Did you know that you probably just are listening to 19 or 18 bit sound?

[Boray]

Did you know that you probably just are listening to 19 or 18 bit sound?

It seems to me that the monitor level knob just is a digital level, just as the master fader. So this volume is probably just adjusted by recalculating the volume digitally after the master fader. But what does this mean in practice? Dividing a digital value by 2 is performed by shifting all the bits one step to the right, leaving the first bit unused. So if you set your monitor level at 50 and have 20 bit converters, then you now will listen to at most 19 bit sound. Here is a little table:


Monitor level: (And when having your master level at 100)

127-101 Should distort the sound if you have mixed optimally.

100-51, You are AT MOST listening to all of your bits! On the VS1680, this is 20 bits.

50-26, You are AT MOST listening to all of your bits MINUS 1, on the VS1680 this is 20-1=19 bits.

25-13, You are AT MOST listening to all of your bits MINUS 2, on the VS1680 this is 20-2=18 bits.

12-7, You are AT MOST listening to all of your bits MINUS 3, on the VS1680 this is 20-3=17 bits.

6-4, You are AT MOST listening to all of your bits MINUS 4, on the VS1680 this is 20-4=16 bits.

3-2, You are AT MOST listening to all of your bits MINUS 5, on the VS1680 this is 20-5=15 bits.

1, You are AT MOST listening to all of your bits MINUS 6, on the VS1680 this is 20-6=14 bits.



And as if that isn't enough, this table also applies to all other faders on the VS! (At least roughly). So if you have a tracks fader at 23%, the master fader at 79% and the monitor knob at 41%. Then take 100*0.23*0.79*0.41 = 7.45... According to the table, you can at most hear 17 bit sound from that channel. Chanses are however that you didn't use all 20 bits when recording it, so it's probably more like 16 bit sound. (If all other channels are quiet). If you have another track playing at level 100 in the same time, then these two would be mixed together... So if you want to know how many bits you are listening to, you should take the track sounding with the highest volume and calculate it in the same way.

It doesn't really matter what mixing algorithm that is used for this table to work. If a sound is of a specific volume (percentage) on a digital bus, then it simply works this way.

/Anders

[Blue Bear Sound]

...(and understand) John Watkinson's "Art of Digital Audio" yet (600+ pages)???

Fuck off with the bullshit until you do.


As usual, you've got a snippet of correct info taken out of context and mixed it in with complete tripe.......

[Son of Mixerman]

Most schematics of digital mixers Ive seen do not do bit depth changes real time. Most folf know that a fader is a type of amplifier, and it changes the volume and nothing more and nothing less. According to your theory, faders are converter channels that are constantly doing bit conversions as you move the fader. In reality that would be too mathmatically intensive for any digital mixer to pull it off, that would be the equivalent of running 16 effect buss' real time on top of the effects on the 2 buss and whatever your running on each channels for sends and returns... WHOA! Try looking up what a fader is, and how they work. Just my 2 bits!


SoMm

[Boray]

Originally posted by Blue Bear Sound
...(and understand) John Watkinson's "Art of Digital Audio" yet (600+ pages)???

Fuck off with the bullshit until you do.


As usual, you've got a snippet of correct info taken out of context and mixed it in with complete tripe.......

What was it you didn't understand? My post or that book? If it was my post, then what part of it was it?

[Boray]

Originally posted by Son of Mixerman
Most schematics of digital mixers Ive seen do not do bit depth changes real time. Most folf know that a fader is a type of amplifier, and it changes the volume and nothing more and nothing less. According to your theory, faders are converter channels that are constantly doing bit conversions as you move the fader. In reality that would be too mathmatically intensive for any digital mixer to pull it off, that would be the equivalent of running 16 effect buss' real time on top of the effects on the 2 buss and whatever your running on each channels for sends and returns... WHOA! Try looking up what a fader is, and how they work. Just my 2 bits!


