Noise floor (Signal to Noise Ratio)...Please explain this term

[Roy Howell]

I've noticed this or other terms similar to this in a post or two. With no schooling other than the DUC, I'm just curious what it means in simple terms. Could someone explain, and thanks..........

Roy

[Mr_Seven]

Simple terms....
(Digital ref.)
_______________________________

-< Clipping
---------------------< 0db
-
-< Head Room
-
---------------------< -18db
-
-
-< Noise Floor
-(background, ambient, hum, etc.)
_______________________________

The S/N ratio is the difference between your recorded level of sound (hopefully within the headroom) and the noise floor. The greater the difference the better i.e. more sound, less noise.

[Roy Howell]

OK, Darren,

But, if I were in a classroom, I would likely ask, "Why is the best audible meat of the sound measured in -db(minus), and Clipping at 0db?

Is that one of those "That's just the way it is, kid" questions?

Thanks, RH

[Sabe]

Roy,

I found an article written by the President of Aphex that might paint a clear, concise picture of what you want to know.

Enjoy [img]images/icons/smile.gif[/img]

Aphex Thermionics White Paper

Title: More Gain. No Pain.

By Marvin Caesar, President of Aphex Systems Ltd.

Synopsis: An extraordinary mic preamp combining new design philosophies allows you to safely run at higher gains without the pains of noise and overload distortion.

I. The Value of a Wide Dynamic Range
Consider dynamic range as a window. The top is the maximum peak level and the bottom is the noise floor. These are physical limits and they exist in the both the analog and the digital worlds. Ideally, the window is wide enough to accommodate the highest input level without any overload distortion while adding as little noise as possible to the signal.

Wide dynamic range for a microphone preamplifier is particularly important inasmuch as the level of the input into the microphone can vary greatly. In order to accommodate these variations the gain in the preamplifier is set so there is no overload distortion on the highest peaks. The difference between that nominal gain setting and the maximum peak level is headroom. Setting the gain too low in the preamplifier, however, will require gain in a later stage. That means that any increase in gain in the later stage will also boost the noise from the preamplifier. Obviously, the lower the noise floor in the preamplifier, the lower the noise on the final output.

If the output of the preamplifier is digitized at too low a level, the conversion will have low resolution. One bit represents 6dB of dynamic range in the digital domain. If the input is converted at -24dBfs the resolution will be four bits less than full resolution. Once the signal is converted there is no way to increase the resolution.

II. Setting Up a Conventional Microphone Preamplifier
A microphone is almost always used to pick up a live acoustic source, e.g.- a voice, an instrument, or ambient sound. Since level variations from these sources can be quite high, it is imperative that a great amount of headroom be set in the conventional preamplifier. This reduces the chances that the preamplifier will be overloaded due to an unexpected increase in input level, but the nominal output level will be very low.

That low output level will have to be boosted in a following gain stage. This, however, shifts the problem of overload distortion to the following stage. That is why it is quite common to see a compressor or a limiter in between the preamplifier and the following stage.

The problem of noise build up, however, becomes quite apparent. As mentioned above, any gain taken on the signal after the preamplifier increases the noise from the preamplifier by the amount of gain in the second stage. In addition, the noise of the second stage itself combines with the input noise. For example, if the noise of the preamplifier is -60dBu and the noise of the following stage is -60dBu with 10dB of gain, the noise at the output of the second stage will be -47dB. Note: When two equal non-correlated noise sources are summed the noise is increased by 3dB.

When a compressor is used, it brings up the lower level signals (including noise) by whatever make up gain is set in the compressor. Adding to the noise in the output of the compressor is the noise of the compressor itself.

As you can see, noise builds up very quickly if the dynamic range of each gain stage is not maximized. That is why it is essential to choose the equipment with the widest possible dynamic range and use that equipment properly. And the most important gain stage is the first gain stage- the microphone preamplifier.

III. Determining the EIN and the Dynamic Range of a Microphone Preamplifier
A very important specification for any microphone preamplifier is the equivalent input noise (EIN). The noise is measured with the input shorted and at a specific gain. That figure is added to the gain. For example, a preamplifier with 60dB of gain has a noise floor of -68dBu. Adding the noise to the gain gives that circuit an EIN of -128dBu.

The dynamic range of that preamplifier at that gain setting is computed by adding the noise and the maximum output level. For example, if the preamplifier has a maximum output level of +27dBu the dynamic range of the preamplifier is 95dB (68 + 27) at that gain setting.

