Topic: Edirol R09 - ADC not good enough for 24bit, right?

[ghellquist]

Just to put things in a bit of perspective.

Good 16bit at 44.1kHz sounds plenty good enough. CD-s use exactly that spec, and when well done they are far superior to the listening equipment used by mere mortals. No taped rock concert will ever be close in taxing the raw perfomance of a good chain of this 16/44.1 as the room + audience noise is high and the music covers a rather small dynamic range. So, if you can find a good 16bit/44.1 recorder use that.

So surely 24 bit has to be better every time? Well, that is what the marketing part is trying to make us believe. No self-respecting recorder today could be sold with only "measly" 16 bit performance. So the marketing people makes the engineers twist. They have to put  a 24 bit AD converter in there, but they cannot afford to feed it properly as the boxes are made to target price. Well, to be truthful even the chip manufacturers are treated by marketing -- there is no chip converter you can buy at any price that really gives you "true" 24 bit definition although all of them goes through all the motions to make all the 24 bits. At least two bits are eaten by the noise level, generally several more. And then the engineers have to starve the circuits in front of the AD as well, not giving them enough energy (saving battery) and also saving money on the components. It is like sitting in a Ferrari in the morning rush hours, the horsepower may be there but the car in front of you sets the pace. In this case the analog parts in front of the AD lowers the "true" range downto to about 16 bits for the R09.

In effect what we have here is that there is not much idea in using more than 16 bits when recording. First the room where you record is noisy, and then the analog circuits are noisy, so the extra bits above 16 will contain about nothing more than noise.

Well, then, when can it be useful to go above 16 bits? First, the chain including room noise, analog circuits and AD has to be up to it. The R9 (or even worse, the R1) is not up to it. A Sound Devices 722 is bit better, a large rig with, say, Millenium HV3D preamps and Lavry Blue converters is up to it even more. Not stealthable though, and would make very little difference in a stadium recording PA amplified music. In recording classical music it may be a good idea as the dynamic range (ie difference between loud and weak sounds) is larger. Or in one more circumstance, using the extra headroom as a safety factor or insurance against laziness -- simply put up some low level on the mic pres and record whatever comes, no need to be too careful about levels as they can always be raised in post production.

In post processing it is generally a good idea to go above 16bits. Storage space is cheap and computers today have no trouble handling the data. I will not dwelve into this, but about all software programs today use at least 32bit floating point for internal processing. Nothin you have to care about when catching the sound though, but import once you start modifying it at home.

All theory though falls flat on its face when you listen, the only thing that counts then is what you hear. Listening is a difficult art though, takes practice, and is also extremely influenced by the exact sound level. So if you want to compare, be extremely careful in normalizing levels -- practical experiments show that listeners always preferr a little louder when comparing two recordings. Now, even that may go too far as witnessed by the squashed stuff we get on some radio stations. And, somehow, in the end it is about taste and some like the Vindaloo spiced curry and others want it milder.

Gunnar

[Gutbucket]

Eloquently stated, Gunnar.

This part is enough for me:

... Or in one more circumstance, using the extra headroom as a safety factor or insurance against laziness -- simply put up some low level on the mic pres and record whatever comes, no need to be too careful about levels as they can always be raised in post production.

That alone lets me let down my guard a bit, forget about recording somewhat and just enjoy the music.

Make mine 4 out of 5 on the spicy scale. 
..and don't get me started about 'radio' or 'TV' sound. 

[Church-Audio]

Just to put things in a bit of perspective.

Good 16bit at 44.1kHz sounds plenty good enough. CD-s use exactly that spec, and when well done they are far superior to the listening equipment used by mere mortals. No taped rock concert will ever be close in taxing the raw perfomance of a good chain of this 16/44.1 as the room + audience noise is high and the music covers a rather small dynamic range. So, if you can find a good 16bit/44.1 recorder use that.

So surely 24 bit has to be better every time? Well, that is what the marketing part is trying to make us believe. No self-respecting recorder today could be sold with only "measly" 16 bit performance. So the marketing people makes the engineers twist. They have to put  a 24 bit AD converter in there, but they cannot afford to feed it properly as the boxes are made to target price. Well, to be truthful even the chip manufacturers are treated by marketing -- there is no chip converter you can buy at any price that really gives you "true" 24 bit definition although all of them goes through all the motions to make all the 24 bits. At least two bits are eaten by the noise level, generally several more. And then the engineers have to starve the circuits in front of the AD as well, not giving them enough energy (saving battery) and also saving money on the components. It is like sitting in a Ferrari in the morning rush hours, the horsepower may be there but the car in front of you sets the pace. In this case the analog parts in front of the AD lowers the "true" range downto to about 16 bits for the R09.