SoMm

If you have a sound on a 24 bit bus that uses 25% of the dynamics (=25% of the total volume), then 2 of the 24 bits are unused (zeros). You don't have to do any bit depth convertions for that. In the actual mixer, there could be floating points bla bla bla used etc, but on the digital bus that the D/A is connected to, this is facts!!! It still is 24 bit, but with the example, only 22 bits of it are actually used and heard. With a lower volume, you are not using all of the dynamics of a D/A - meaning - you are not using all of it's bits.

[Son of Mixerman]

Can someone please explain to me how not using all the bits isn't a bit conversion? Not using bits that are originally available smells like truncation, which is commonly used by Roland, Akai and Korg. It sounds as if your D/A converters are truncating down to 16 bit. Alot of your DAW's like the VS series use non-linear converters anyhow, so your probably only getting 18 bit at best and a 2 bit truncation on the master.
Volume and dynamics are different are they not?

Boray, sounds like you need a change in scenery. Meaning...Dump the VS stuff. Try moving up a Yamaha Aw4416. Its uses 24 bit linear converters and you have the option of truncation or dithering using one of the Waves add on cards. Even some of the best Mastering Engineers have been saying the 4416 has provided the highest quality digital end products of all the DAW's available. Or, get one of the newer Hardisk recorders from Alesis or Mackie. Seriously. When I was considering a DAW I did alot of research of the documentation and the internal circuitry. The Yamaha was no different, but they fixed the truncation problem. Some of the Rolands added some fixes, but a majority of the VS made compromises with the converters. Don't get upset, I bought an MD8 from Yamaha and whew....not the smartest purchase. Ive regretted it quite abit.
Marketing schemes!

SoMm

[Boray]

Originally posted by Son of Mixerman
Can someone please explain to me how not using all the bits isn't a bit conversion? Not using bits that are originally available smells like truncation, which is commonly used by Roland, Akai and Korg. It sounds as if your D/A converters are truncating down to 16 bit. Alot of your DAW's like the VS series use non-linear converters anyhow, so your probably only getting 18 bit at best and a 2 bit truncation on the master.
Volume and dynamics are different are they not?

Boray, sounds like you need a change in scenery. Meaning...Dump the VS stuff. Try moving up a Yamaha Aw4416. Its uses 24 bit linear converters and you have the option of truncation or dithering using one of the Waves add on cards. Even some of the best Mastering Engineers have been saying the 4416 has provided the highest quality digital end products of all the DAW's available. Or, get one of the newer Hardisk recorders from Alesis or Mackie. Seriously. When I was considering a DAW I did alot of research of the documentation and the internal circuitry. The Yamaha was no different, but they fixed the truncation problem. Some of the Rolands added some fixes, but a majority of the VS made compromises with the converters. Don't get upset, I bought an MD8 from Yamaha and whew....not the smartest purchase. Ive regretted it quite abit.
Marketing schemes!

SoMm

You know about binary values, right? With normal ("10-based") values, for every new digit you add, you will get 10 times as many different values to use... For example, with one digit, you can represent 10 different values (0-9), with two digits you can represent 100 different values (0-99). With binary values, you will get only twice as many values to use for every new bit... With one bit, you can represent 2 values (0-1), with two, you can represent 4 values (0-3), with three bits, you can represent 8 values (0-7) etc....

You know about samples, right? If you want to simplify, then you could say that digital recording in principle is to store volume values really fast. A waveform is "volume" or "waveform amplitude" changing over time. So if you record a sound quite low compared to how loud it could be recorded (without getting distorted) then it will result in lower values stored.