IV. Designing the Aphex Model 1100 Tube Microphone Preamplifier

1. Noise
One of the primary design goals of the Model 1100 was to have as wide a dynamic range as possible. Several key inventions combined with a no-compromise selection of components (see below) create a microphone preamplifier with unprecedented performance. The EIN with 65dB of gain is an incredible -135dBu. That means that the Model 1100 adds less that 1dB of noise to the natural self noise of a 150 ohm microphone. The worst case dynamic range is 97dB and is a high as 101dB. But low noise is only part of the story.

2. 20dB extra headroom
As described above, a conventional preamplifier must be set for sufficient headroom in order to avoid overload. The Model 1100 has two inventions that actually provide up to 20dB of extra input headroom so that it is virtually impossible to overload the preamplifier. A third invention does not directly increase headroom, but maximizes available headroom in the digital domain.

a. MicLim
The first invention is the Microphone Limiter (MicLim), first used on the Model 1788. It comprises a custom designed optical attenuator directly on the microphone input line. It smoothly limits the microphone output signal prior to the preamplification by up to 20dB. The peak limit detector is located after the preamplifier input stage and feeds a control current back to the attenuator so that the input signal remains below clipping. MicLim has no effect whatsoever on the input signal until the preamplifier’s output approaches clipping.

b. Low Frequency Cancellation Filter (LoCafTM)
The second invention is a tunable low frequency cancellation filter (LoCaf). It is a second order (12dB/octave) modified Butterworth filter meshed into the nodal intersections of the first and second amplifying stages in a servo configuration. The servo affects only frequencies below the corner frequency, thus it contributes nothing to the audible signal. Imposing the servo filter in such a manner gives the preamplifier about 20dB more overload headroom in the low cut range as compared to conventional techniques. Additionally, the added low frequency headroom eliminates the need for the MicLim to trigger earlier than necessary from excessive low frequency energy.

c. Drift Stabilized A/D Converter
Conventional analog to digital converters utilize high pass filters in the digital domain to block any DC generated in the conversion process or already in input signal. While this is effective in eliminating the DC, it requires extra headroom in the converter to allow for the DC. The patented drift stabilized A/D eliminates the DC in the analog domain so that the input can be at the true maximum level. Since there is no high pass filter in the digital domain, all ringing from that filter is also eliminated.

3. Other Features
a. Full Featured AES/EBU Digital Audio Output
AES/EBU XLR output is standard. Clock synchronization options allow for locking to standard “word clock” and to AES/EBU clock received at the standard BNC clock input jack. Internal clock options provide low jitter 44.1, 48, and 96HKZ sample rates. When a unit is set for internal clock, its internal word clock reference is sent to the rear panel word clock BNC output jack to serve as clock reference to other units. When the unit is set for an external clock reference, the clock input BNC jack is directly tied to the output clock BNC jack for easy daisy-chaining of Model 1100 units from the master clock source. All digital audio settings are controlled and displayed on the front panel.

The A/D converter receives signal from the soft mute stage just prior to the analog output level control and triode output stage. This means you can use both the digital and analog outputs independently, with full and proper calibration of both regardless of the analog output level settings. The analog and digital outputs respond equally to the input gain, low-cut filter, and all front-end conditioning effects.

d. Bifurcated 20dB Pad and Phase Reverse
Many preamplifiers, even the more expensive models, switch microphone level signals directly through switch contacts. It is well known that even the best quality switches will eventually suffer from dry contact diode effects causing noise and distortion. The Model 1100 uses high-grade, bifurcated gold contact relays which do not develop these problems. The Gain and Low Cut controls are sealed gold contact rotary switches.

e. Precision Three Turn Output Level Attenuator
In order to match the analog output of the Model 1100 to the user’s system level, the output gain is adjustable from zero dB (max gain) to -14dB. The user will appreciate the smoothness and precision afforded by the front panel 3-turn high-grade potentiometer adjustment.

f. 48-volt Phantom Power Circuit
Very slow rise and fall of the phantom voltage is used to eliminate turn-on and turn-off thumps. Industry standard resistances of 6.81 K-Ohms supply the highly filtered 48-volt source to pin 2 and pin 3 through a voltage ramping active buffer. The phantom powering system can withstand a short circuit to ground on both microphone jacks indefinitely.

g. Series-Shunt, Optical Soft Mute Attenuator
The second-stage output signal passes through a specialized series-shunt optocoupler circuit to provide a soft mute while introducing no distortion or noise.

