A1MixMan

John

And 96kHz extends the dynamic range. People will argue whether "humans" can hear it, but it goes beyond that. Nature doesn't stop at 44.1 kHz, so why should I?

No, there is no extension of dynamic range. There is a bandwidth increase. Two very different things.

Whoops, you're right. I meant to say 24bit extends the dynamic range. That is correct, right?

Correct. Sort of. 24 bits gives one more steps in the dynamic range. The loudest sound and the lowest sound could be the same with 16 bits. You just have more levels in between with 24 bits. Here the graphic analogy works somewhat.  One could say more shades of grey in a bitmap with 24 bits and less with 16 bits. Sample rate has no analogy in the graphics domain though.

One important fact is that 0 db is 0 db whether its a 16 bit file or a 24 bit file.

UnderTow

rabeach

UnderTow

rabeach

The Shanon-Nyquist theorem says you only need twice the audible bandwidth to perfectly reproduce any audible signal.

That is absolutely not what it says.

I love these statements with absolutely no backing.

From http://en.wikipedia.org/wiki/Nyquist%E2%80%93Shannon_sampling_theorem :

"the theorem shows that a bandlimited analog signal that has been sampled can be perfectly reconstructed from an infinite sequence of samples if the sampling rate exceeds 2B samples per second, where B is the highest frequency in the original signal."

UnderTow

And now you see your mistake. You should really try to post factual information.

It IS factual. I wrote "The Shanon-Nyquist theorem says you only need twice the audible bandwidth to perfectly reproduce any audible signal." I could have left out the word audible but that would not have made it any more factual. If something is true for ALL signals, then it must be true for a subset, the audible subset, of those signals.

Do you really not understand that?


UnderTow

"if the sampling rate exceeds 2B samples" are you really not able to comprehend that.

jyeager11

UnderTow

Or because we understand how sampling works. There is no such thing as resolution in audio.

Umm, you understand what a metaphor is, right? Like, when someone says "easy as pie", there isn't always pie involved? That's what we're doing here when we're talking about resolution.

It is a bad metaphor.

it is better to record and mix in 24b than it is in 16b, even if the final output will be 16b.

Yes it is better to record in 24 bits. I have agreed with that a long time ago if you would have been paying attention...

UnderTow

Loptec

Freddie H

Loptec

best possibl

Freddie H


Okay boys and girls are you all ready for the truth?



Okay I use 24 bit, 32bit, 64bit 48kHz or 96kHz and all kinds of dithering and hoghend AD DA converters too etc......but can we hear the difference...?
You can't hear any dithering go on at all , infact you can't even hear the difference between 12-14 bit or 16 bit whatever...
Still we should always use the best possible quality we have..but you should be aware of the facts why and what's real or not..



Watch this Video and you will learn alot of all kinds AUDIO myths floating around out there...
http://www.youtube.com/watch?v=BYTlN6wjcvQ

Sorry I can’t agree with this.
You hear a huge difference just by changing from 16bit 44..1kHz to 24bit 48kHz!
And it’s not just what you hear in the basic recording it’s what you can do with the sound afterwards..
It’s like editing an image..

Let’s say you have a HUGE photo of a flower with a fly sitting on it.

* If you have great resolution you can zoom in and draw a hat on the fly.

* With low resolution all you can do is putting a blob in the fly’s head and say.. “well.. it kind of looks like a hat”

In other words you get SO MUCH more control over the sound when working with higher resolution. And even if you compress the audio to mp3 or whatever after, the final product sounds SO much better if it wasn’t all blurry from the start

My friend... I'm with you.. I'm not saing that...
Its Sunday take your time watching the VIDEO I posted.. I think you will like it. It it will take one hour to watch the Video.

Haha.. Yeah, I'll do that ..but later :)
First I think I'll get out in the sun for a bit.. :) ..If I'm not mistaking you're from sweden too, yes? Then you know that if you
want to get any sun before winter you'd better be fast! :)

Also, I didn't see the line "We should always urge and use the best possible quality we have..but you should be aware of the facts why and what's real or not.." in your message before. I couldn't agree more with this! :)

True my friend! .....By the way...  just a friendly advice...no point trying to explain your point on this FORUM... let the people get to their own conclutions...even though later----> their own conclutions will be perhaps exactly the same you are trying to tell them...
 
Think about it!

UnderTow

jyeager11

Umm, you understand what a metaphor is, right? Like, when someone says "easy as pie", there isn't always pie involved? That's what we're doing here when we're talking about resolution.

It is a bad metaphor.

YOU'RE a bad metaphor.

it is better to record and mix in 24b than it is in 16b, even if the final output will be 16b.

Yes it is better to record in 24 bits.

For reasons very well illustrated by Loptec's metaphor.

The end.

jyeager11

John

Correct. Sort of. 24 bits gives one more steps in the dynamic range. The loudest sound and the lowest sound could be the same with 16 bits. You just have more levels in between with 24 bits.

Hence Loptec's use of the word RESOLUTION. And it is a spot-on analogy.

Here the graphic analogy works somewhat. 

Not "somewhat", John. It just works. It is the very definition of resolution.

Notice though I am talking about BIT DEPTH not sample rate. So don't think I am agreeing with you.

rabeach


"if the sampling rate exceeds 2B samples" are you really not able to comprehend that.

