 bit file. The extra 8 bits are just filled with zeros. There is no requantization of the audio. A (loud) signal which lives in the upper bits of a 16 bit audio file does not gain any resolution by having extra zeros added in the extra bits.

Quote:

---

The first time I heard 24-bit, recording Bela Fleck and the Flecktones, I expected to hear more sheen and high-frequency clarity. I was incorrect. The most noticeable difference was in the low frequencies: the mouthwatering sound of the bass, the breathtaking realism of the kick drum, the clarity of vocals, and the low-end improvements that finally make the banjo worth recording. It took me a while to figure out why this was the case, but I finally got to the bottom of it.

---

This is anecdotal and isn't based on any facts. He seems to base his whole premise on "bad bass" in 16 bit audio on this recording.

Quote:

---

This 256 times higher resolution of a 24-bit sample is in effect everywhere on the waveform, from the lowest levels to the highest peaks. A sample point nearing 0dB full scale is 256 times more accurate than the same sample recorded at 16-bit.

---

This is incorrect. The resolution is only added at the lower end of the amplitude scale.

Quote:

---

To cut down on the confusion with bit sizes, let's use the size of the smallest bit in the 24-bit scale as a reference and call it a step. The difference between Sample A and Sample B in the 24-bit recording is 16 steps. The difference between the same samples in the 16-bit recording is 112 steps. That is 96 steps away from where it should have been — a 700 percent error in a low-frequency signal.

---

Not only can you not compare 24 and 16 bit signals like this (you need to to add dither during conversion to linearise the quantization steps) but the man can't even calculate percentages!

Quote:

---

The voltages generated in a converter are from small resistors that are trimmed by lasers during manufacturing. The small bits that are prone to error are different for every converter, even from the same manufacturer. This means that bit 23 on the converter on track 1 may be different than bit 23 on the converter on track 12.

---

Is he talking about segmented DACs? He does have this paragraph in his article:

Quote:

---

Current oversampling converters do not have the linearity and tracking problems that conventional converters have. Usually you are dealing with a 1-bit converter that samples at 256 times the sample rate. This type of converter just looks at the incoming signal and compares it to a reference voltage. The comparing circuit then decides whether the reference voltage should go up or down. This is done very fast many times between samples. Now the reference voltage that matched the incoming signal determines the resulting PCM value. Only one bit, and no linearity errors.

---

But nearly every soundcard you buy these days uses oversampling. I might be wrong but he seems to be looking at antiquated technology and basing a whole article on that. This just confuses things.

A few comments to the response:

Quote:

---

I've not read the article in question but I can spot faulty reasoning a mile off.

---

Not supporting my comment is not faulty reasoning. It's just lack of evidence.

Quote:

---

You have so far failed to provide:

1) Specific parts of the article with which you disagree; all you have said is "the article is not correct"
2) Evidence to support your assertion that "the article [or part of it] is not correct"
3) Evidence to support your assertion that "Nichols does not understand digital audio"
4) Reasons why Nichols "should not be writing technical articles for SOS"

---

See above.

Quote:

---

5) Your real name and credentials

---

This is totaly irrelevant. I prefer to let the facts talk rather than reputation. Actually I am against any form of reputation or lack thereof clouding the issue.

Quote:

---

If you wish to start a sensible dialogue on this article (or anything for that matter) you first need to learn how to construct an argument - simply referencing other people's writing on the general topic of digital audio gives no credibility to your unsupported claims.

---

Oh stop being so patronising. You have no idea if I can construct an argument or not as I didn't even attempt to argue! (Talking about faulty reasoning ...) Maybe you should have read the article before responding ...

Those articles are long and complex and it would be completely pointless and extremely time consuming to rewrite those articles in my own words just for this purpouse.

UnderTow

Quote tex:

---

Undertow knows hardly anything. He saw a book once and he heard his teacher say something about Nyquist theorem. Oh! He is only 12.

---

And another person jumping to conclusions without any facts ...

UnderTow

You made some decent technical points, but all we were saying was- maybe you should have prepared your backup arguments and made the post complete from the start rather than just starting a fire, fanning the flame a little, and walking away.

--------------------
Gear list: If you can't find it, grind it

Quote MadManDan:

---

You made some decent technical points, but all we were saying was- maybe you should have prepared your backup arguments and made the post complete from the start rather than just starting a fire, fanning the flame a little, and walking away.

---

Yes well, I was shocked. Anyway, the fire is still burning.

