NOTION Music Forum

This topic is focused on exploring the problem of noisy string samples and perhaps providing a few solutions . . .

[NOTE: For reference, "noise" generally is defined to be any undesired sound . . . ]

I did a few experiments, and this problem happens with the Notion 4 string sections, but it also happens with the solo violin, solo viola, and solo cello, although some of the troublesome notes are different for the solo strings . . .

[NOTE: You can view the complete image via right-clicking on it with Firefox ("View Image") or Safari ("Open Image in New Tab") . . . ]



"N4-String-Sections.notion" ~ (approximately 61 KB)

The solution is to use a VSTi virtual instrument for the strings, but the VSL strings that come with Kontakt 5 (Native Instruments) and the strings that come with MachFive 3 (MOTU) have the same problem, as do some but not all of the Miroslav Philharmonik string ensembles (running with NOTION 4 in 32-bit mode), but as noted Miroslav Philharmonik is 32-bit mode only, although IK Multimedia is taking pre-orders for SampleTank 3 which will be released in July 2014 and plays the various Miroslav Philharmonik sampled sounds . . .

As best as I can determine, violins and violas generally are troublesome, but this is based on the various sound sample libraries I have here in the sound isolation studio . . .

Intuitively, I think that Vienna Symphonic Library (VSL) should have some good violin and viola sample sound libraries, but I have no way to verify this at present . . .

The section violins that come with Finale 2012 (Make Music) sound pretty good, and they are Garritan Personal Instruments for Finale, which is encouraging; and you can use them in NOTION 4 via the Aria VSTi virtual instrument, noting that I have this, hence was able to do a few tests with it, and it runs in 64-bit mode, which is important . . .

[NOTE: This discussion started in a different topic, but I think it is important and deserves its own topic. This is one of the relevant posts, which I quote and copy here . . . ]

idiotSavant wrote:

Surfwhammy wrote:I also like the instrumentation but with one exception, which specifically is the instrument that has what appears to be a prominent white noise component . . .

Surfwhammy, I think what you're hearing is the sample's rendition of the sound of the bow scraping the strings. I've noticed that this can be grating on many violin patches from various sound libraries (VSL included) and also noticed it in John's Clouds piece.

I usually apply a focused parametric EQ to eliminate this effect, sometimes successfully, sometimes not. I think this adds to what many refer to as a "chilly" sound from the strings, especially the violins.

THOUGHTS

Relevant items of interest include (a) solutions; (b) sampled sound libraries that do not have the problem; (c) workarounds and techniques for reducing the noise; and so forth . . .

However, since psychoacoustics is a key aspect, I created yet another topic to focus on psychoacoustics, where for reference "psychoacoustics" focuses on the perception of sound and is quite fascinating, especially with respect to creating and identifying auditory illusions (a personal favorite) . . .

Psychoacoustics and Digital Music Production (Notion Music FORUM)

Lots of FUN!

These are a few relevant quotes from the "Clouds" topic, which is where the idea for this topic on string samples and noise originated . . .

JohnF wrote:Surf - I think you're right, it's the upper register of the cello's that's creating that sound. I did compare with my exported .wma and it is more prevalent on soundcloud, due to compression I suppose. BTW, these are not notion samples, they are Hollywood Strings. Thanks for the detailed write-up!

idiotSavant wrote:

JohnF wrote:
idiotsavant - . . . Do you eq all the time?

John, when I first started using VSL I struggled with the "cold" sound of the string samples. I started playing with EQ about a year ago to warm up the sound. Then recently someone here posted something that was quite a revelation for me. They said that many sample libraries (and this might include yours) are what I will call "frequency-inclusive". In other words, the samples are recorded, usually in a sound isolation chamber, to capture the full range of an instrument. But listeners in "real" environments never hear the instrument this way - they hear it with frequencies attenuated by the music stand, by the player in front of the player, by the players in front of those players, by the diffusion created by the materials in the hall or room, by the audience in front of the listener, etc. etc. You get the picture.

