Blackness - how quiet does it need to be?

Jitter in playback of digital source is basically noise proportional to amplitude of the signal and not detectable without signal.

If I got this right, the equipment is the conduit and the music is the signal. That signal can sound cleaner depending on the equipment. One familiar enough with a recording can discern a lower noise floor using the same signal (music) while comparing different makes of equipment. And a really good piece of equipment can detect the noise floor limits of a so-so recording. It works both ways as it would be difficult to decide if you have a really good recording if your equipment is not up to par with the recording.

Familiarity with your equipment is necessary. Now if you hear a slight hiss with your ear at the speaker (really slight) and not at your listening spot, I wouldn't worry. If your system were to mask the hissing of a poor recording, I would worry. Sound to me like your system is okay.

On Diana Kralls Love Scenes CD I can hear when she comes in and out of the music. She is recorded separately as the noise floor on her voice is messier compared to the rest of the recording. Right before she comes in, the noise floor is comparatively low and when she does come in, you can hear the hiss levels rise along with her voice. When I first heard it, it was quite disconcerting. As she stops singing, the hissing goes away to reveal a quieter background as her backup musicians play on.

That's why I tend to go for better recordings which can really limit what I listen to. That is the price you pay for your endeavors.

All the best,
Nonoise

This is not complicated. Your right on the mark, Last Lemming. The recording has nothing to do with it. We're talking about gear. However, noise floor is not the only factor influencing low level resolution. Typically, if you put your ear up to the speaker at 0-half volume and don't hear anything except the tweeter, the amp has a low noise floor. Especially on an efficient speaker like a Klipsh. If it's noisy, it stands to reason that resolution is going to suffer commensurate with the level of that noise since the amp is going to amplify whatever it's being fed. At "couch" distance it's irrelevant unless of course you can hear it there! Differences are apparent when A/B ing amps with the same recordings and other gear. So in answer to your question; yes, they mean the equipment is quieter. From your description it looks like you have a nice quiet amp.

Csontos - CDP is also an equipment and can be very noisy when playing but quiet when not. That way noise is undetectable without music (proportional to volume). In addition amplifier might create noise of its own by intermodulation etc. (speakers can also intermodulate).

Last_lemming, your amp might appear as quiet but in reality can be very noisy.

Sure, but I was referring exclusively to amps of considerable quality where differences are relatively small.

The OP's question, regarding how noise that is not audible beyond a very short distance from the speaker when no music is playing might have audible significance, is an excellent one, that I've pondered myself at times.

I think that Kijanki's answers are on the mark.

In the digital domain, the explanation is easy, namely jitter effects, as he indicated.

In the analog domain, it is not that clear, but the one explanation that occurs to me relates to intermodulation effects, as he also indicated. The ear is much more sensitive to some frequencies than to others, as can be seen in the figure in this Wikipedia writeup on the Fletcher-Munson Effect. Non-linearities in the speakers, and perhaps also in the electronic components that are in the analog signal path, will result to some degree in intermodulation effects, producing (at very low but conceivably significant levels) new frequencies corresponding to the sum and difference between all of the frequency components that are present. Hiss typically contains a mix of essentially all frequencies within some broad range, especially in the upper treble region (and beyond, at ultrasonic frequencies that are inaudible in themselves). Perhaps intermodulation of some of the frequency components of the music (and perhaps also frequency components of recorded noise, tape hiss, and/or LP surface noise) with those upper treble and ultrasonic system noise components results in difference frequencies in the lower treble or mid-range regions, where the ear is more sensitive.

That's my speculation, anyway, elaborating on what IMO were excellent answers by Kijanki.

Regards,
-- Al

I'm trying to understand the concept of the CDP contributing to noise while playing vs. when not. How would one hear this noise while the CDP is playing? In my case it's a transport to DAC. What I do know, is that on high quality hi'er res material like 96k/24 bit, on quiet passages, the hiss level hardly goes up, even on more than average listen levels. In other words I can only hear the hiss if Im very near the speakers, but certainly not when sitting. Now this only holds true for well recorded sources. Of couse I have hi res recordings that have hiss in the recording like my Miles Davis, Kind of Blue that I can hear from my couch, but this is not the same thing that Kijanki is speaking of.

I don't quite understand what IMD is, in layman terms. So Im off to look that up.

Last_Lemming I'm hesitating if I should even start since Al can explain it much better and more coherent way. If he does follow his explanation.