UnderTow

I'd like this thread brought to the attention of the article writer to clarrify the points. I'd hate to think the magazine was printing out inaccurate articles, it's also an interesting subject to be debated.

Carry on

Quote UnderTow:

---

Quote tex:

---

Undertow knows hardly anything. He saw a book once and he heard his teacher say something about Nyquist theorem. Oh! He is only 12.

---

And another person jumping to conclusions without any facts ...

UnderTow

---

Well I wasn't getting any facts from you at the time and I was interested to see the whole argument. Now I have thank you. I didn't say what 12.

--------------------
Success is round the corner. It's also round the bend.

wow...

interesting...

(i'm trying to figure out the correct percentage of that quoted figure)

(i'm a guitar player, it might take a while)

Dear UnderTow,

Your actual arguments are most valid. The article wavers between objective and subjective statements with plenty of inaccurate information. The editor must have had a couple of pages to fill to meet a deadline. Too many dependable staff writers away at shows. What?

Next time for brevity on your part simply pose your argument as a question in the first place i.e 'Nichols on 24 Bit,Wrong or Not?', point us towards the article and the sites for the real facts. Having thrown open the topic to investigation, argument, feedback & discussion here on what is a forum afterall we'd eventually come to our own decision which may concur with your's or not.

Regards to all (never intentionally patronising)
The Producer.

Quote The Producer:

---

Dear UnderTow,

Your actual arguments are most valid. The article wavers between objective and subjective statements with plenty of inaccurate information. The editor must have had a couple of pages to fill to meet a deadline. Too many dependable staff writers away at shows. What?

---

Apparantly Nichols has a monthly blurb in SOS since one or two issues. I found the previous one questionable at best. (Something about harddrives loosing their magnetization if you leave them on a shelf for three months...)

I remain with the notion that Nichols does not have the technical knowledge and understanding to write technical articles for SOS.

Quote:

---

Next time ...

---

Lets hope there isn't a next time ...

Quote:

---

for brevity on your part simply pose your argument as a question in the first place

---

I would if it was a question.

I know my approach was probably counter productive. On the other hand, it might convey my shock and dismay at reading such an article in a magazine I respect.

Quote:

---

Regards to all (never intentionally patronising)
The Producer.

---

Thanks for confirming my technical arguments.

An article about the merits of 24 bits vs 16 bits recording and not a single time is the word dither mentioned. I'm still amazed ...

UnderTow

What an odd rant. So Mr. Nichols (let's be polite, please) thinks long and hard, and comes up with a good, logically consistent theory about bass sounds with 24 bit. And writes an interesting and brain-engaging article.

It may not fit with some learned 'facts' as learnt by an anonymous ranting poster.

'nuff said

G

--------------------
PA stuff on FB

Quote Guy Johnson:

---

What an odd rant. So Mr. Nichols (let's be polite, please) thinks long and hard,

---

That is his job when writting such articles. Unfortunately he forgot one important aspect of his job: checking his facts.

Quote:

---

and comes up with a good, logically consistent theory about bass sounds with 24 bit.

---

He did nothing of the kind.

Quote:

---

And writes an interesting and brain-engaging article.

---

The part of my brain most engaged after reading the article was my amygdala.

Quote:

---

It may not fit with some learned 'facts' as learnt by an anonymous ranting poster.

---

Read those articles in the links I posted.

Quote:

---

'nuff said
G

---

That is the whole problem. Readers of SOS presumably trust SOS to bring them informative articles based on correct technical data. When they fail to do this for whatever reason, they fail as an informative source about Music Recording, and thus they fail at their primary purpouse.

You seem to be blindly believing the article simply because it is printed in SOS which prooves my point and confirms my reasons for concern.

Seriously, read the articles I linked in my first post. That should dispel any ideas that the article in question is technicaly correct.

UnderTow

i take issue with the belief that the extra resolution only applies to low-level signals, this is surely not correct. there are 256 times as many integers available between -FSD and +FSD, so the resolution is available to all signals.

it is of particular value to low-level signals, yes, but is available across the whole range. what an odd argument.

--------------------
planet nine
lincoln, uk.

Quote Nathan:

---

i take issue with the belief that the extra resolution only applies to low-level signals, this is surely not correct. there are 256 times as many integers available between -FSD and +FSD, so the resolution is available to all signals.