Since you can't add frequencies that don't exist in a sample, the intent with a "frequency-inclusive" sample is that virtual instrument recordists (us) can use EQ in a subtractive way to start to filter frequencies. The frequencies that I end up filtering are the highs in very particular ranges with tight Q values (what I call "surgical EQ") to remove some of the harshness that I find associated with the samples I'm using. The "scrape" of the high range in a cello is a perfect example. I used to try to add low end to warm things up but found that everything was becoming a muddy mess.

I also will often boost the +/- 250 hz range in strings which equates to the "wood" of the instrument. This makes the sound more realistic to my ears since it brings out the sound of the hollow wood body, which is where warmth comes from for strings. I also sometimes boost the highs at about 1.5-3.0 kHz (depending on the instrument) to bring out the sound of the bow moving across the string if that's the effect I want.

Very long answer to your short question, but I hope you find it helpful. Almost any DAW has built in EQ, and there are a lot of free and inexpensive EQ plug-ins out there. I'm searching now for the best EQ tools. I just tested a very nice VST EQ from Hofa. Free evaluation download, check it out.

http://hofa-plugins.de/pages/start_en/hofa-iq-eq_en.php

THOUGHTS

As JohnF indicated, this list of strings with noise issues includes Hollywood Strings (EWQL), and I continue to be intrigued by the experiments and observations of idiotSavant . . .

When I listen to some of the staring samples I have here in the sound isolation studio, I cannot help but wonder whether some of the recordings were enhanced with compressor-limiters, which if this is the case certainly is a mistake insofar as I am concerned, because while I love and use compressor-limiters for very specific purposes, the facts of the matter are (a) that compressor-limiters destroy and annihilate dynamics and (b) that compressor-limiters have the great potential to make noise considerably more prominent, which in the digital music production universe is a huge problem . . .

Why?

Great question!

In contrast to the way everything works in the analog magnetic tape universe, in the digital music production universe there is no "generational loss" with respect to tone, but this significant improvement in the realities of the analog magnetic tape universe where there is "generational loss", this benefit comes at a price; and the price is the accumulation of noise . . .

The background on this is that in the analog magnetic tape universe, each time an audio track is "bounced" from one analog magnetic tape machine and corresponding magnetic tape to another--which can be the same analog magnetic tape machine and its corresponding magnetic tape--some of the original tone is lost, and this is called "generational loss", since each "bounce" is considered to create a new "generation" . . .

[NOTE: When two analog magnetic tape machines are used, the technique is called "ping-ponging", and this technique was used in Abbey Road Studios with four-track analog magnetic tape machines during the production of "Sgt. Pepper's Lonely Hearts Club Band" (Beatles). Phil Spector used this technique to create his "Wall of Sound" in the monaural universe; and a key part of Phil Spector's success was his ability to determine with great accuracy the correct way to record instruments and voices so that the desired tone was correct after a planned set of "bouncing" or "ping-ponging". Both these producers built songs in layers, and layering requires a bit of advance planning. However, in modern times the general technique traces to Les Paul, who used these techniques long before Phil Spector and George Martin . . . ]

If you have an analog or digital sound pressure level (SPL) meter, then you probably know that there is a good bit of ambient sound in rooms, when it appears that nothing is happening. The only exception is an anechoic chamber, but these typically are found only in research laboratories . . .

Using a NADY DSM-1 Digital SPL Meter, the ambient noise here in the sound isolation studio is approximately 40 dB SPL when measured with dBA weighting, and I consider the sound isolation studio to be very quiet. It also is excellent both (a) for preventing the incursion of external sounds and (b) for preventing the excursion of internal sounds, where for example when I had the Really Bigger Drumkit in the sound isolation studio I could play it and with all the doors closed someone outside at a distance of 20 feet could not hear it. in fact, you can run a gasoline powered Stihl chainsaw in the sound isolation studio with the same results, although a bit of vibrating is felt . . .

