Syntrillium Backup Forums :: View topic

AndyH · Posts: 1425

A few questions arise from my first experiment with microphone recording. For this I recorded from my living room speakers, with a single microphone set up on front of my normal listening position, pointing towards what I perceive as the center of the sound stage.

This was not oriented towards the subject of my earlier post about comparing subtle differences in equipment by recording the output via microphone. I simply needed some source to record, so I put some Sara K on the CD player and listened to the same song over and over as I tried to get things working.

Various times I seen written here that CE's VU meters are dependent on the soundcard driver, that they can only report what they are told, whether it be right or wrong. Most of my recordings are of LPs, directly off the phono section. I've never had need to pay much attention to the VU meters because that input just happens to be right where it needs to be. With the microphone input, clipping was easily (in a manner of speaking) obtained, so I noticed some things I never looked for before.

First, when I adjusted the input level so that the highest peaks just reached 0 on CE's VU meter, with the clip indicator never coming on, the recording still had lots of clipping, which does not sound very good. Bring up the soundcard control panel, I frequently saw clipping indicated thereon (with no clipping on CE's meters). Although the soundcard meter display is much smaller than CE's, and thus not so easily fine tuned, it should not be a problem to use it, but I am curious. Is this discrepancy between the card's VU meters and CE's VU meters normal?

I have both an Audiophile 2496 and a recently obtained Mia in the same machine. The first music recording with the Mia was an LP, via its analogue input. The next day I recorded the same LP into the Audiophile to compare results. (This is obviously not about microphone recording, but somewhat relevant). At first I thought the recordings sounded significantly different, but I soon realized that the Audiophile recording was louder. The RMS average difference was only 1.5dB, according to CE, but my general impression was of a more extreme difference.

For most of my microphone tests, I used the S/PDIF output of the M-Audio DUO (the microphone preamp) running into the S/PDIF inputs of the Mia and the Audiophile. The connection was via a cable that is supposedly high quality and made specifically for this purpose (i.e. not simply an analogue audio interconnect cable). Recordings were at 32/88.2. The DUO is switched to standalone mode for S/PDIF output, and the sample rate is set by hardware switch.

I set the input so that CE's VU meters were running just up to 0dB on the highest peaks, no clipping indicated, recording via the Mia. This produced a recording that actually peaked at -6dB (CE's statistics), but that had many clipped samples. The clipping was observable in CE's Edit view by virtue of the fact that many peaks reached this -6dB level, and in that these peaks looked sort-of squared off. In Spectral view these peaks showed many click-like spikes that went all the way to the 48kHz Nyquist limit.

When I switched the input cable to the Audiophile, but changed no external settings, CE's meters now showed frequent clipping (which they had not with the Mia). The resultant recording had a much higher general level than the Mia's recording, and showed much clipping at 0dB in CE's edit view. This was probably due to different settings in the soundcard's control panels, but the difference in the LP recordings was not due to the same thing.

The above stuff is mostly just curiosity. I realize that my two to three hours of experimentation, the first time I ever used a microphone, cannot begin to compare with the experience necessary to make good recordings. It is just my way of getting started. There is one thing more serious, however.

In order to get a useable recording level, I had the DUO's level control(s) set almost at maximum. This preamp is supposed to provide 60dB of gain. The gain is not calibrated, but there are division markings behind the knob. Based on 1/2 circle being divided into six parts, with one more division beyond 180 degrees in both directions, the scale extends through 240 degrees. The knob actually moves a little beyond both ends of the scale.

CE doesn't register anything with the gain set below the last division, and even then the signal level is very low unless the knob is almost at the end of its travel. In the test I described above, the Mia did not indicate clipping in CE (but did on its own meters) and the Audiophile showed frequent clipping on CE's meters. Using the Audiophile, I backed off the preamp's gain control so that the Audiophile meters stopped clipping, which also caused CE's meters to fall significantly.

Adjusting as carefully as possible to avoid going too far down, I actually barely moved the gain knob at all. It was in fact so little that I could not be certain, from touch, than any movement occurred, but the VU meters said it had. On the Audiophile's meters, the peaks fell from frequent clipping to mostly below -6dB. The resultant recording had one peak at -3.5dB, one at -4.5dB, and everything else much lower. Surely this is not normal behavior for a microphone preamp?