---

There are indeed 16777216 values in 24 bit instead of the 65536 values of 16 bit but the extra 16711680 values are all packed down at the bottom of the amplitude scale.

Check out this image of a sine wave: http://home.casema.nl/ajohnston/Bits.png

Look at the dB scale on the right of the image. Now look at the same wave heavily zoomed in verticaly: http://home.casema.nl/ajohnston/Bits-Zoomed.png

All those extra 16711680 values are packed arround the red center line between the green lines that represent -96 dB. They are not evenly distributed over the whole range.

If these extra values were distributed over the whole range, you would need to requantize when converting a 16 bit file to a 24 bit file when in reality the 8 extra bits added are just filled with zeros.

There is even a small section of the Nichols article that supports this:

Quote:

---

Each bit of a 24-bit sample has a different voltage value assigned to it. The voltage value of each bit is supposed to be exactly half the value of the bit above it.

---

So each bit you add to the bit depth of a digital word has an equivalent voltage value of half that of the previous least significant bit.

You can actually test this yourself: Take a 16 bit wave file. Any file will do. Convert it to 24 bits and then convert it back to 16 bits without doing any kind of processing (including any gain change or adding any dither!). The resulting file is 100% identical to the file you started with. Every sample has the exact same value. You can do this as many times as you want. It won't affect the signal in any way whatsoever.

Does this explain things or does it add more confusion?


UnderTow

I'm gonna wait for Hugh to wake up so I can start reading this thread properly.

Night

Quote Music Manic:

---

I'm gonna wait for Hugh to wake up so I can start reading this thread properly.

---

Hugh's gonna wake up to a headache.

Although, I do have to admit that upon reading the article I am inclined to share Under Tow's concerns about Mr Nicols reasoning. Subjects/principles such as dither, Nyquist, and reconstruction seem not to apply...

Mike

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

i've got a headace already...

(still working on that percentage thingy)

har har

Quote UnderTow:

---

There are indeed 16777216 values in 24 bit instead of the 65536 values of 16 bit but the extra 16711680 values are all packed down at the bottom of the amplitude scale.

Check out this image of a sine wave: http://home.casema.nl/ajohnston/Bits.png

Look at the dB scale on the right of the image. Now look at the same wave heavily zoomed in verticaly: http://home.casema.nl/ajohnston/Bits-Zoomed.png

All those extra 16711680 values are packed arround the red center line between the green lines that represent -96 dB. They are not evenly distributed over the whole range.

If these extra values were distributed over the whole range, you would need to requantize when converting a 16 bit file to a 24 bit file when in reality the 8 extra bits added are just filled with zeros.

UnderTow

---

There's a fundemental misunderstanding of digital audio right here. Back to basics:

Lets start with a 2 bit signal (easier), assuming full scale is 1 volt (and ignoring signing for the moment):

00 - 0v
01 - 0.33v
10 - 0.67v
11 - 1v

Add another bit:

000 - 0v
001 - 0.14v
010 - 0.29v
011 - 0.43v
100 - 0.57v
101 - 0.71v
110 - 0.86v
111 - 1v

The extra bit doesn't just add extra resolution at the bottom end of the scale - it adds it throughout. Just because the extra zeros are added to the least significant end of the data doesn't mean that it just represents the quietest part of the converted signal.

I really wouldn't put much weight on what screen grabs of cool edit show you - a decent book (not internet site) on digital audio will explain all this far better than I can. Failing that, I'm sure Hugh will soon dispatch this when he returns.

Sheriton

--------------------
There's nothing we can't face... Except for bunnies

the bit about each less significant bit being half of the value of the one preceding it is just binary. in the same way that each number to the right is a tenth of the value of the preceding one in denary (base10).

you could think of 16-bit as being to the nearest integer, whereas 24-bit uses a few decimal places (this is an analogy, note). the extra resolution is available whatever the number, but yes, it would be more significant for lower numbers because they are smaller. when you convert from 24 to 16-bit and back again, it's like adding those decimal places (or working to them) and then lopping them off again -the number doesn't change if you didn't rescale (they're still "zeros"), the problems arise when you lop off (truncate) the "fractional" bit that is not zero. (end of analogy)

don't forget that those graphs show a dB logarithmic ruler on what i think is a linear scale, hence the dB numbers are going to cram in once you get nearer the zero line. the digital number representing a sample is linear, bottom to top, it represents the voltage not the dB value -just with more resolution (accuracy if you like) if it's a 24-bit sample.