I do comprehend but you clearly don't. It means you need a sampling rate that is at least twice the highest frequency you want to sample. So if you want to record frequencies up to 20Khz, you need a sampling rate of at least 40Khz.

What did you think it meant?

UnderTow

John

Notice though I am talking about BIT DEPTH not sample rate. So don't think I am agreeing with you.

Do you even know what my argument is at this point, John?

I made no distinction between bit depth and sample rates. I'm stating that Loptec's metaphor is a good one for understanding why higher resolutions during production is always a better idea than working strictly in what the final output will be. I did not go more in-depth than that. It's people's blanket objection to the metaphor that I have a problem with.

More levels between two points is the very definition of resolution, and it's a wonderful analogy for anyone who doesn't understand why he should be recording and mixing in anything except 16 bits / 44.1 kHz because that is what the listener will be listening to.

As to John's quote above. It is true in the strictest sense in a properly dithered system but one has to realise that the noise floor of a 16 bit system will be higher than that of a 24 bit system. That means that although you can indeed use a 16 bit system to record signals that are as low as that of a 24 bit system but those signals will be buried in noise.

True.

I was trying to avoid the noise floor issue. LOL

UnderTow

rabeach


"if the sampling rate exceeds 2B samples" are you really not able to comprehend that.

I do comprehend but you clearly don't. It means you need a sampling rate that is at least twice the highest frequency you want to sample. So if you want to record frequencies up to 20Khz, you need a sampling rate of at least 40Khz.

What did you think it meant?

UnderTow

Correct! True in theory... He also saying there are benefits using more then 44.1 kHz. 24bit infact 48kHz or higher are need of capture overtones and oversampling...example EQ filters and so on...1976 there were nothing like CD:s or MP3 what ever...
 
Remember Nyqvist died 1976 so what he had to say about all this today would be probably a lot different...
 
http://en.wikipedia.org/wiki/Harry_Nyquist

UnderTow

rabeach


"if the sampling rate exceeds 2B samples" are you really not able to comprehend that.

I do comprehend but you clearly don't. It means you need a sampling rate that is at least twice the highest frequency you want to sample. So if you want to record frequencies up to 20Khz, you need a sampling rate of at least 40Khz.

What did you think it meant?

UnderTow

It doesn't mean that at all. And clearly you don't comprehend. It means the sampling rate must exceed twice the highest frequency you want to sample.

jyeager11

John

Notice though I am talking about BIT DEPTH not sample rate. So don't think I am agreeing with you.

Do you even know what my argument is at this point, John?

I made no distinction between bit depth and sample rates. I'm stating that Loptec's metaphor is a good one for understanding why higher resolutions during production is always a better idea than working strictly in what the final output will be. I did not go more in-depth than that. It's people's blanket objection to the metaphor that I have a problem with.

More levels between two points is the very definition of resolution, and it's a wonderful analogy for anyone who doesn't understand why he should be recording and mixing in anything except 16 bits / 44.1 kHz because that is what the listener will be listening to.

This is the problem. You should make a distinction between them to understand them.  I know how film works and I also know how digital imaging works. Although it seems that digital audio and digital imaging have a lot of things in common they really don't. There is no equivalent to sample rate what so ever in digital imaging. You should know that.

The other problem is your misunderstanding about "More levels between two points is the very definition of resolution" This only applies to bit depth not sample rate. Getting that understood will get you on the road to clarity.

By confusing the two you are not understanding what they really are. I know it doesn't seem that is the way it is but it is.

Freddie H

Correct! True in theory... He also saying there are benefits using more then 44.1 kHz.

Actually the theorem doesn't. It is a mathematical theorem and as such is always theoretical. The 44.1Khz is an engineering decision based purely on the equipment available at the time. (U-Matic video recorders. You can read all about it here: http://www.exp-math.uni-e...mink/pdf/beethoven.htm )

24bit infact 48kHz or higher are need of capture overtones and oversampling...example EQ filters and so on...1976 there were nothing like CD:s or MP3 what ever...

Actually EQs and filters don't need any oversampling. Non-linear processing like compression can benefit from oversampling. As for overtones, they are only needed if they can be heard.

  Remember Nyqvist died 1976 so what he had to say about all this today would be probably a lot different...
 
http://en.wikipedia.org/wiki/Harry_Nyquist

He would be kicking and screaming at all the advertisement ****! But you know that this sampling theorem is used in all sorts of fields? Telecommunications, radar, sonar (the ones on boats and submarines I mean ;-), video etc etc etc. They all follow the same principles set out by Shanon and Nyquist many years ago. It is ONLY in audio that we have these ridiculous products that offer things like 384 Khz sampling.

The reason for these ridiculous products is because of a lack of understanding and the effect of placebo and expectation bias. It is easy to fool people with this. (Everyone has experienced tweaking an EQ so that it sounds just right to then discover it was in bypass all the time...) The people that are most sure about what they hear are the easiest to fool because they are ego driven and don't have a healthy amount of self-doubt.

UnderTow

Freddie H

FastBikerBoy

 I'll go for some... Do do you want Coke or Sprite my friend? 

Sprite for me please. I've got to pop out for a while. I'm sure this'll be up to 7 pages by the time I get back.

UndertowYou only need two points to fully define a circle. (The centre and any point on the diameter). Adding more points does not in anyway define the circle better. The same thing goes for sampling.

 
Hmmm... sounds familiar:
 
http://forum.cakewalk.com/fb.ashx?m=1865385