Topic: Noise floor question

[MikeMannZ]

In the m10 thread there's a good portion dedicated to discussing the noise floor of the unit compared to other units.  For the sake of arguments let's take these models. 

Sony PCM-M10
Edirol R-09HR
Tascam DR-1 and the 07
Olympus LS-11
Zoom H2 and H4N

All in the under 300 buck range.  Are there really noticeable difference when recording live music with the internals?  Now I've already read the H2 is awful with mics plugged in, but what about the other units and if one is just recording live music are there any major differences between the lot of em (noise floor wise)?  I understand the M10 is going to produce a more mono mix, but that aside and just taking noise floor into consideration for live music (as opposed to sounds of nature) are there huge decernible differences?

Thanks
Mike

[chrise]

What sort of music ?

[MikeMannZ]

What sort of music ?

Does it matter?  If you're implying that if it were to be say classical or jazz (standards) there would be lower volumes and space where the floor could be heard then not so much.  Even still are some in that list unacceptable for such an application?  On the other side of the coin how about funk, rock and the blues how would they fair against each other in that type of situation?

[chrise]

The self-noise on the most of these units will often be irrelevant recording music, unless very quiet passages are involved e.g classical music with quiet passages.

The internal mics are handy for making "fun" recordings.  For more serious recording, you may need to  use external mics, and it would make sense to check out the mic input noise level tests:

  http://www.avisoft.com/recordertests.htm

[chrise]

Is it the noise of the external mics indicated or the M-10 internal pre-amp causing this hiss?

In theory, the PCM-M10 mic input pre-amp is among the quietest of the affordable recorders.  See the measurements here...

 http://www.avisoft.com/recordertests.htm

[MikeMannZ]

In theory, the PCM-M10 mic input pre-amp is among the quietest of the affordable recorders.  See the measurements here...
http://www.avisoft.com/recordertests.htm

But as you can hear from pools test http://taperssection.com/index.php?topic=132752.0 there is quite a bit of hiss from both the line and mic inputs.  If one were recording classical I would find the hiss unacceptable and annoying.

[spzkt]

You'll have to excuse my live taping ignorance, but I've banged my head against the noise floor thing for quite a while in relation to nature sound recording so hopeful I can shed a bit of light!

It's really important to keep in mind the preamp figures are a bit useless when considered in isolation from the noise output level of the mics you are using. It's only when you look at the sensitivity and self noise of the mic in relation to the mic preamps noise that you can understand what is contributing to the level of hiss/noise that you are hearing.

I can't find the specs for OKM that Pool was using in other thread but the Soundman site does have specs for the OKM II Klassik and Rock.  The signal to noise is usually referenced to 94dBa to give a self noise figure - in this case the Klassic has a self noise of 33dBa (94dBa - 61dBa) and a sensitivity of 300mv/Pa.   The OKM II Rock have self noise of 53dBa  (94dBa - 41dBa) and sensitivity of 30mv/Pa.

If you could punch the sensitivity and self-noise into chart 3 given at http://www.rane.com/note148.html you'll get the output noise of the mic which can be directly related to the preamp EIN.  Unfortunately the OKM II mics are off the scale due to their high levels of self noise.  Punching the figures into a spreadsheet I use to calculate output noise the Klassic is around -68dBu. The OKM II Rock has the same noise floor as Klassic.

Now if we bring the M10 back into the picture we can get a good idea of whether the mic pres are the main contributor to the hiss in Pool's recordings.  Using the mic input noise figures given at Avisoft we can see that the mic pre has a self-noise of -122dBu A-weighted, while the OKM II's have a noise output of -68dBu. This means that the noise output of the OKM II would be 54dBu higher than noise floor of the M10 preamp. The recommendation made by Rane (and elsewhere) is that choosing a mic with a noise output about 7dBa-10dBa higher than the preamp EIN will ensure the preamp doesn't degrade the noise floor of the mic. So in the case of the OKM II's you would be hearing zero contribution from the preamp or line ins - 100% of the hiss will be coming from the mics.