Occasionally, there are rodeo parades that go by the front of the house, approximately 100 feet from the sound isolation studio, which includes marching bands, horses, and so forth; but when I am inside the sound isolation studio none of it is heard, which is fine with me and is one of the reasons I designed and built the sound isolation studio . . .

Noise exists, and the key is to manage it, where one of the important aspects of noise management when recording is the "signal-to-noise ratio", where the basic rule is that the desired audio material (the "signal") needs to be sufficiently greater than the undesired audio material (the "noise"); and when this rule is followed, everything is good and the noise is so far in the background that for perceptual purposes there is no noise . . .

This is where the accumulation of noise in the digital music universe becomes problematic, because the consequence when noise is not handled correctly is that noise becomes dominant; and while this also was a problem to a lesser extent in the analog magnetic tape universe, it is a significant problem in the digital music production universe, hence requires a different set of rules . . .

One strategy which originated in the analog magnetic tape universe is to use devices called "noise gates", which allow sounds above a certain volume level but allow no sounds when the volume level drops below a predetermined threshold; and noise gates are an excellent way to reduce ambient studio noises when recording . . .

Brickwall equalizers are another useful tool, and they can be set to allow only those frequencies above a specified lowest frequency and only those frequencies below a specified highest frequency, where one use is to remove ambient low frequency rumble and other noises, but there are additional ways to use brickwall equalizers, primarily to partition or to focus instruments and voices to specific ranges of frequencies . . .

~ ~ ~ Continued in the next post ~ ~ ~

~ ~ ~ Continued from the previous post ~ ~ ~

Summarizing, when there is noise in string samples and there are a lot of string instruments, what happens is that the noise which travels with each instrument staff or track accumulates; and what might acceptable when there is only one string instrument becomes unacceptable when there are several string instruments each of which has the same or similar level and spectral footprint of noise tagging along, because at some point the accumulated noise actually is louder than some of the instruments, which is something that happens primarily in the digital music production universe due to the ease of working with layered or "stacked" audio clips . . .

[NOTE: This also happens in the analog magnetic tape universe; but due to the "generational loss" problem, managing noise is part of the overall strategy . . . ]

Managing noise certainly is a key aspect of the overall strategy in the digital music production universe, but it only becomes a key aspect when you know about it; and for the most part the only clue is found in an overview where it usually is a footnote to the wonders of digital music production with respect there being no "generational loss" with digital audio . . .

Lots of FUN!

P. S. It is important to understand that finding solutions to the problem of strings and noise primarily is a producing and audio engineering activity, since for the most part there is no control over the way the sampled sounds were recorded and digitized . . .

Using an analogy, metaphor, or simile, it is like devising a way to record and to produce a wonderful lyrical soprano who for some unknown reason has a portable white noise generator which she turns ON every time she stops singing and occasionally does this when she shifts from singing forte to pianissimo . . .



One can use a noise gate to deal with the noise when singing stops, but a noise gate provides no help when there is noise during a ramp-up or ramp-down of singing, which makes this a different problem . . .

Another key bit of information is that reverberation generally aggravates the problem; and in this context one might suggest that in the digital music production universe, reverberation is the vacuum tube of the analog universe, at least in the sense of providing a bit of controlled "blur", which can be a good thing depending upon how it is configured and used, where the primary problem with reverberation is that it requires extensive space actually to do anything useful, where in this context "space" refers to volume, panning, and spectral footprint, with the general rules being similar to the rules for whammying, which for practical purposes is a two-part activity that usually needs to be started before it is heard and then needs to extend long enough for the pitch variation to occur and to be heard, all of which requires considerable sonic space, really . . .

Really!

Clearly this is a topic widely discussed by many. A noisy unnatural sample is quite distracting.
"Sample noise" can certainly distract from a creative idea.
It's just not pleasant to work with.