In both cases, digital and analog, we're talking about modulation that creates new frequencies.

Two frequencies in presence of nonlinear element (nonlinear transistor, nonlinear motion of speaker's membrane) produce additional signals of frequencies that are sum and the difference of original two frequencies. You can think of it as noise since it wasn't in the music and amplifier or speaker manufactured it. Great amplifiers or speakers modulate very little but we also "learned to listen" and can detect smallest differences in clarity.

Time jitter of your transport digital signal when playing single frequency also creates two additional frequencies (sidebands), but this time they are spaced on both sides of original/root frequency by modulation frequency distance (frequency of time jitter). Al mentioned once, that this modulation frequency can be completely random at the moment (random noise affecting digital signal timing) and we call it uncorrelated or can be cause by particular frequency like noise from switching power supply - then it is correlated. Usually it is mix of both. When complex signal (music) contains a lot of frequencies there is a lot of byproducts - basically a hash proportional to amplitude of the music. When music stops hash stops. It is detectable only as a lack of clarity and as such affects pretty much everything from tonality to imaging. Jitter usually has very small amplitude and byproducts (sidebands) are very, very small but still audible since they are not harmonically related to root frequency (like in harmonic distortion) and our ears are sensitive to it.

We're dealing wit two aspects of jitter: amplitude (modulation index) and the frequency. Small modulation frequency (like 60Hz) creates sidebands near root frequency that are easily masked therefore is less dangerous but high frequencies of the jitter create byproduct far apart from the root frequency and very audible. Amplitude of the jitter affects amplitude of the byproduct up to the point where even more than two sidebands can be created. Bad CDP have often a few nanoseconds jitter amplitude while good transports perhaps one tenth of it. I shouldn't really attribute it to transport but rather to system (transport, cable, DAC, ambient noise, power noise etc.)

Al, did I miss something?

Thank you Kijanki. I do believe I get the gist in what you are saying. Since there can only be jitter when a musical signal is present, then noise can't be heard as a "hiss" per say, but as distortion on some level in the music reproduction.

Al, did I miss something?

No, I think you covered it well. The only thing I would add is to re-emphasize the point I had made that some of the spurious difference frequencies that may be created, especially as a result of intermodulation of recorded hiss or music and system-generated hiss in the speakers or other analog parts of the system, may be at frequencies that are more audible to us than the original frequencies.

With respect to the purely analog parts of the system, I'll qualify my comments by saying that I have no particular quantitative feel for how audibly significant noise intermodulation may typically be. (The significance of jitter at the point where digital is converted to analog is well established, of course). But what I am saying is that if the subjective perception of background blackness can in fact be improved by reduction of noise that is essentially inaudible when nothing is being played, the explanations that Kijanki and I have offered are the only ones that come to mind.

Best regards,
-- Al

Is the noise floor system dependent? or is it directly the sum of the noise floor of the individual components?

Thing is, when I turn on my system and no music is going through it, I can't hear anything, unless I put my ear right up to the speaker and the AC isn't running and the fan isn't on, etc. And with music on the only thing I hear is any recorded hiss that might be from the recording. So what I don’t get is when they say a piece of equipment sounds quieter, do they mean somehow that the hiss on the recording is lower? I can’t see how that would be possible...

I agree that, on the face of it, this seems impossible. But I’ve had a number of experiences that have led me to conclude otherwise.

For a long time I thought that my system was dead quiet, due to the fact that there was no audible noise floor when the system was on and no music was playing. As a result, whenever I heard hiss during playback, I believed that the hiss was in the recording.

Several months ago, I made an effort to reduce RFI/EMI in my system, which is discussed in this thread. One of the consequences of those changes was that hiss during playback was dramatically reduced. But the reduction of hiss was audible ONLY WHEN MUSIC WAS PLAYING.

From this I concluded that a significant fraction of the noise floor that I thought was IN THE RECORDING was actually IN THE SYSTEM. But the noise floor in the system was not perceptible when the system was "at idle."

I know that seems paradoxical, but that's what I experienced.

Bryon

Bryon, Jitter modulation produces all sorts of garbage across all frequencies raising noise floor. Normally you would not be able to hear it as a hiss because it would be covered/masked by loud signals and not audible during gaps or soft passages. Let's assume that jitter product is for instance, always -60dB down from the loudest music level. When you play softer passage jitter products are -60dB down from that being inaudible. The louder you play the more apparent it is showing as a lack of clarity (loud trumpet is not as pure as it could be etc.)