Putting it another way you'd need a mic with a noise output of -112dBa or lower to be hearing the influence of the M10's preamp noise.   Of the mics listed below only the Schoeps CMC5-MK41 would be degraded by the M10 preamps. On the other hand the noise output of the OKM's would swamp the preamp noise of the M10.

Mic	Self-noise	Sensitivity	Output Noise
Senn. MKH40	12dB(A)	25mv/Pa	-111.8dBu
Schoeps CMC5-MK41	15dB(A)	13mv/Pa	-114.5dBu
Neumann K140	16dB(A)	15mv/Pa	-112.2dBu
DPA 4060	23dB(A)	20mv/Pa	-102dBu
OKM II Rock	54dB(A)	30mv/Pa	-68dBu

hope this helps.

cheers
Paul

[marvin100]

Great comment.

But this is confusing. I think the thread should be merged with http://taperssection.com/index.php?topic=132752.0

Mods?

[Ozpeter]

The discernable differences are going to be more down to the pickup pattern of the mics and their frequency response, I suspect.

In the M10 noise thread I'm about to post a sample which to me shows that the M10 internal mics appear to be very low noise, but they also appear to be rolled off at the high frequency end, with some consequent rolling off of the noise too.  Kind of cheating!

It's almost impossible to compare one against the other apart from simply saying which you prefer - there's too many factors in play to say which is "best", and for me the stereo image characteristic is at least as important as frequency response and noise level.  And in that imaging respect, again, there are bound to be wide variations.

[MikeMannZ]

In the M10 noise thread I'm about to post a sample which to me shows that the M10 internal mics appear to be very low noise, but they also appear to be rolled off at the high frequency end, with some consequent rolling off of the noise too.  Kind of cheating!

It's almost impossible to compare one against the other apart from simply saying which you prefer - there's too many factors in play to say which is "best", and for me the stereo image characteristic is at least as important as frequency response and noise level.  And in that imaging respect, again, there are bound to be wide variations.

So it's interesting that you say this, because when I first heard the noise floor samples of the m10s internal mics  at wingfield I noticed exactly that...Highs cut, and warmth of the voice.  When I went from the top to the bottom of the samples provided for each recorder the internals of the m10 seemed to be the best next to the d50.  Other units which I was considering (like the ls11) seemed to tinny or high end-E.  One thing Stephanie said to me in an email was she thought the ls11 was giving a more accurate picture of what was being recorded and that to her the m10 seemed muddy. 

Mike     

[pool]

In theory, the PCM-M10 mic input pre-amp is among the quietest of the affordable recorders.  See the measurements here...
http://www.avisoft.com/recordertests.htm

But as you can hear from pools test http://taperssection.com/index.php?topic=132752.0 there is quite a bit of hiss from both the line and mic inputs.  If one were recording classical I would find the hiss unacceptable and annoying.

there is noise in the mic inputs. In the line its practically nil. ill post samles of a ticking clock of internal, om mic in and line in. line in very very satisfactory if using the ac3 (which acts like a small premap) allows recording of low performances. however im willing to try dpa 4060's.

[DSatz]

Admittedly I don't have the patience to read everything that has been posted in this thread and/or the other thread referred to above. But I'm aware of a great deal of confusion on this forum about mike and line inputs, and I want to make sure people understand a couple of basic points.

One point is that you can't tell ANYTHING by leaving a mike or line input unconnected and listening for noise. If anyone is worried about the fact that they hear hiss (perhaps even a lot of hiss) that way from a given preamp or recorder, please stop; that's no proof of anything. An unconnected input is being "driven" in effect to an extremely low voltage by an extremely high impedance, and high impedance signal sources (even virtual ones as in this case) are full of noise. A microphone has (or should have) a low output impedance which causes the input circuit to behave very differently.