This web discussion may be of interest in regard to noisy string sample libraries.
"white noise" in string samples? (VSL, LASS, etc.)
http://www.vi-control.net/forum/viewtopic_print.php?t=20461&start=0&sid=1e7f938b58cb816051e7a5d38b4dd1d9

Here's a link to a web article that describes some basic terminology and a few of the challenges associated with "real" live string instrument recording. I found some of the terminology helpful in understanding the source instrument sound.
Recording Strings
Tips & Tricks
Technique : Recording / Mixing
http://www.soundonsound.com/sos/apr99/articles/recstrings.htm

IMO as far as sampling is concerned microphone proximity effect plays a large part in the noticeable noise produced by bow hair, rosin and string interaction.

The EWQLSO Platinum Library I use doesn't produce the scratchy sound nearly as much as other libraries I've heard and have used. Not saying it's the best but it doesn't have the "scratchy sound".
EQ_ing the Notion String sounds as most suggest may help. VSL seems to have a plug-in or EQ specifically for there string library that addresses this issue. It's mentioned somewhere in the vi-control.net discussion link above.

,Johnny

Johnny wrote:Clearly this is a topic widely discussed by many. A noisy unnatural sample is quite distracting.
"Sample noise" can certainly distract from a creative idea.
It's just not pleasant to work with.

Absolutely!

This is encouraging, since initially in the "Clouds" (JohnF) topic I was somewhat reluctant to mention the "white noise" aspect, with the concern being that it might be a bit obsessive; but after pondering everything for a while, I decided that it was a relevant observation, since (a) I like the composition and (b) it is not a composing problem . . .

JohnF wrote:Surf - I think you're right, it's the upper register of the cello's that's creating that sound. I did compare with my exported .wma and it is more prevalent on soundcloud, due to compression I suppose.

[NOTE: I have a calibrated full-range studio monitor system, which maps to being able to hear everything--at least within the constraints of my hearing abilities, which are good to about 14,000-Hz, more or less--so in some instances I make observations based on the premise that folks who do not have calibrated full-range studio monitor systems cannot hear some of the things I hear, which is a fact and is something I am discussing in another new topic ("Psychoacoustics and Digital Music Production"), since creating well designed auditory illusions requires being able to hear the audio accurately. As an example, if a studio monitor system goes only so low as 50-Hz, then it is not going to reproduce lower frequencies accurately, which for the most part eliminates the low-pitch "A" and low-pitch "E" strings of a string bass or electric bass (55-Hz and 41.203-Hz, respectively). The lower frequencies are there, but they are not there at equal loudness, which is a significant problem that causes people to use the "Missing Fundamental" auditory illusion to compensate, with the consequence that midrange and high frequencies become skewed, because creating the "Missing Fundamental" auditory illusion requires adjusting the midrange and high frequencies, hence one of the indications is that mixes done this way only sound good in the specific studio played through the particular studio monitor system. Explained another way, if you want songs to sound at least pretty good everywhere, then you need to work with a calibrated full-range studio monitor system, since it is the only way you can hear everything accurately, noting that headphones will not work for this purpose . . . ]

Psychoacoustics and Digital Music Production (Notion Music FORUM)

Johnny wrote:IMO as far as sampling is concerned microphone proximity effect plays a large part in the noticeable noise produced by bow hair, rosin and string interaction.

I think so, too!

As noted in one of my previous posts to this topic, I hear relics of compressor-limiter usage in some of the string samples I have here in the sound isolation studio (NOTION 4, Kontakt 5/VSL Strings, MachFive 3, Garratin Instruments for Finale 2012); and using a compressor-limiter magnifies the problem . . .

[NOTE: The "Violins KS" samples provided by Garratin Instruments for Finale 2012 are very nice and do not have the white noise tails, but there are problems with some of the other strings, solo and sections; where one of the solo violin playing styles has a few noisy attacks (or "heads") rather than noisy tails (or "releases"). And there are nice notes for all the other virtual instrument string libraries (NOTION 4, Kontakt 5/VSL Strings, and MachFive 3), as well as troublesome notes, which among other suggest the possible solution of using specific virtual strings based on the key of the song, where some key signatures probably have no string samples that are troublesome for a particular virtual instrument string library but are troublesome for other key signatures . . . ]

idiotSavant explained it another way, which I think provides a clue to a few potential solution strategies:

idiotSavant wrote:. . . listeners in "real" environments never hear the instrument this way - they hear it with frequencies attenuated by the music stand, by the player in front of the player, by the players in front of those players, by the diffusion created by the materials in the hall or room, by the audience in front of the listener . . .