--------------------
planet nine
lincoln, uk.

We've had a lot of comments about this article, and Roger is preparing a response or clarification for a future issue. I'll leave that to him, but a couple of points about UnderTow's claims:

Quote UnderTow:

---

Quote:

---

The first impression is that you are just changing the noise floor from -144dB to -96dB, which isn't going to hurt the punchy high-level mix that you spent so much time on. But there is much more to this conversion than meets the ear.

---

This is incomplete to say the least. You could have a recording with a noise floor way above -96 dB FS. It won't suddenly go down to -144 dB FS just by converting to 24 bits.

---

Roger Nichols is talking about converting from 24-bit to 16-bit. You seem to be talking about converting the other way.

Quote UnderTow:

---

Quote:

---

Just to refresh your memory, 24-bit audio gives you 256 times more resolution in the position of each sample on the waveform.

---

This isn't correct. It only gives more resolution to low-level signals. This can be easily demonstrated by converting a 16 bit file to a 24 bit file. The extra 8 bits are just filled with zeros. There is no requantization of the audio. A (loud) signal which lives in the upper bits of a 16 bit audio file does not gain any resolution by having extra zeros added in the extra bits.

---

I don't think he's claiming that converting a 16-bit recording to 24-bit will magically add resolution; he's simply stating that 24-bit recording is more accurate.

Quote UnderTow:

---

Quote:

---

The first time I heard 24-bit, recording Bela Fleck and the Flecktones, I expected to hear more sheen and high-frequency clarity. I was incorrect. The most noticeable difference was in the low frequencies: the mouthwatering sound of the bass, the breathtaking realism of the kick drum, the clarity of vocals, and the low-end improvements that finally make the banjo worth recording. It took me a while to figure out why this was the case, but I finally got to the bottom of it.

---

This is anecdotal and isn't based on any facts. He seems to base his whole premise on "bad bass" in 16 bit audio on this recording.

---

I think he is just using that record as an example of the kinds of difference he hears in 24-bit recording, not as the sole evidence for his theory. In any case, whether things sound better or worse is inevitably going to be subjective. What other 'fact' could there be about whether the bass sounded better to Roger Nichols on that particular recording?

Quote Sheriton:

---

Quote UnderTow:

---

the extra 16711680 values are all packed down at the bottom of the amplitude scale.

---

00 - 0v
01 - 0.33v
10 - 0.67v
11 - 1v

Add another bit:

000 - 0v
001 - 0.14v
010 - 0.29v
011 - 0.43v
100 - 0.57v
101 - 0.71v
110 - 0.86v
111 - 1v

The extra bit doesn't just add extra resolution at the bottom end of the scale - it adds it throughout. Just because the extra zeros are added to the least significant end of the data doesn't mean that it just represents the quietest part of the converted signal.

---

AFAIK… In the digital domain the interpretation of the bitpattern is:
- bit 1/MSB carries info up to 0dBFS
- bit 2 (that happens to be LSB in example 1) carries info about what happens up to -6dBFS
- bit 3 (that happens to be LSB in example 2) carries info about what happens up to -12dBFS

So, the added bit has added four new values at the bottom of the amplitude range. Logarithmic, not linear. Shouldn't it be so in the analouge domain too?

Lars

This is a clear cut case of having technical abilities, strengths and knowledge but absolutely no people, influencing or presentational skills.

Had the poster of this thread raised the subject in a adult, polite, fascinating way I would have been open to listen but instead I just picture some sad geek sitting with his mathematical formulas screaming "you're wrong you're wrong!".

There is absolutely no excuse for referring to a writer by their surname only - wherther you are a fan or a foe.


(Perhaps there are issues around the writer of this thread wanting to write for music magazines but being unsuccessful).

The key challenge around the accuracy of facts may have been correct but like any spoilt child that starts screaming, we're not so much interested in the noise coming out of their mouth.

So Mr Angry - why not shock the world and apologise for your brattish conduct, then perhaps we can start a sensible debate about 24bit to 16bit conversion ?

--------------------
What it says on the tin...

Interesting fracas! And hopefully everyone can keep cool so we can have an informative discussion which will leave us all older and wiser!

I am not a scientific genius about digital recording; I started life with tape recorders and razor blades. My judgments are based on does it sound good, and does it work reliably time after time without a lot of farting about with the technology getting in the way of making music.

