Signal/Noise Ratio

[mulester7]

.....how high of a signal/noise ratio "number" is safe, and about what number would be REAL good?....those active equalizers I went to that guy's house over, were rated at S/N 108, and 110....someone relate those to a rating that would be "safe", please......

 

[WmAx]

.....how high of a signal/noise ratio "number" is safe, and about what number would be REAL good?....those active equalizers I went to that guy's house over, were rated at S/N 108, and 110....someone relate those to a rating that would be "safe", please......

This depends on the actual effective SNR that you are achieving with your set up. For example, you can have a pre-amp with a 110 dB SNR, but this is measured with a test signal of specific voltage magnitude. If the input signal from the music is of lesser magnitude, the effective SNR for music playback will not be as high. In fact, this must always be the case, as the music signal has to allow for headroom for the dynamics. The test signals on the otherhand, commonly are performed at full scale amplitude before clipping the device under test. Some equipment manufactures may test SNR at realistic playback conditions, but this is rare. This goes for every part of the electronics chain, btw. But there is good news: according to perceptual research, a 74dB effective SNR was sufficient for transparent human playback of speakers in controlled lab setting. A slightly higher SNR is required was required for headphones(as the headphone blocks some sound externally, allowing for a lower noise floor). Also, most music has far higher noise floor than electronics. An 80dB SNR in actual music recording would be phenomenal, for reference. It should also be noted that most domestic environments have a relatively high noise floor, therefor 74dB would probably be overkill in most of those cases.

-Chris

 

[mulester7]

This depends on the actual effective SNR that you are achieving with your set up. For example, you can have a pre-amp with a 110 dB SNR, but this is measured with a given voltage amplitude. If the input signal from the music is of lesser voltage, the actual SNR will not be as high. This goes for every part of the electronics chain, btw. But there is good news: according to perceptual research, a 74dB effective SNR was sufficient for transparent human playback of speakers in controlled lab setting. A slightly higher SNR is required was required for headphones(as the headphone blocks some sound externally, allowing for a lower noise floor). Also, most music has far higher noise floor than electronics. An 80dB SNR in actual music recording would be phenomenal, for reference. It should also be noted that most domestic environments have a relatively high noise floor, therefor 74dB would probably be overkill in most of those cases.

-Chris

.....interesting indeed, Chris, but isn't a chain only as strong as it's weakest link?....if one component of your system is rated barely acceptably clean by S/N ratio, does the sound quality going through the component suffer from the effects of possible carryover from the component barely making the 100 cut?....like a glow carryover from the cut or something?.....

 

[WmAx]

.....interesting indeed, Chris, but isn't a chain only as strong as it's weakest link?....if one component of your system is rated barely acceptably clean by S/N ratio, does the sound quality going through the component suffer from the effects of possible carryover from the component barely making the 100 cut?....like a glow carryover from the cut or something?.....

It is correct that the system can only be as good as the weakest link in this case. But this parameter is easily observable: audible noise during what is supposed to be silence. Press pause or stop on the CD player. If you don't hear anything, there is no problem for you, practically speaking. You can not judge some digital devices in this manner, however, as they may have a noise gate circuit, which mutes all output automatically when no signal is present. I am aware of only one example doing this, but there may be more. That example was a computer soundcard(Audigy 2) that used this as a trick to apparently get better static noise measurement performance.

-Chris

 

[mulester7]

It is correct that the system can only be as good as the weakest link in this case. But this parameter is easily observable: audible noise during what is supposed to be silence. Press pause or stop on the CD player. If you don't hear anything, there is no problem for you, practically speaking. You can not judge some digital devices in this manner, however, as they may have a noise gate circuit, which mutes all output automatically when no signal is present. I am aware of only one example doing this, but there may be more. That example was a computer soundcard(Audigy 2) that used this as a trick to apparently get better static noise measurement performance.

-Chris

.....ok....so, what is the S/N number considered "safe", that is assigned to a pushing component or a source?....isn't it 100?....can we really trust 100?......

 

[mulester7]

It is correct that the system can only be as good as the weakest link in this case. But this parameter is easily observable: audible noise during what is supposed to be silence. Press pause or stop on the CD player. If you don't hear anything, there is no problem for you, practically speaking. You can not judge some digital devices in this manner, however, as they may have a noise gate circuit, which mutes all output automatically when no signal is present. I am aware of only one example doing this, but there may be more. That example was a computer soundcard(Audigy 2) that used this as a trick to apparently get better static noise measurement performance.

-Chris

.....for second reading.....

 

[WmAx]

.....ok....so, what is the S/N number considered "safe", that is assigned to a pushing component or a source?....isn't it 100?....can we really trust 100?......

There is no way to know the actual SNR you are achieving in actual use without actually measuring the total system as it being used in real application. Because of the variables mentioned previously in this thread, there is no way I can suggest a safe SNR value as listed by the manufacturers[their SNR measurement method is not likely going to be the same as you are actually achieving in real use].

-Chris

 

[mulester7]

There is no way to know the actual SNR you are achieving in actual use without actually measuring the total system as it being used in real application. Because of the variables mentioned previously in this thread, there is no way I can suggest a safe SNR value as listed by the manufacturers[their SNR measurement method is not likely going to be the same as you are actually achieving in real use].

-Chris

.....WmAx, you're slippery as owl snot....I have always been told and heard from those who should have been correct through experience, that 100 is the supposed line for noise to be inaudible via the SNR....but, I seem to have always gotten the cleanest and best results when the number is on up around 108-112....anyhow, this thread was initiated only to possibly help those new in the hobby....over and out-tuh......

 

[WmAx]

.....WmAx, you're slippery as owl snot....I have always been told and heard from those who should have been correct through experience, that 100 is the supposed line for noise to be inaudible via the SNR....

I doubt those people can be exptected to even know the actual SNR they are achieving in application[vs. the claimed SNR by the manufacturer, which is usually measured under non-realsitic conditions].

A claimed manfacturer spec of 100+ db SNR, at least, leaves a margin for error in the manufacturer's measurement method, so that you will probably not have any issue(s).

-Chris

 

[mulester7]

A claimed manfacturer spec of 100+ db SNR, at least, leaves a margin for error in the manufacturer's measurement method, so that you will probably not have any issue(s).

-Chris

.....there it is.....