In effect what we have here is that there is not much idea in using more than 16 bits when recording. First the room where you record is noisy, and then the analog circuits are noisy, so the extra bits above 16 will contain about nothing more than noise.

Well, then, when can it be useful to go above 16 bits? First, the chain including room noise, analog circuits and AD has to be up to it. The R9 (or even worse, the R1) is not up to it. A Sound Devices 722 is bit better, a large rig with, say, Millenium HV3D preamps and Lavry Blue converters is up to it even more. Not stealthable though, and would make very little difference in a stadium recording PA amplified music. In recording classical music it may be a good idea as the dynamic range (ie difference between loud and weak sounds) is larger. Or in one more circumstance, using the extra headroom as a safety factor or insurance against laziness -- simply put up some low level on the mic pres and record whatever comes, no need to be too careful about levels as they can always be raised in post production.

In post processing it is generally a good idea to go above 16bits. Storage space is cheap and computers today have no trouble handling the data. I will not dwelve into this, but about all software programs today use at least 32bit floating point for internal processing. Nothin you have to care about when catching the sound though, but import once you start modifying it at home.

All theory though falls flat on its face when you listen, the only thing that counts then is what you hear. Listening is a difficult art though, takes practice, and is also extremely influenced by the exact sound level. So if you want to compare, be extremely careful in normalizing levels -- practical experiments show that listeners always preferr a little louder when comparing two recordings. Now, even that may go too far as witnessed by the squashed stuff we get on some radio stations. And, somehow, in the end it is about taste and some like the Vindaloo spiced curry and others want it milder.

Gunnar

Well said.. Great post. +T

[SparkE!]

Just to put things in a bit of perspective.

<A whole series of basic truths>

Gunnar

Well said.. Great post. +T

Yup.  +T for sure.  Thanks for taking the time to write such a well-conceived post on this topic.  Most people have such a hard time sorting fact from fiction when it comes to issues that are affected by resolution, accuracy and noise.

[Javier Cinakowski]

Gunnar, I would like to subscribe to your newsletter.

T+

[spyder9]

Just to put things in a bit of perspective.

Good 16bit at 44.1kHz sounds plenty good enough.

+T Gunnar.  Great post.

[Brian Skalinder]

Good 16bit at 44.1kHz sounds plenty good enough. CD-s use exactly that spec, and when well done they are far superior to the listening equipment used by mere mortals. No taped rock concert will ever be close in taxing the raw perfomance of a good chain of this 16/44.1 as the room + audience noise is high and the music covers a rather small dynamic range. So, if you can find a good 16bit/44.1 recorder use that.

< snip >

In effect what we have here is that there is not much idea in using more than 16 bits when recording. First the room where you record is noisy, and then the analog circuits are noisy, so the extra bits above 16 will contain about nothing more than noise.

Well, then, when can it be useful to go above 16 bits? First, the chain including room noise, analog circuits and AD has to be up to it. The R9 (or even worse, the R1) is not up to it. A Sound Devices 722 is bit better, a large rig with, say, Millenium HV3D preamps and Lavry Blue converters is up to it even more. Not stealthable though, and would make very little difference in a stadium recording PA amplified music.

Nice feedback, Gunnar.  Curious about your take on my previous post (and the one to which I link in the post):

http://taperssection.com/index.php/topic,79375.msg1066030.html#msg1066030

Summary:  as I understand it, there's more at issue here than dynamic range - i.e. more samples in the vertical axis (4x at 24-bit v. 16-bit) used to describe the waveform so it more closely approximates the analog waveform - and so there are benefits to 24-bit recording regardless of whether one has the analog gear and / or recording environment to take advantage of the broader dynamic range.

[ghellquist]

Hi Brian,
       this is definitely not simple matters. I am not sure I really understand it and definitely certain that I cannot really explain it. I´ll do my best though. The matter is not quite intuitive. Some parts of it seems to be quite against common sense.

First, let us assume that we are measuring a sound source going between -1 and +1. It could be Volts, the units is not the important thing here.

At every moment in time we want to measure it as exactly as possible. Depending on the number of bits we have in the AD the resolution we have is (figures are rounded):
   16 bit :  0.00003
   20 bit :  0.000002
   24 bit :  0.0000001

What this means is that we cannot "see" a difference smaller than that.

Now, add noise into the figure. If we measure a Sm (measured signal) it is actually composed of S + N (the true Signal + Noise). The noise is a (more or less) random figure, it could be either plus or minus. Let us assume that the N part goes between +0.0001 and -0.0001. What this tells us is that, in general, the smallest bits we are measuring are totally swamped by noise. Regardless if we use 16 or 24 bits in that example, the last few bits are still only noise.