When a song is played in a concert hall by a real orchestra and there is an audience, the ambient noise is more than sufficient to mask nearly all the noise produced by the orchestra, and the result is the ability to have smooth strings when smoothness is desired . . .

In the digital music universe, everything is different; and the reality is that nothing is real . . .

Instead, everything is virtual, which is fine with me, but the rules for virtual reality are different . . .

In this respect, several of the posts in the discussion found at the first link you provided note that the perspective is different depending on the genre, which makes sense to me; and the specific example consistently was Pop music, where the focus is distinctly different and is something I noted, as well . . .

From my perspective, the relevant classic example in the Pop genre is "Eleanor Rigby" (Beatles), and while this was done in the analog magnetic tape universe with musicians playing real instruments, (a) there is no white noise and (b) George Martin specifically had the various strings recorded with very close microphone placement to such an extent that in an interview, he noted that the microphones annoyed the musicians and in some instances required them to devise novel ways to hold and to play their instruments . . .

"Eleanor Rigby" (Beatles) ~ YouTube music video

I agree that microphone proximity is a significant aspect of the problem, but it is not solely a matter of microphone proximity, otherwise there would be white noise at various times in "Eleanor Rigby" . . .

In the Pop genre, this is not a problem, because the general rule is that everything needs to be heard clearly, which basically maps to "loud" . . .

But for genres where having a wide dynamic range is important, this is a problem; and as time allows, I am planning to do some experiments, where one might suppose that the worst case scenario involves getting the NOTION 4 generated audio into a Digital Audio Workstation (DAW) application via a ReWire 2 session and then manually editing the tails of notes in the audio clip editor . . .

However, I think there probably are other strategies that are not so time intensive . . .

Lots of FUN!

The "proximity effect is most of the problem" IMO.
I use a DPA 4099b Super-cardioid microphone on my bass.




Because the microphone is positioned facing 180 degrees away from the instrument strings and it's proximity to the top of the body most all of the bowed sound is much lower in amplitude in comparison to the sound emanating from the instrument body which defuses the bow's sound yet carries the high energy string vibration.
The body of the instrument is what the listeners ears hear in general. The instrument body producing the sound fundamental and partials from sympathetic resonance.
I don't get that scratchy sound artifact in my captured recorded sound. And if I were to create samples from my recordings they wouldn't have that noisy scratchy artifact in them either. That said, neither should costly expensive sound sample libraries where the goal and intent is for semi-high end or hi-end recording purposes.

This diagram shows the pattern that the DPA microphone uses.
Notice the "off access" frequency response and roll-off at the 180 degree position. Believe me the bowed noise level is ultra-low in the captured sound. Again, the positioning and proximity have a large effect.If samples of string sections used this type of sound capture I doubt we'd be having this discussion. At least not about unusable scratchy sounding noise in samples intended to be used to create semi-pro/professional recordings.
This type of microphone would be of great benefit in capturing "solo" instrument sounds and even section sounds if they were mixed effectively.



,Johnny

P.s. The EWQLSO platinum section samples I use don't seem to have that much noise. I am going to try EQ processing my other sound libraries that have measurable noise levels to see if they might be better in a mix.
Sound Forge is a great tool to test this.
I'm hoping that using multiple sound sample libraries will produce an even richer more varying sound i.e. carefully mixing EWQL, Notion and Garritan samples to get a more complex and varied wall of sound.

Johnny wrote:The "proximity effect is most of the problem" IMO.

I prefer the perspective that attributes the problem to the combination of (a) instruments, (b) musicians, (c) studio or concert hall location and room acoustics, (d) microphone selection and placement, (e) recording technique, (f) mixing and mastering technique, (g) digitizing technique, (h) virtual instrument engine rendering technique, (i) producing, mixing, and mastering techniques when using the sampled sound library, and (j) the key signature, although the last two are after the fact . . .