So if you record something as loudly as permitted (before distorting) using 24 bits would result in binary values like:

100101011010010110100111
010010100100101110101100
011101010001010011010011
110100101101001101001010
010001011101001101001011

Recording the same sound at a lower volume (at 25% of the volume of last recording) would result in:

001001010110100101101001
000100101001001011101011
000111010100010100110100
001101001011010011010010
000100010111010011010010

As you can see, two bits are always zero at the beginning of every 24 bit word, meaning only 22 bits are needed to represent the sound. Just as I described in the beginning, every new bit doubles the number of values that you can use. As you now don't need all those values to represent the sound (just because those high amplitude changes now not are that high any more), then the two first bits will remain zero.

Lets now say that this sound is played and you have the master fader at 50% and the tracks fader at 100%, then this is what is output to the converters:

000100101011010010110100
000010010100100101110101
000011101010001010011010
000110100101101001101001
000010001011101001101001

Another zero at the left. Now let's say the converters are 20 bit... THIS is where the truncation takes place (4 bits are cut away from the right):

00010010101101001011
00001001010010010111
00001110101000101001
00011010010110100110
00001000101110100110

20 bits are outputted, but only 17 of them are sounding.

This example is for positive values. If both positive and negavive values are used (and they probably are), then the first bit indicates if it's a positive or negative value. The result is the same, but for the rest of the bits (not the sign bit).


Most converters tries to be linear, but better quality converters does a better job of it. I doubt Roland would say that their converters are non-linear. There is btw dithering in the Mastering toolkit of the VS machines.... (not the 840, but the rest...)

But this whole deal has nothing to do with how good the mixer of a DAW is! If a sound is played over a n bit D/A converter with 50% of the volume that it could have been played with before distorting - then one of the bits of the converter is always 0 - meaning, that only 50% of the converter's dynamic range is used - meaning that the sound just as well could have been played with a n-1 bit converter. That's the laws of maths.

About truncation and dithering: What everyone seems to mix up is that they think dithering means converting from 24 to 16 bit. That's not true. You get full 16 bit sound just by cutting the last 8 bits off from the 24 bits. This is what happens when you burn a CD, even when you dither! Dithering adds a noise to the 24 bit sound - that's all! Really! ...The cool part is that this noise makes that the 16 bits that will be left after your CD burning will have a higher dynamics than that of 16 bits.... Maybe like 17 or 18 bits..... So when you set your mastering effect to "dither to 16 bits", then random numbers (noise) are added to the sound, numbers as big as they carry parts of the sums over to the bits that will be left. Try to set it to 8 bit and you will hear this really loud noise because now really big random numbers are added to the sound. At first, I also thought that "dithering" ment converting... But it doesn't....

Truncation is (according to maths) a linear conversion. A sound with dithering added to it before it's truncated is not a mathematically linear conversion just because additional noise now is added to it.

When truncating from 24 to 16 bits, the first 256 values (0-255) in the 24 bit sound becomes 0 in the 16 bit sound. 256-511 becomes 1, 512-767 becomes 2, etc, etc... A linear conversion of 256 steps for each 16 bit number. You can test this by take any number from the 24 bit range (0 to 16.8 million), calculate how many percent of the total it is, convert it to binary, cut the last 8 bits off, convert it back (or simply divide the decimal value by 256), then calculate how many percent this value is of the 16 bit range (0 to 65535). You will get the same percent from both.

Not that I mean that dithering is bad, I use it all the time!!!

/Anders

[TexRoadkill]

Like most users I monitor through an amplifier that is hooked up to analog outs on a digital recorder. Why would adjusting an analog volume send effect the bit depth?

Your whole argument assumes that DAW's adjust the volume before DAC which may be true for some all in one boxes but if you use one of those boxes you probably aren't all that concerned with quality anyway.

[Blue Bear Sound]

Anders.... you're completely pathetic....

You don't even realize what it is you don't know.............. and even what you do know, you get wrong!

Hopeless...............

[Boray]

Originally posted by TexRoadkill
Like most users I monitor through an amplifier that is hooked up to analog outs on a digital recorder. Why would adjusting an analog volume send effect the bit depth?

Your whole argument assumes that DAW's adjust the volume before DAC which may be true for some all in one boxes but if you use one of those boxes you probably aren't all that concerned with quality anyway.