h. Front Panel Peak Headroom Meter and Function Controls
Each channel contains a 20-segment LED headroom meter, making it easy to optimize the performance of the preamplifier. The headroom meter is calibrated in decibels below clipping, where 0dB is the analog clipping point. This coincides with the A/D converter’s maximum input level, so the headroom meter also indicates the digital audio level accurately. Each channel contains its own independent controls over every function.

i. Rear Panel Mute Jack
The mute function may be activated by the front panel push-button, or by a remote switch plugged into the mute jack. In the absence of a phone plug, the mute jack serves as a closed circuit, and only the front panel push-button has control. In the presence of a phone plug, an open circuit mutes the preamplifier while a closed circuit un-mutes. This facilitates the convenient use of mute controls such as floor mat switches (step on to un-mute) and musician’s footswitches.

j. Internal Linear Low Noise Power Supply
Though heavier and more expensive, we designed the Model 1100 with a high quality, fully regulated, linear power supply to maintain the highest audio performance. Every output is regulated including the tube filament voltage. The whole power supply is internally contained for maximum user convenience.

V. Sound Quality by Design
While specifications and functions are important, the most important characteristic of a piece of audio gear, particularly a microphone preamplifier, is how it sounds. The sound of the Model 1100 is clean, clear, present, open and solid. It is extraordinarily detailed and spacious. The low end stands up without any muddiness and the high end is very extended without any harshness.

This sound is achieved through the use of proprietary designs, careful engineering and the highest quality components. It is a combination of Class A discrete components and patented tube circuitry as briefly described below.

1. Ultra Low Noise, Transformerless, Discrete Class A, Bipolar PNP, Variable Gain Differential Input Stage
No outer feedback is used, thus eliminating the possibility of any dynamic interaction with the microphone’s self-impedance. The input impedance remains passive, providing an optimal load for any microphone. The solid state, class-A PNP bipolar design achieves high common-mode rejection with extremely low noise, wide bandwidth, and low distortion at all gain settings. Robust input overload protection assures that all performance features will be retained indefinitely. The gain is adjustable in 4dB precision steps from 21 to 65dB.

2. Tube, Discrete Class A Differential Second Stage
The unique, Aphex patented, “Reflected Plate Amplifier” tube circuit is configured as a single-triode differential opamp to further enhance the preamplifier’s common mode rejection. This novel circuit topology subtly imparts the tube’s sonic warmth and character while retaining relatively long and stable operating life.

3. Tube, Discrete Class A Output Stage
An Aphex patented “Reflected Plate Amplifier” tube circuit is configured as a low distortion triode buffer having a very low output impedance and high output current drive. The maximum output level of +27dBu meets the needs of any professional application. Matched-impedance balancing assures peak performance whether driving balanced or unbalanced lines. A rear panel switch is assigned to insert a 12dB low impedance pad into the output line for systems based on IHF (semi-pro) operating levels. Rear panel XLR and quarter-inch phone jacks are both provided for balanced output.

VI. Now The Bad Stuff
The Model 1100 consumes approximately 35 watts. That means that it gets hot. It requires air flow above and below the chassis. Under no circumstances should units be stacked on one another. The front panel is aluminum polished to a glass finish. While beautiful, it will show smudges. Keep it clean.

VII. Summary
Every circuit and component that went into the Model 1100 was studied and scrutinized for optimum performance. The result of the innovations and careful engineering is a uniquely excellent preamplifier.


Copyright 2000. All rights reserved.

[Sabe]

I apologize if the above post seemed as an advertisement, but there is very good info within it. Particularly about the EIN and what it means to the end user.

In case you couldn't tell, I own one of these units and so far so good. It is everything they claim and more.

Side note: My signal chain is as follows.

AKG C414 (mic)- Aphexx 1100 (mic pre)- TC Electronics Finalizer (channel strip/converter)- ProTools breakout box.

With the exception of the cabling between the AKG and the mic pre, all signals are proceessed at 96k. The Finalizer then resamples to 48k for ProTools to utilize.

[David001]

Post deleted by David001

[Roy Howell]

David,

I sincerely appreciate your explanation of each term, because you took the time to really make me understand.

You're obviously a good man, and I thank you so much for this lesson.

Roy

[Roy Howell]

Sabe,

No need to apologize. Thank you for the education. I'm open to learn now that I'm engineering my own stuff.

Yours sounds like a simple, effective setup.

Thanks, Roy