Technically, the "white noise" problem has nearly nothing to do with the proximity effect, but I understand the way you are using the term "proximity effect", which for reference is not the standard usage . . .

Specifically the standard usage of the term "proximity effect" is to refer the the fascinating phenomenon that enhances the bass when a microphone is used very near or "proximal" to a source . . .

The proximity effect in audio is an increase in bass or low frequency response when a sound source is close to a microphone.

Proximity Effect (Wikipedia)

[NOTE: It helps to understand that I study dictionaries and generally am fascinated with the different ways words are used. And I make an effort to be as precise as possible in my use of words, which can be a bit confusing when folks are speed-reading--a problem that I also have occasionally even though I make an effort not to read too rapidly. Specifically, in the earlier post when I wrote that I agreed with your observation, I was agreeing about "microphone proximity" being a significant problem, which is entirely different from "proximity effect". Additionally, as the consequence primarily of having an epiphany during a course on Boolean Logic, if the order taker at McDonald's asks, "Is this for here or to go?", then my usual reply tends to be "Yes!" if I sense that the order taker is human, hence naturally will follow-up with a question for which the answer actually provides useful information, which basically maps to asking half of the question, for example, "Is this for here?'" or "Is this to go?" . . . ]

I think you are referring to microphone type and placement; and this is what I consider to be part of the activity I call "recording", which is distinctly different from the activity I call "digitizing" . . .

If instead of using the term "proximity effect", we use "microphone selection and placement", then I think we generally agree . . .

[NOTE: I use the word "selection" in "microphone selection and placement", because there are different types of microphones, as well as different patterns, and it is not simply a matter of one aspect, hence the need to select a specific set of aspects. It is reasonable to presume that a professional recording studio will not be using high-impedance dynamic microphones for recording strings, but there are several types of low-impedance microphones, which is another aspect. Here in the sound isolation studio, I use condenser microphones and ribbon microphones, and they require phantom power, but it depends on what is being recorded, as well as its dynamic range. Some condenser microphones and nearly all ribbon microphones are very sensitive, which is good for recording singers, but these types of microphones do not respond so well to loud noises and actually can be damaged by loud noises. It also is important to provide a bit of instruction for singers regarding the correct ways to "work" a condenser microphone or ribbon microphone, but the rules are not difficult. High sensitivity is good for recording singers but bad for recording kick drums, which is where low-impedance dynamic microphones are handy. There also are different microphone directional patterns, which is yet another aspect to consider when selecting a microphone. Additionally, microphones have different frequency responses and some of them have different types of padding or damping options, which are two more aspects . . . ]

Using the wrong type of microphone and putting a microphone in the wrong location are two significant causes of problems when recording, but recording is only the first major step of the overall process, although it certainly is an important part . . .

The additional major steps are (a) mixing and mastering the recorded audio where the goal is to create a set of raw audio clips, each of which is a sampled note or perhaps a series of sampled notes and (b) digitizing the recorded and mastered raw audio clips . . .

The "mixing and mastering" aspects refer to the need to ensure that the raw audio is consistent, where for example all the audio clips for notes played in the staccato articulation actually are played staccato, but also that the dynamics are consistent, as is the tone, and so forth. Some of this things require a bit of producing and audio engineering even though the "song" actually is a single note . . .

In this context, "digitizing" refers to doing the additional editing required to transform the raw audio clips into logically organized digital samples, an activity which typically involves editing audio clips to some type of standard, which sometimes is one of the WAVE formats, but can be just about anything and in some instances is a proprietary standard . . .

[NOTE: As best as I can determine, NOTION 4 uses "WAV16" for its bundled sampled sounds, and I base this on information found in the XML data section of NOTION 4 *.prox" files . . . ]

Native Instruments provides standards for the types of sample libraries and other files with which it works; and for example Bolder Sounds is a third-party company that produces sample sound libraries that work with Kontakt 5. There are quite a few other third-party companies that produce sample sound libraries for Kontakt, which is one of the things that makes Kontakt attractive . . .