But it seems to me that some, if not indeed much, of Under Tow's argument is based on converting a 16 bit signal to 24 and, lo and behold! - it doesn't sound any better. Well of course it doesn't. You can't magically put back something that has already been taken away, never to return.

However, if you put a really great mic through a great pre-amp into a great AD converter, then what you record will sound a lot better at 24 bit than at 16 bit. And your noise floor is much better, and multi-tracking e.g. drums or even a whole band is easier because you are less concerned about the risk of clipping, because you can track with a lot more headroom.

If you stay in 24-bit land for as long as possilbe (mixing, adding DSP effects, reverbs EQ etc), that is all going to sound better than at 16 bits. If you save "downgrading" to 16-bit till the latest possible stage in the chain (burning to CD), you will have a better recording.

This is what my ears tell me.

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

ok, i'm american so maybe there is a language barrier here...

but what is the big deal about calling the author Bloggs as opposed to Joe Bloggs?

from my limited education, i thought it was common to refer to authors by surname only?

perhaps this is from a different sector?

although you can say Joe Bloggs at first mention and then after just say Bloggs?

anyway, what's the big deal?

its the internet, dude...

lighten up...

i think its time for a group hug...

Well I just use the toys to make a noise... and I'm slowly getting to grips with some of the maths that's actually behind what we do, so I've found this thread very interesting, if a little heated !

For what it's worth, I still record at 16 bit, simply because-

1. I'm essentially quite lazy, and take a while to get 'up to date' unless my clients are complaining (and they're not...)

2. Given that most people seem to be listening on a lossy format anyway(mp3) I'm more concerned with the 'meat' of the tune- melody, chords, feel, etc, not the 'sound', as long as it's adequate- (tho' I do understand that having a 'better' recording in the first place might improve the final compressed result).

3. To my ears, my 16 bit productions sound just as 'good', and very often (to bang my own drum for once) a helluva lot better (sound-wise) to a lot of modern productions which I can only assume were recorded at 24 bit...


With all due respect, folks... storm in a teacup ?


tomafd

--------------------
http://anotherfineday.bandcamp.com/ http://anotherfineday.co.uk http://apollomusic.co.uk

Plenty of successful recording engineers have blind spots when it comes to technical details. But equally it appears that Undertow has a few blind spots too. Undertow - I would suggest you read and digest a decent text book on digital representation of analogue signals before starting a thread like this. There are some very fundamental flaws in your reasoning - the major one being that the increased resolution affects the whole dynamic range - not just the quiet parts.

Cheers

James.

--------------------
JRP Music - Audio Mastering and Restoration.
http://www.jrpmusic.net

Lars, the maths gets a bit offset here because so few bits are being used, the more bits you add, the closer binary1000... gets to 0.5 - and don't forget that the sample has to represent positive and negative excursions, so here B100 would be the zero line (tho I believe that for 16 bit ADCs the twos-complement method is used, where the MSB represents the polarity of the sample).

it depends what you mean by "cariies info up to", but bit 1 represents about halfway up the scale, hence -6dBFS, bit 2 6dB less. remember that bit3 is toggling (and therefore representing numbers) all the way thru this table, it might represent -24dB changes, but thats from 0 to full scale.

--------------------
planet nine
lincoln, uk.

UnderTow Quote:

---

Just to refresh your memory, 24-bit audio gives you 256 times more resolution in the position of each sample on the waveform.

---

Quote UnderTow:

---

This isn't correct. It only gives more resolution to low-level signals. This can be easily demonstrated by converting a 16 bit file to a 24 bit file. The extra 8 bits are just filled with zeros. There is no requantization of the audio. A (loud) signal which lives in the upper bits of a 16 bit audio file does not gain any resolution by having extra zeros added in the extra bits.

---

I don't particularly want to defend the author of the original article, but your contradiction here isn't strictly correct either. Quantising to 24 bits theoretically provides 256 times more resolution than quantising to 16 bits because the amplitude differences between quantising levels is smaller. So the statement as quoted is correct.

However, I also take your point that simply placing a 16 bit file into a 24 bit system will not change the quantising resolution of the original file (unless you also perform some processing on that signal, in which case the result will be output with true 24 bit resolution).

UnderTow Quote:

---

This 256 times higher resolution of a 24-bit sample is in effect everywhere on the waveform, from the lowest levels to the highest peaks. A sample point nearing 0dB full scale is 256 times more accurate than the same sample recorded at 16-bit.