The other point is that you should reasonably expect a mike input to SEEM to have more hiss than a line input simply because the function of the mike input is to amplify any incoming signals much more than the line input does. If you use the line input with the same signal source, you'll have to turn up the recording levels considerably--and that will only increase the noise. So again this situation seems to lead some people to highly confident but totally wrong conclusions.

To determine whether a line or mike input is noisy or quiet or in-between, you need two things. You need to know what levels (in terms of voltage) are typical for the output of your microphones when recording whatever you record; you also need a audio signal generator with low output impedance and adjustable level. (Note: If you're already saying to yourself that you can't make tests like this, fine--but that means that you truly don't know which inputs are how noisy. and anything you think or believe on the subject is only a guess.)

If the typical output levels of your microphones when recording loud music are, say, 40 mV then set the signal generator to put out 40 mV at some non-challenging frequency such as 400 Hz. Then feed that signal into the mike input of whatever preamp, mixer or recorder you're testing. Set the record levels (if it's a recorder) so that you get close to 0 dB with the signal coming in. Measure the signal coming out of the recorder at its headphone or line outputs, and/or set up a monitoring arrangement with headphones or loudspeakers so that the 400 Hz tone is as loud as the loudest sound you would normally expect to hear.

Now turn the signal generator off but leave it connected (remember, open inputs don't count), and measure and/or listen again. That's the noise floor to be concerned about. Repeat the experiment with the generator connected to the line input, and I'd be surprised if you can even turn things up far enough to get the meters to read near 0 dB--but even if you can do so, there's a very good chance that you'll find the line input to offer MORE noise rather than less, because of how far you have to crank the gain up.

--best regards

[ghibliss]

Since most mic inputs on these recorders are the weakest link in the chain why not mention this?  I am using an external A/D converter to bring the mic signal into my M-Audio 24/96 specifically to bypass this weak link.  The differences are huge when doing this.  This is the main reason why I selected the M-Audio 24/96 for its spdif input which most other machines lack.  The other is that it provides 24 bit 96 KHz recording which I do not find offered on any other recorder which is of comparable size. 

Recording concerts for most of us is something which requires a stealthy sized machine which most machines do not fill the bill on this note.  As much as i would like a Sound Devices 722 they are hardly stealthy and do not offer anything that the M-Audio unit can't do with a good external A/D converter which I use, aside from having a large hard drive.  I use a 16 GB compact flash card which is large enough capacity typically for 4-5 shows.  I have an external lithium battery rated at 7200 mah which is far greater then the recorder requires for this length of time as well.

I use DPA 4061 microphones which are designed for high sound pressure environments such as rock concerts and other loud amplified music.  This is why they are intentionally less sensitive then the microphones listed in one of the earlier responses.  The high sound pressure level of the concert allows for having the gain of the A/D converter very low which contributes almost no self noise.  By comparison using a microphone connected directly to the mic inputs of the recorder would have significantly higher noise on the recording from the noisy preamplifier circuitry.

Twenty Four bit recording also allows one to use substantially lower gain then sixteen bit so you are able to record at much lower levels and reduce the self noise applied to the recording.  Increasing the gain +6 db in post will not add  noise to the recording if you have the inclination to to increase this after the fact. 

[Ozpeter]

Increasing the gain +6 db in post will not add  noise to the recording if you have the inclination to to increase this after the fact. 

It will increase the noise by the same 6dB.  However, the contribution from the AD > DA will be of such low level that you are unlikely to notice it.  The noise from the analog stages feeding the original recording will of course increase by 6dB too, and that you might notice.  There's no free lunch.

[aaronji]

The other is that it provides 24 bit 96 KHz recording which I do not find offered on any other recorder which is of comparable size. 

Recording concerts for most of us is something which requires a stealthy sized machine which most machines do not fill the bill on this note.

I am guessing you haven't been shopping for a recorder recently?  I am pretty sure that all of the ones mentioned in this thread have 24/96, except for the Tascam.  Most are "stealthy sized" too (basically the same size as the MT, give or take a little)...