Johnny wrote:I use a DPA 4099b Super-cardioid microphone on my bass.

Very nice!

Johnny wrote:The EWQLSO platinum section samples I use don't seem to have that much noise. I am going to try EQ processing my other sound libraries that have measurable noise levels to see if they might be better in a mix.
Sound Forge is a great tool to test this.
I'm hoping that using multiple sound sample libraries will produce an even richer more varying sound i.e. carefully mixing EWQL, Notion and Garritan samples to get a more complex and varied wall of sound.

This makes sense to me!

My use of the term "White Noise" is in reference the mechanical sound present in the process of sound generation from a bowed string instrument. This "noise" is for the most part undesirable because of it's "percussive" non-harmonic sound property. I therefore use "White Noise" to define it since it is constant in a poorly sampled sound file (instrument or audio file).

"Proximity effect" causes a increase in bass frequency response. That is why I'm suggesting the effect covers or masks the lower amplitude high frequency of a bow sliding (scratching) across a vibrating string not accentuating it.
Thus the fundamental sound frequency and it's harmonic partials are recorded amplified and produce louder sound levels than the noise that surrounds the source.
On my bass if I move the microphone farther away from the body the frequency response changes and a wider range of the mechanical noise level will be heard in the reproduced sound.



,Johnny

Johnny wrote:My use of the term "White Noise" is in reference to a mechanical sound which is always present in the process of sound generation from a bowed string instrument. This "noise" is for the most part undesirable because of it's "percussive" non-harmonic effect. I therefore use "White Noise" to define it since it is constant in a poorly sampled sound file (instrument or audio file).

This is the way I use the term "white noise", and there is an example audio clip of white noise on the right side of the page . . .

[NOTE: White noise is the noise that ocean waves make at the beach, and it also is made by wind blowing through leaves and by rain. In some scenarios it is comforting and is used in portable machines that mask city noises and by doing so help people sleep without being distracted. It also is part of the sound made by waterfalls. However, at higher volume levels, it is annoying . . . ]

White Noise (Wikipedia)

There are several different colors of noise, as explained at this link:

[NOTE: Pink Noise is used when configuring a sound system to tailor it to a specific venue, and this also is the case when calibrating a studio monitor system using a real-time analyzer and equalizer like the Behringer DEQ2496 UltraCurve Pro Mastering Processor (a personal favorite) . . . ]

Colors of Noise: Technical Definitions (Wikipedia)

Johnny wrote:"Proximity effect" causes a increase in bass frequency response. That is why I'm suggesting the effect covers or masks the lower amplitude of a bow sliding across a vibrating string not accentuating it.

In some respects, this might work for a string bass and possibly for a cello, although not so much for a viola or violin; but the basic problem with the hypothesis is that the offending noise is not low-frequency--instead it is midrange and high frequency . . .

Intuitively, I would expect less undesired bowing noise with the DPA 4099b pointed downward rather than upward, and this is the way it is shown in the photograph . . .

There is a good bit of room for doing the same thing with a cello, but the available room decreases significantly with a viola or violin, as does the ability of the instruments to produce bass tones . . .

There are similar small microphones for violas and violins, but they usually are used in Pop, Country Western, Bluegrass, Rhythm and Blues, Jazz, and other genres where there is a sound reinforcement system and all the musicians and singers are amplified for purposes of performing before an audience; but this certainly can include recording studio usage, although not necessarily, since there are different considerations in the recording studio scenarios . . .

As best as I can determine, most of the sampled sound libraries are recorded either (a) in concert halls under controlled conditions with professional recording equipment or (b) in professional recording studios with essentially the same type of equipment; and the fact of the matter regarding microphones is that some of them cost more than everything in a typical home recording studio (analog or digital), and this includes instruments, microphones, computers, software, various hardware, furniture, acoustic treatments, studio monitor systems, cables, and so forth . . .