That is correct! That is exactly what my post is about! Having powered monitors, you can't easily change the amplifiers volume. And as it seems like the monitor knob of the VS indeed is just another digital "fader" before the D/A, this problem is the result. Also check the VSPlanet thread: http://www.vsplanet.com/ubb/ultimate...c&f=1&t=013032

/Anders

[Boray]

Originally posted by Boray
000100101011010010110100
000010010100100101110101
000011101010001010011010
000110100101101001101001
000010001011101001101001

This example is a bit invalid, but it shows what I mean anyway... It is possible to store a waveform with just positive values. It's just to design the converters to have their 0 volts offset at half of the dynamic range (at 8.4 million for a 24 bit converter). I think there even are "unsigned" audio data files.... But anyway.... Then one specific bit would not be zero like in that example. With the signed approach it would. The point is that when playing something at 50% then only half of the dynamics are used regardless of how the bits are stored. If the converters are of the signed type, then one specific bit would remain unused/zero.

/Anders

[Boray]

Originally posted by Blue Bear Sound
Anders.... you're completely pathetic....

You don't even realize what it is you don't know.............. and even what you do know, you get wrong!

Hopeless...............

I wrote my first digital recording/sampling program in 1988 or 1989. When was that book you talked about written? I even predicted that it would be possible to play sound with a higher resolution than 8 bit on the Amiga computer by adding the 4 channels to two. I talked about this several years before the first programs appeard that actually did this trick. Maybe your problem is that you think that you know more about digital recording than you actually do?

/Anders

[Blue Bear Sound]

Originally posted by Boray
I wrote my first digital recording/sampling program in 1988 or 1989.

I can just imagine how accurate THAT was...

Look you flake - every single thing I've ever seen you post has been questionable at best. And this recent post takes the cake - what a load of shit you've managed to spew up all over this nice clean website..........

Incidently, John Watkinson's credentials need no proof - he is a well-known, international audio/video recording consultant... you - clueless, on the other hand, have absolutely ZERO credibility - as you keep proving over and over again with your bullshit posts..........

[Boray]

Originally posted by Blue Bear Sound
I can just imagine how accurate THAT was...

Look you flake - every single thing I've ever seen you post has been questionable at best. And this recent post takes the cake - what a load of shit you've managed to spew up all over this nice clean website..........

Incidently, John Watkinson's credentials need no proof - he is a well-known, international audio/video recording consultant... you - clueless, on the other hand, have absolutely ZERO credibility - as you keep proving over and over again with your bullshit posts..........

You who know so much, please point out to me what is not correct in this thread (except for what I have pointed out myself). You are very welcome to quote that book if you like. I have probably missused a term or so, but if I am absolutely wrong about something, then I would really like to know about that. I'm not joking. So if you see something wrong, then just point it out and explain how it really works. If you just say "that's all bullshit", then I will never learn, if I'm incorrect that is.

About saying that sound is "volume changing" is not the correct term of course. It's a wave in the air. But volume and the waveforms amplitude is different sides of the same coin. An analog volume control that is controlled digitally is almost exactly the same thing as a D/A converter. Both sets an analog level digitally. The only way to play samples on really old computers is to play the sample through the volume control of the soundchip. It wasn't intended this way, but it works...

/Anders

[TexRoadkill]

Originally posted by Boray
An analog volume control that is controlled digitally is almost exactly the same thing as a D/A converter. Both sets an analog level digitally.

A digital control over an analog circuit is not the same thing as a DAC. Resolution of volume steps would be limited to the resolution of the digital control but that has nothing to do with DAC bit depth.

[Blue Bear Sound]

Originally posted by TexRoadkill
A digital control over an analog circuit is not the same thing as a DAC. Resolution of volume steps would be limited to the resolution of the digital control but that has nothing to do with DAC bit depth.

Exactly............