One possibility is that they suppress microphones' hiss before and after but I would agree with Al, that what you hear might be differential of high inaudible frequencies modulated on the tweeter or amplifier. These frequencies come from many sources including digital playback itself (quantization noise). New techniques of oversampling push quantization noise outside of audible range but it might come back when system is not perfectly linear and has additional noise sources (poorly filtered switching power supplies, RFI pickup etc). These high frequencies can have extremely high amplitudes. I remember that some of SACD players have bandwidth limiting switch (filter) to protect weaker tweeters from overheating just from inaudible quantization noise. Stereophile review mentioned that it was left by manufacturer to customers with warning of possible risk of turning off the filter switch. Tweeter membrane does not move much at very high inaudible frequencies but as long as it does it will modulate because its displacement vs. voltage is never perfectly linear.

Guys,

I'm a long time audiophile and I just wanna say that this has been one of the most informative, easy-to-read threads about "noise" as it pertains to equipment that I've ever read. Kudo's to all and thanks for taking the time

Vance

There's one more thing I'd like to add to this and it's from another thread about believing in magic. It's in the last paragraph before the last line.
http://www.physicsforums.com/archive/index.php/t-122756.html

This lead me to ask if you hear this hiss with all recordings or with just some.
I knew that some recordings have a lot of processing but I never knew that lots of recordings have white noise pumped into them. You could very well be hearing negative artifacts from the recording process as well as the jitter errors mentioned above.

All the best,
Nonoise

It may have been covered to some degree by your discussion in the
context of digital, but to me, i think it goes way beyond hiss- it is the
resolving power of the system to pluck out the nuances of the music or
voice against the background at low levels. The more 'absent' the
background, like ambient noise, the better you can hear all the little things
that are going on in the recording. (rough analogy would be trying to hear a
conversation in a noisy room, compared to quiet space). i think there is a
tendency to
turn up the volume to get more reality, and while that makes it 'louder,' it
doesn't make it more real sounding- in fact, sometimes the opposite
happens; the natural volume of the recorded performance is exceeded and
rather than sounding more lifelike, the system sounds like a sound
reproduction system. So, one 'test' (not really a test in the scientific sense,
but sort of a 'by thumb' way of getting the sense of a system) is how well it
resolves stuff at low volume. Not 'detail' per se, but the nuances, the
shading, the dynamics (differences between loud and soft) at a lower
overall volume level. if a system can do that well, i think it has a low 'noise
floor' in the sense that the music is presented with less ambient junk
around it and
stands out more clearly against the so-called 'black background.'
The point above about the 'natural volume' is a little different, and is just
that there seems to be a db level on playback where the recording just
sounds right, given how it was made, your system, room acoustics, the
type of music, etc. Kinda 'clicks in' just like getting the VTA on a tonearm
has a 'right spot' for a particular record.
And while the recording itself may have nothing to do with the noise floor of
the system, a system with a lower noise floor should be able to benefit
more from a good recording, if the above makes any sense.
my 2 centavos.

"I knew that some recordings have a lot of processing but I never knew that lots of recordings have white noise pumped into them."

The hiss I'm talking about is only on a few select cd recordings of older music. Like my Miles Davis, Kind of Blue HDCD. Most music sounds pretty quiet.

Well, there goes that theory.

I hear an ever so slight hiss or veil (for lack of a better word) on my Kind of Blue CD as well. It's a Columbia/Legacy SBM (Super Bit Mapping) recording.

I've read where Steve Hoffman said that older masters from tape and tube (if they were done that way) don't make the transition to a newly remastered state if done with SS equipment unless certain software is used, correctly. He said it will not sound quite right. Maybe that's it.

All the best,
Nonoise

But tape hiss from an older recording is not really about the 'noise floor' of the system.

Vance (Vhiner), thanks for your comment. This has indeed been an excellent discussion.

08-09-12: Nick_sr
Is the noise floor system dependent? or is it directly the sum of the noise floor of the individual components?

Nick, yes, there is a good deal of system dependency involved.

With respect to the analog parts of the signal path, the audible significance that ultimately results from noise that is generated, introduced, or picked up at any point in the signal path will vary depending on the relation between the signal as it exists at that point and that noise, especially the ratio of their amplitudes. Which in turn will depend on the gains, sensitivities, and output levels of other components in the signal path.