Let us go back to the figure S/N or signal to noise ratio. We have to do a mathematical operation here to go from number of bits to an S/N value. But as large mathematicians has already done that we can rely on them. It turns out that one bit represents about 6 dB. So if we have 16 bits we can have a maximum of 6 x 16 = 96dB of S / N. In effect this is a different way to say that the smallest measurement we can make is 0.00003.

As the noise is composed of many parts, we might want to look at them. First there is noise in the room -- a very quiet room may have 30 to 40 dB of noise. And if the maximum sound volume we want to have is, say 120 dB, then the room has something like 80dB S/N or about 13 bits. (And it is not quite as simple).

Next comes the mic. Smaller mics generally has more noise than larger (this is partly due the laws of nature, maybe to be talked about at a later time). Good small diameter condensers turn in at about 20dB equivalent noise level, really small electret mics may have much more. Add to that mic pre amps that can only amplify and add noise, never remove it.

The problem with noise signals is that they add together (adding dB is not a simple adding operation though in this case, as we are adding different noises). In the end we might find out that the 40dB Noise of the room has gone up to, say 50dB, but we still have a maximum of 120dB. So we are now down to 70dB of S/N, ie around 12 bits.

Now, nothing is as simple as the above examples. The ear will actually hear things a bit below the noise floor as expressed in bits. We will probably waste one or two bits at the top as we keep signal levels conservative to insure that we never clip the signal (which sounds horrible). But all in all, for typical amplified music, 16 bits is definitely OK if we are sure that our chain gives us that.

For classical music, we want a little more. 20 bits is about enough for most all classical music, an about what we can get from top-level stuff today.

Personally I preferr 20 bits (well, ok then, my machine calls it 24 bits, but 20 is about what I get). The simple reason is that I want to be safe and lazy. Safe against going up into the red territory where things start clipping. Lazy enough to simply set a recording level and then let it run. I generally aim at -12dB for the peaks, they tend to go a little over that from some reason. This will "eat" about 3 bits, but there is still 17 left to go, plenty really.

Could there be other reasons to record with more than 16 bits? Probably, for ease of postprocessing or perhaps simply because the box we record might sound different at different bit depth.

Sorry, not quite clear presentation, done my best though.

Gunnar

[poorlyconditioned]

Hi Brian,
       this is definitely not simple matters. I am not sure I really understand it and definitely certain that I cannot really explain it. I´ll do my best though. The matter is not quite intuitive. Some parts of it seems to be quite against common sense.

First, let us assume that we are measuring a sound source going between -1 and +1. It could be Volts, the units is not the important thing here.

At every moment in time we want to measure it as exactly as possible. Depending on the number of bits we have in the AD the resolution we have is (figures are rounded):
   16 bit :  0.00003
   20 bit :  0.000002
   24 bit :  0.0000001

What this means is that we cannot "see" a difference smaller than that.

Now, add noise into the figure. If we measure a Sm (measured signal) it is actually composed of S + N (the true Signal + Noise). The noise is a (more or less) random figure, it could be either plus or minus. Let us assume that the N part goes between +0.0001 and -0.0001. What this tells us is that, in general, the smallest bits we are measuring are totally swamped by noise. Regardless if we use 16 or 24 bits in that example, the last few bits are still only noise.

Let us go back to the figure S/N or signal to noise ratio. We have to do a mathematical operation here to go from number of bits to an S/N value. But as large mathematicians has already done that we can rely on them. It turns out that one bit represents about 6 dB. So if we have 16 bits we can have a maximum of 6 x 16 = 96dB of S / N. In effect this is a different way to say that the smallest measurement we can make is 0.00003.

As the noise is composed of many parts, we might want to look at them. First there is noise in the room -- a very quiet room may have 30 to 40 dB of noise. And if the maximum sound volume we want to have is, say 120 dB, then the room has something like 80dB S/N or about 13 bits. (And it is not quite as simple).

Next comes the mic. Smaller mics generally has more noise than larger (this is partly due the laws of nature, maybe to be talked about at a later time). Good small diameter condensers turn in at about 20dB equivalent noise level, really small electret mics may have much more. Add to that mic pre amps that can only amplify and add noise, never remove it.

The problem with noise signals is that they add together (adding dB is not a simple adding operation though in this case, as we are adding different noises). In the end we might find out that the 40dB Noise of the room has gone up to, say 50dB, but we still have a maximum of 120dB. So we are now down to 70dB of S/N, ie around 12 bits.

Now, nothing is as simple as the above examples. The ear will actually hear things a bit below the noise floor as expressed in bits. We will probably waste one or two bits at the top as we keep signal levels conservative to insure that we never clip the signal (which sounds horrible). But all in all, for typical amplified music, 16 bits is definitely OK if we are sure that our chain gives us that.

For classical music, we want a little more. 20 bits is about enough for most all classical music, an about what we can get from top-level stuff today.