'nuf said...........

[Boray]

Originally posted by TexRoadkill
A digital control over an analog circuit is not the same thing as a DAC.

Not exactly the same thing, but both sets an analog level digitally.

Originally posted by TexRoadkill
Resolution of volume steps would be limited to the resolution of the digital control but that has nothing to do with DAC bit depth.

That has very much to do with bit depth. An 8 bit volume control can set an analog level to 256 different values. So can an 8 bit D/A.

[Boray]

Originally posted by Blue Bear Sound
Exactly............

'nuf said...........

Was that all you had to say? So you think everything else in this thread is correct then? Swell!

[Blue Bear Sound]

yeah... sure... whatever you say, moron...

[volltreffer]

Hey Boray...

Originally posted by Boray
Not exactly the same thing, but both sets an analog level digitally.

Sure, that's true, but...

Originally posted by Boray
That has very much to do with bit depth. An 8 bit volume control can set an analog level to 256 different values. So can an 8 bit D/A.

There is a huge difference... If you go through a 8 bit volume control with a 16 bit signal, you're gonna multiply the 16bit signal (which is dynamic) with the (static) 8-Bit-signal. This means that the output will have a 16-bit resolution no matter to which of your 256 volume steps you set the volume control. This fact MAY change if you change the volume control dynamically (e.g. in order to build a compressor). But if you choose one static gain of your vol control, you'll just choose one of 256 values your LSB will have, the next one will be factor two a.s.o. ... Hope I could express what I meant...

aXel

BTW.: I found a very interesting article in Swedish on the internet today... Not that I would be able to understand too much of it, but I thought it might be nice to add some random swedish sentences to my posts replying to yours - maybe you'd be able to tell me what I said afterwards...

Ciao,

aXel

P.S.: Re-read this thread, and have to partially review my thought. If I understood you right, you say that the monitor out level control of the vs IS placed BEFORE the DAC? Then it would do it's volume control via some real 'arithmetic' multiplication?? Then it WOULD actually simply reduce the bit depth... Oh my god, why did I ever buy a VS??? I still hope that you're wrong... There are lots of linear ICs like the philips TDA1524 that do control volume, bass, treble etc. of an audio signal and are triggered via DC - for that DC, a simple and cheap 8 bit DAC would be enough... If they wouldn't do it that way, they'd have to spend the costs for a second 20 bit DAC for the monitor out only... Dont' imagine they would... The monitor out should just get its input from one the other DACs...

aXel

aXel

[sweetnubs]

so you've been programming digital audio since the 80's and you've just discovered that if your signal is not near 0dbfs you are not getting maximum bit-dpeth? My favorite part is your only point of reference is a roland vs880! Hey boray, would you please answer my monitor question.

[acidrock]

I'd add my two bits but that would probably confuse things even more.

[Boray]

Axel, Yes, it's all about the monitor knob. That part about that a volume control could be used as an D/A doesn't have anything to do with it at all. (it's not about the monitor knob, but about old computers). Not that I understand your comment about it, but anyway... Check the VSPlanet thread, it's a lot better: http://www.vsplanet.com/ubb/ultimate...c&f=1&t=013032

sweetnubs, It's about if the monitor knob is before or after the D/A. ....VS880???? I don't have one?????

/Anders

[Blue Bear Sound]

Clueless..........

So you think that when you perform any gain-related DSP functions, you are affecting the inherent bit-depth of the recorded digital signal????????????

bwa-ha-ha!!!!!!!!!!!!!!!!!!!!!

That really is funny..........!

(It also once more demonstrates your complete and total ignorance on the subject.............)

[Boray]

Originally posted by Blue Bear Sound
So you think that when you perform any gain-related DSP functions, you are affecting the inherent bit-depth of the recorded digital signal????????????

Not of the recorded signal, but when it reaches the D/A converter... Then it won't use it's whole dynamic range if it's not played loud enough. Don't you agree?