With respect to digital parts of the signal path, many variables and interactions that are relevant to noise sensitivity and its ultimate effects on jitter come into play between the transport or other source component, the component in which D/A conversion is performed, and the cable connecting them. Obviously the jitter rejection characteristics of the DAC are one. Also, the risetime and falltime of the output signal from the transport, which are parameters that are usually unspecified, will significantly affect the consequences resulting from noise that is present at the interface between the transport and DAC. Impedance mismatches between the two components and the interconnect cable will inevitably be present to at least some small degree, and might affect the consequences of noise at that interface, if the mismatches result in distortion of critical parts of the waveform. Also, noise that may be introduced by ground loop effects will be affected by the interaction of various technical characteristics of both components.

Best regards,
-- Al

I don't think Steve Hoffman was referring to the tape hiss from the original recording. That could be somewhat masked, but not entirely eliminated. What he referred to was the difficulty of taking an original master and getting it sound right on modern equipment.

Here is his view:http://www.tapeop.com/articles/85/steve-hoffman/.

I guess the only alternative is to try the DCC CD version that he made of the Miles Davis 'Kind of Blue' CD.

All the best,
Nonoise

Many recordings were made with pre-emphasis to greatly reduce tape noise in playback. CDP or DAC has processor for it (de-emphasis) and every CD has bit switching it on or off in CDP or DAC. Today almost all recordings are digital and pre-emphasis is never used. Because of that many CDPs or DACs don't even have de-emphasis. None of Lavry DACs, for instance, has de-empasis but my Benchmark DAC1 does. If your CDP or DAC doesn't have de-emphasis and you listen to one of few recordings with pre-emphasis it will sound noisy. I'm not sure if tapes were recorded with pre-empasis (10db increase between 3kHz-10kHz) before invention of CD. Some people tested rippers and found that EAC does not correct for pre-emphasis (leaves recording intact) while Itunes does apply correction. Also some MP3 compressors like for instance LAME have de-emphasis option.

It sounds like we are victims of the times. Not everything is keeping apace and older formatted music is not sounding right, all the time, on modern equipment.

I like Hoffman's story about trying to figure out how to get Hotel California to sound right and it took getting similar vintage JBL speakers to hear how the original engineer mastered it to understand why it was so bass heavy on modern speakers. He had to work backwards with that understanding to do it.

One needs an extensive historical understanding of recording, techniques, playback, equipment and what have you to bridge that gap of old and new in order to preserve the old and enjoy it anew.

All the best,
Nonoise

great thread guys!. something i've often wondered about myself but never asked about. a very interesting read!

I read the Steve Hoffman piece, and my takeaway is as follows (but I'm still not sure this has anything to do with the 'noise floor' of a playback system):
1. given the original limits of record cutting and playback equipment at the time, stuff was heavily EQ'd and dynamics were often constricted, essentially 'gain riding' to bring up the level of the soft passages to overcome inherent noise in vinyl playback.
2. trying to remaster these older recordings on more modern, wider bandwidth equipment, including using differeent playback equipment than a circa mid-70's JBL monitors, means that some of the fiddling originally done with the tapes, described above, has to be 'undone.'
A.If I have that right, I'm still not sure what that has to do with the noise floor of the system- perhaps you are saying that these older recordings are going to sound noisier on a modern system, whether or not remastered. I certainly hear that on old RCA records, which had notoriously noisy surfaces (but gloriously natural sound in many cases).
B. If remastered, there is a danger that the life of the recording can be lost. I have found this- and i don't think i'm alone, in the case of a number of 'remasters'- the other day, i listened to a recent remaster of Janis Ian's 'Between the Lines,' it sounded sterile, lifeless. I pulled out a 1975 pressing that i bought recently, sealed, and the magic was there. Same experience with other audiophile warhorses, like Tea. The old UK pressings are far more natural sounding to me than the reissues or remasters, including the old, expensive UHQR.
C. Granted, I'm listening to vinyl only and don't have the technical expertise to address what happens in a digital processing environment, so I'm not extending my comments to that.
D. But having said all of the above, how does this relate to the noise floor of the system itself?

Whart,
I think your "A. If i have that right..." explains your "D. But having said that...".
You were correct at "A" .
If I understand this.