---

Quote UnderTow:

---

This is incorrect. The resolution is only added at the lower end of the amplitude scale.

---

Again, not strictly true. Quantising with more bits reduces the level of the noise floor and allows quieter signal to be heard more easily, but the resolution is actually improved across the entire dynamic range.

Think of it this way. You have an analogue signalto quantise that runs between 0mV at its quietest and 1000mV at its loudest. If you quantise with 16 bits, you effectively have to draw a grid over the audio waveform with 65536 equally spaced horizontal lines, each line representating the level of a valid quantising value.

If you quantise with 24 bits, that grid has to contain 16,777,216 lines -- in the same space. We don't only add the extra 16,711,680 lines underneath the bottom of the 16 bit grid! So yes, the resolution across the entire dynamic range is improved. We can define the position of a sample peak with considerably greater resolution regardless of its particular amplitude. That is why the quantising error is smaller, and that's why the noise floor is lower.

Quote UnderTow:

---

But nearly every soundcard you buy these days uses oversampling. I might be wrong but he seems to be looking at antiquated technology and basing a whole article on that. This just confuses things.

---

He seems to be talking about very simple delta-sigma converters, and while he right about the principle advantages, the practical application is a little different and there are also some inherent problems. There's not such thing as a free lunch...

It certainly was a, shall we say, interesting article...

hugh

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

yes but
"The most noticeable difference was in the low frequencies: the mouthwatering sound of the bass, the breathtaking realism of the kick drum, the clarity of vocals, and the low-end improvements that finally make the banjo worth recording."