Personally I preferr 20 bits (well, ok then, my machine calls it 24 bits, but 20 is about what I get). The simple reason is that I want to be safe and lazy. Safe against going up into the red territory where things start clipping. Lazy enough to simply set a recording level and then let it run. I generally aim at -12dB for the peaks, they tend to go a little over that from some reason. This will "eat" about 3 bits, but there is still 17 left to go, plenty really.

Could there be other reasons to record with more than 16 bits? Probably, for ease of postprocessing or perhaps simply because the box we record might sound different at different bit depth.

Sorry, not quite clear presentation, done my best though.

Gunnar

Yep, I agree with this.  If the noise floor is -90dB, or about 15bits, which I have measured, any signal below that is going to be buried in the noise.  I suppose you might get some gain by averaging out the noise with bunch of samples measured at 24bits, but I don't expect much improvement.  Certainly not worth all the hassle (and fake marketing as a 24 bit recorder!).  As I said, Burr-Brown does not even call this a 24 bit chip!

  Richard

[trajhip2000]

Summary:  as I understand it, there's more at issue here than dynamic range - i.e. more samples in the vertical axis (4x at 24-bit v. 16-bit) used to describe the waveform so it more closely approximates the analog waveform - and so there are benefits to 24-bit recording regardless of whether one has the analog gear and / or recording environment to take advantage of the broader dynamic range.

I don't think this is how linear quantization works - it's 1 bit per 6 dB of dynamic range, and that's fixed. So by going to 24-bit you're not dividing the same dynamic range into more bits, you're adding more dynamic range by (theoretically, anyways) decreasing the noise floor of the A/D chip. At least, that's how I understand it...

Steve

[SparkE!]

Summary:  as I understand it, there's more at issue here than dynamic range - i.e. more samples in the vertical axis (4x at 24-bit v. 16-bit) used to describe the waveform so it more closely approximates the analog waveform - and so there are benefits to 24-bit recording regardless of whether one has the analog gear and / or recording environment to take advantage of the broader dynamic range.

I don't think this is how linear quantization works - it's 1 bit per 6 dB of dynamic range, and that's fixed. So by going to 24-bit you're not dividing the same dynamic range into more bits, you're adding more dynamic range by (theoretically, anyways) decreasing the noise floor of the A/D chip. At least, that's how I understand it...

Steve

I think the point is that noise causes the lower 4 or 5 bits to be random 1's and 0's which effectively causes a 24 bit converter to act as a 20 or 19 bit converter.  If noise masks the least significant bits like that, you don't gain the benefit of the additional bits.  You might as well  call it a 73 bit converter and add an additional string of 49 randomly set bits on the end of each 24 bit sample.  The result is still the same.  You only get the benefit of the most significant 19 or 20 bits.

[Brian Skalinder]

Thanks for the feedback Gunnar, Richard, Steve.  Since the sound source / environment of my concert recordings typically covers only ~50 dB of dynamic range (or ~8-bits) - meaning (assuming I understand you correctly) at 16-bit the least significant 8-bits capture only noise, and at 24-bit the least significant 16-bits (!) capture only noise (well, more like ~19-bit actual & least significant 11-bits of noise) - then I'm at a loss for explaining why my 16-bit recordings sound better than 8-bit recordings, and why 24-bit (~19-bit actual) recordings sound better than 16-bit.



Maybe I'll get my head around it one of these days...

[Nick's Picks]

DSD outs?
doesn't the tascam have a HDD in it ?  I thought it was offered in newer models, and as an upgrade for the older ones.

my plan was to record DSD on the MR-1, and then play it back on the Tascam (after transferring files to its internal HDD).  Unless i'm way off base...that was my goal.

[Nick's Picks]

another thing...

the "1bit" DSD thingy...which certainly trumps the PCM world...
where does the noise go ?

I believe the process involved here w/DSD is so unlike PCM, that my conventional pulse control brain waves just can't wrap around it.

[BayTaynt3d]

I've said it before, and I'll say it again... I believe 24 bits is worth it for resolution alone for anyone who so much as touches their recordings in post. But one of my main contentions has been for a while now (with no one seeming to pile on to this point ever), is that if you are adding any digital gain in post via compression (plus make up gain) or normalization, than 24-bit recording is going to be your friend. Speaking of friends, our friends at Sound Devices seem to have created a nice example of JUST WHAT I'VE BEEN SAYING for a while now on this subject. They've got 16-bit and 24-bit examples, and then they have a set that was recorded a WHOPPING 40 db down from peaking, then they normalized them all. If running peaks 40 db down doesn't convince you of the headroom, then I'm not sure what will. So, take a listen for yourself here:

http://www.sounddevices.com/tech/24-bit.htm