[NOTE: The highest price microphones are older, but even with new microphones there are quite a few in the $3,000 (US) to $7,500 (US) range. Generally, I think that a reasonably complete home studio tailored primarily for working with virtual instruments; a few but not so many real instruments; and real singing costs somewhere in the range of $15,000 (US) to $25,000 (US), including a few professional quality instruments used primarily in popular music genres. On a related note, I like some of the inexpensive MXL condenser and ribbon microphones that cost approximately $50 (US) each when Musician's Friend has a super discount sale, which is considerably easier of the budget . . . ]

If the "white noise" problem were universal, then it would happen with recordings of orchestras and string ensembles in a variety of scenarios, including concert halls and professional recording studios like Abbey Road Studios, but it does not appear to happen in any of those settings--except in the specific scenario where instruments are being recorded for purposes of making sampled sound libraries . . .

SUMMARY

Everything about microphones is important, and this includes the proximity effect; but this is part of the "recording" step, which is not the only step . . .

Once the raw audio is recorded, produced, and mastered for purposes of being the input to the next step, which is the digitizing step, the factors change; the focus is different; and microphones are not part of this step . . .

Whatever the microphones did already is done; and it is made permanent as part of the audio clips--specifically the audio clips that are the output of the recording step but are the input to the digitizing step . . .

In other words, at the recording step, microphones are a variable; but when the recording step is completed and the audio clips are finalized, the microphones become a constant rather than a variable . . .

Based on what I have heard so far, the way samples are recorded in some instances might be a problem; but there are other indications that the real problems most likely occur during the digitizing step rather than during the recording step; and some of the problems occur based on the ways end users produce songs that have virtual instruments, which is something that JohnF suggested as a possibility, where the example is adding a compressor-limiter, which typically would raise the level of the tails of notes arbitrarily and by doing so would amplify the perhaps otherwise tolerable white noise in the background, thereby moving it to the foreground, noting that I do not think this is a matter of dithering or not dithering . . .

Lots of FUN!

P. S. Regarding the graph of the proximity effect and the DPA 4099b condenser microphone, it is important to consider the equal loudness curves when putting the information into context . . .

[NOTE: Based on the information provided in the "proximity effect" graph for the DPA 4099, it has no proximity effect below 40-Hz, which essentially is just a tiny bit below the low-pitch "E" string (41.203-Hz) at standard tuning ("Concert A" = 440-Hz) and the proximity effect from 40-Hz to 80-Hz is represented by dotted or dashed lines rather than the solid lines that start at 80-Hz, where the key bit of information is that while the "frequency range" of the microphone is stated to be "20-Hz to 20,000-Hz", there also is a second frequency range stated for plus or minus 2 dB, but it is 80-Hz to 15-kHz, which is fine with me, although I am a bit puzzled by the dotted or dashed lines and what they indicates as contrasted to the solid lines . . . ]





Unless the bass notes are in the range of 85 dB SPL to 90 dB SPL, they will not have much "masking" affect on midrange and high frequencies . . .

The troublesome frequency range with respect to noise is the "dip" centered at approximately 3,000-Hz, which is where normal human hearing is most sensitive, and this is part of the midrange region . . .

Bass also is troublesome, but it is a different type of trouble, which specifically is that as shown by the equal loudness curve(s), it requires considerably more power to hear bass at the same perceived loudness as midrange and high frequencies, which among other things is the reason that with only one exception (JBL Professional M2 Master Reference Studio Monitors, which cost $20,000 [US]), commercial-off-the-shelf (COTS) studio monitor systems need to be augmented with a pair of deep bass subwoofers, which essentially double the price of a studio monitor system. Absent the pair of deep bass subwoofers, COTS studio monitor systems are not full-range, which is a big problem, because the only way to hear music accurately when producing, mixing, and mastering is to listen to it played through a calibrated full-range studio monitor system. Headphones do not work, and there is no other way to hear the music accurately . . .