MR NICHOLS IS ACTUALLY RUSS ANDREWS, AND I CLAIM MY FIVE POUNDS

"mouthwatering bass" indeed

--------------------
~~~ A weasel hath not such a deal of spleen as you are tossed with! www.feline1.co.uk ~~~

And on a more serious topic by far, the technology has yet to be invented which would "finally make the banjo worth recording"

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

Quote UnderTow:

---

There are indeed 16777216 values in 24 bit instead of the 65536 values of 16 bit but the extra 16711680 values are all packed down at the bottom of the amplitude scale.

---

I haven't bothered to check out those sites, but on an amplitude scale, yes, it might appear that way. However, this is not strictly relevant to the quantising process.

What the quantiser is trying to do is define the amplitude of a sample. With a 16 bit quantiser the 65536 available scale points are spaced out equally across the full dynamic range. If you switch up to 24 bit quantising, you effectively add a further 256 new levels between each and every line onthe original 16 bit scale.

It's a bit like measuring the size of your living room for a new carpet. If you measure the room to the nearest metre and cut the carpet to that figure the measurement is pretty crude and you might end up with lots of carpet curling up over the skirting boards, or a big gap around the edge of the room! Measure the room to the nearest millimetre and cut the carpet accordingly, and the fit will be virtually perfect.

If you check out a tape measure, I don't think you'll find all the millimietre markings squeezed in under the first metre mark and none above it... They are evenely distributed along the entire length of the tape.

Quote UnderTow:

---

If these extra values were distributed over the whole range, you would need to requantize when converting a 16 bit file to a 24 bit file when in reality the 8 extra bits added are just filled with zeros.

---

Again, a misunderstanding on your part. Going back to my carpet analogy, I can quite happily work with figures in whole metres on a ruler marked in metres and millimetres without any trouble.... I don't have to do anything to the whole metres figure before I can use it... but it clearly doesn't make use of the precision available in a ruler with millimetre markings.

For example, we might have measured the carpet to the nearest metre as being 4 metres long. When measured to the nearest millimetre it might have been 4.236 metres.

The last number has the equivalent (for the sake of this argument) resolution as a 24 bit quantise, while the 4.000 figure is the equivalent of the 16 bit quantise. Notice that there are three zeros appended to metre-only figure when denoted with the same potential resolution as my millimetre scale.

Quote UnderTow:

---

There is even a small section of the Nichols article that supports this:

Quote:

---

Each bit of a 24-bit sample has a different voltage value assigned to it. The voltage value of each bit is supposed to be exactly half the value of the bit above it.

---

So each bit you add to the bit depth of a digital word has an equivalent voltage value of half that of the previous least significant bit.

---

Again, what Nichols says in this quote is technically correct. Your interpretation is wrong.

Each additional bit you add to the wordlength allows you to count twice as many quantising levels. Which means the space between each quatisation level is halved if you asdd an extra bit to the word length. Which means the accuracy of measuement has doubled, and the potential error is halved... which is why the noise floor (noise caused by random errors in the quantising measurement) halves (drops by 6dB) for each additional bit added to the wordlength.

hugh

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

QED?

Quote Hugh Robjohns:

---

It certainly was a, shall we say, interesting article...

hugh

---

I think misleading would be a better word. How you can discuss this subject without mentioning dither is beyong me. And surely the 'bass benefits more' idea is just plain wrong - the noise floor is reduced for all frequencies equally (until you get into shaping anyway).

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


Edited by cc. (27/04/06 10:40 AM)

yeah i think that other dude made the most valid point...

banjo? are you serious?

Quote:

---

ok, i'm american so maybe there is a language barrier here...

but what is the big deal about calling the author Bloggs as opposed to Joe Bloggs?

from my limited education, i thought it was common to refer to authors by surname only?

perhaps this is from a different sector?

although you can say Joe Bloggs at first mention and then after just say Bloggs?

anyway, what's the big deal?

its the internet, dude...

lighten up...

i think its time for a group hug...

---

Yo Gerard hey hows it hanging dude,

It's a question of common courtesy and hey here's always hoping that this will be one of the few forums that doesn't get completely Americanized (but that's a discussion for another day). Using just a surname could be done as an appropriate, good gestured reference but in this case it clearly wasn't, read the opening post again, he makes it personal even calling for the person to be sacked from their job!! - That's the point Dude / Homie.

I'm not defending the author of the SOS article or his views - I don't know him from Adam... It just pee's me off to see people who are just interested in starting a flaming session... all it takes is some manners and respect for people... And surely theres no distinction there between British and American communication.

Rant over..



--------------------
What it says on the tin...

Quote Lars Farm:

---

AFAIK… In the digital domain the interpretation of the bitpattern is:
- bit 1/MSB carries info up to 0dBFS
- bit 2 (that happens to be LSB in example 1) carries info about what happens up to -6dBFS
- bit 3 (that happens to be LSB in example 2) carries info about what happens up to -12dBFS

So, the added bit has added four new values at the bottom of the amplitude range. Logarithmic, not linear. Shouldn't it be so in the analouge domain too?

Lars

---

That would be a misunderstanding on how binary works. As most digital audio is stored in floating point values between 1 and -1 binary allows that number behind the decimal to be much more accurate, but in its most basic form binary describes any number in a set number of on/off signals.

Meaning in a 8 Bit/1 Byte number which can be up to 255, the first digit describes wether or not the number is at least 128, the second describes wether there is another 64 added onto that, the third describes wether there is another 32 added onto that etc. This allows numbers such as 129 to be possible by having a 1 in the first digit and last digit. So therefore it allows for more accurate depth description along the entire scale if you have a limited range, and larger number if dealing with integers(Well ever integer from 0 on up) with a non limited range.

So if most digital audio is dealing with floating point numbers between 1 and -1, increasing from 16 to 24 bits allows you to get the same dynamic range out of it overall, just much more accurate of one, which does allow for extended descriptions into the quieter range, but also extended descriptions between the limit of 65535 of a 16 bit number, and slightly quieter which in 16 bit would be 65534, but in 24 bit would be some number I cant think of right off hand that is 65534.999something. That can be a significant difference when dealing with a scale as large as the dB scale for our hearing as we hear a fairly large range of differences in sound.

But in other words, it doesnt just add the ability to hear quieter sounds, it allows for more accurate representation of what those quiet sounds actually are in terms of levels, allowing your noisefloor to shift from 0000000000000001 to 000000000000000000000001. This affects your entire scale if a signal is recorded in 24 bit allowing for much more 'depth' to your recordings. Still doesnt change the voltage value for the equivalent of every digit being 1 in a 16 bit or 24 bit recording, both of those will still be the exact same.

Hope that made sense at all.

Seablade

Not very good at explaining what goes on in his head;)

NOTE: Not all Digital audio deals in floating point numbers, it really depends on how a program was coded, but in general I find it better and it does seem to be a majority of software I deal with deals in it. It could easily be defined as a range between -32767 and +32767

NOTE 2: In many of my examples above to keep things simple I used unsigned bit numbers, however technically it owuld be a signed bit number as I noted, so one of those bits would be dedicated to identifying wether the signal was positive or negative. I cant remember right off hand which one, however it isnt exactly important unless you deal with the code on it;)