All the best,
Nonoise

NoNoise, I'm not trying to be contentious. But, I don't think system noise
floor is entirely program material-dependent. If you had a very good
recording (let's stick with vinyl since that's what I'm more familiar with),
quiet pressing, good quality recording, etc., a system with a lower noise
floor is going to reveal more that is on that recording - more music -and all
the descriptors that come with good fidelity- dynamic shadings, tonality, etc.
with less ambient 'gunk' (my scientific term) from the system itself to
obscure what is on the record.
I base this on my experience in using horns with extreme efficiency- where
the sound of the electronics, the AC power, etc. is revealed with
ruthlessness unless the system is very, very quiet. I am raising this not to
advocate a position, or to say that horns or vinyl or whatever is better, but
just to focus specifically on the question of 'noise floor' of the system itself
and how it bears on musicality of home reproduction. And, as mentioned, i
judge it not from how 'noisy' the system is at steady state without a record
playing (although that can be relevant i guess) but how revealing the
system is at low volume with music playing.

BTW, i appreciate your response.

Whart,

I think I see your point: system noise floor is independent of whatever is playing.
And until something is playing, you can't really judge or appreciate it.

Dense, that I am, I still believe that a system such as yours (and I'd like to think mine) with a very low noise floor will reveal the shortcomings (higher noise floor) of a recording rather than mask it.

I think we are both saying the same thing but in different ways or I'm not quite getting it. Oh well, off to work and have a great day.

All the best,
Nonoise

Blackness has to do with intermodulation distortion. This is a property of all preamps and amplifiers. The lower the IM, the more the unit will be perceived as having a 'black background'. In principle it functions the same way in analog gear as Kijanki described in digital gear.

To lower IM, you have to increase linearity. The supplies have to be quiet, and incapable of modulation due to signal level.

Of course, the equipment should be low noise too :)

Note:

1. given the original limits of record cutting and playback equipment at the time, stuff was heavily EQ'd and dynamics were often constricted, essentially 'gain riding' to bring up the level of the soft passages to overcome inherent noise in vinyl playback.

This statement is incorrect. LP mastering is one of the more unlimited audio processes in existence. The LP cutter has dynamic range far in excess of digital or tape and challenges the best microphones. It also has crazy bandwidth which is why CD-4 LPs were able to be made in the 1970s (CD-4 used a 50KHz subcarrier FM-stereo modulated into the grooves).

The limit of the LPs is in the playback. A cutter can make grooves no tone arm or cartridge could ever hope to track and it can do so without a hint of overload. But such grooves might throw a stylus and arm right across the record. All the processing comes in (dynamic limiting and the like) so you don't have to be all that careful to exceed playback limitations. Its just the record labels being lazy.

There is signal and there is noise. Nothing else.

Signal is the good. Noise is the bad.

If more noise leads to better signal, that does not make the noise good. It just means there is a correlation between the two for some reason and the benefits of the signal outweigh the disadvantage of the noise.

If I hear ANY noise in my system, it bothers me. It means something is not right and needs to be fixed. Even noise that you may not hear explicitly is detrimental to what you might be hearing otherwise.

In summary, noise is something that must be dealt with but it is ALWAYS BAD, NEVER GOOD.

Ralph, i was summarizing what was said in that Hoffman article that Nonoise
posted a link to. Perhaps I summarized it inaccurately, to the extent that the
limits placed by the engineers were solely due to playback limitations, not to
limits in what could be cut into the groove in the first place. Here's one relevant
passage: "The one nice thing about LPs is that there's a certain set of rules and
regulations that you have to follow. You can't add a ton of high end and you
can't add a bunch of low thumping bass, because you just can't! Unless Einstein
was all wrong, there's just a certain type of groove that you can cut. And that
groove can only have so much of this and so much of that. In one way, that's a
good thing, because that keeps engineers honest. They can't screw around. On
the other hand, it's a pain in the butt. You want to get as much level as you can
on the record, so the music is louder than the surface noise of the record. So
you want to get it as loud as possible. But in order to preserve dynamics, you
need to make sure that you're not overmodulating anything. In the '50s and '
60s the engineers just reduced the dynamic range by using analog
compression, which is what is on most EQ cutting masters of master tapes. The
ones marked "master" are the ones that they used to actually cut the LPs, the
ones that have been dynamically compromised. So that's how they got away
with it. They kept the levels above the surface noise of the record just by
reducing the dynamics. In order for me to sleep at night, I want to use the
original master tape, which has dynamics. So then what do I do? It's just
compromise."

In another part of the article he discusses how he tried to 'undo' EQ from tapes
that were mixed over older equipment.