I also did a little recording via USB. This only worked in CE at 44.1 (I thought I had recorded at 48kHz some time ago, but I could not now get any signal in CE except at 44.1). Gain control on the DUO had to be near maximum with USB, just as for S/PDIF, although I did not try to make any measured comparisons.

To answer a few potential questions. The Hi/Fi was set to my normal listening level. This may be lower than in a studio recording, but Sara K. sounds about right to me, as far as I imagine how loud any half-way normal person might sing. The DUO has one green LED for signal, and one red LED for clipping. The green LED flashes occasionally. Within those few degrees of gain control that are actually useable, I can turn it up high enough that the red LED flashes on the loudest peaks.

During the 6+ minute recording that produce the -3.5dB maximum, I went down the hall and into the back room to make a phone call. At the end of the song, as it faded out, my normal level phone speaking was recorded along with the Hi/Fi speaker output (although at a lower level) -- that from behind the microphone.

The Mia's control panel was set for +4 for these experiments. I did try -10 at first, but I think that was before I realized I needed to turn the preamp gain up all the way, so I may not have actually record anything at that setting. I never though of it after finally getting a signal to register at the computer and beginning recording. For the LP recording, the Mia was set at -10. The Audiophile input has no level options; it should be more or less equivalent to the -10dB level, I think. I can see the Mia's setting making a difference for the microphone recordings, but not for the LP recording.

The DUO has switches for +4 and -10, both input and output independently. Those apparently do not effect the S/PDIF output, although switching while recording does produce clicks. The manual says they effect the line input and line output.

There are -20 dB pads on the DUO (one per channel). Pressing those switches does reduce the input level, but I did not attempt to measure how much. Likewise, the microphone itself has a -10dB pad switch. That also works; it reduces the level. Except for brief tests to see the effect, I did all recording without the pad reductions.

Most of the recordings were done via microphone channel 1, but I did try channel 2. It seems to be identical to 1 as far as needing near maximum gain to produce a useable signal.

SteveG · Location: United Kingdom Posts: 6695

According to the Duo spec, you can set the global output to be either -10dBv or +4dBu. Obviously, you need the Mia set to the same setting as the Duo. Is this what you did?

SteveG · Location: United Kingdom Posts: 6695

Perhaps rather more to the point, did you ever get an answer to the problem you were having this time last year with the Duo's S/PDIF output? I just re-read 'digital to digital difficulty'...

AndyH · Posts: 1425

some reading and some thoughts last night after posting this mess.

The +4/-10 switch on the Mia only applies to the analogue I/O. It should have no effect on the digital, therefore it should not be involved in my getting S/PDIF signal into CE that differs in level depending on whether I input via the Mia or via the Audiophile. However, I do get different levels with the two cards, but the +4/-10 setting should not be any part of the reason. Likewise, the +4/-10 settings on the DUO does not apply to the microphone input or to the S/PDIF output.

My original expectation was that the level into CE would not vary because of diferences in the gain setting of the DUO. Only a digital signal is being sent from the DUO to the computer via S/PDIF -- a string of numbers. However, that is not what I experienced. This makes me more confused than ever I (realized I) was before about just what goes on with S/PDIF.

I believe the answer to SteveG's question about the record/playback speed posting from last year is 'Yes, that did get straightened out.' The straightening out was somewhat mysterious. I seem to remember something about the sample rate switches on the DUO not working correctly until I got a smaller screwdriver to toggle them. They are very tiny and apparently were not getting moved completly into the different positions with my original blunt instrument.

Anyway, getting the correct sample rate for record and playback was not a problem yesterday. I've not really had occassion to use the DUO's S/PDIF output before except to run some tests. My memory, which is unfortunately never completly reliable, says I believed there was no problem about different sample rates working correctly.

The major problem is that the microphone preamp gain needs to be turned up essentially all the way to get a useable signal out the S/PDIF. At and near that maximum setting there is some variation, but control is quite hair-trigger, thus difficult to manage.

I expected the gain control to act more like a volume control on analogue HI/Fi equipment. I expected the "correct" setting to be somewhere around, or before, the middle of the range, and I also expected to be able to turn the dial a few degrees or more from the ideal position with only minor variations in results. I also expected the gain control to effect only the signal level delivered to the DUO's DAC, not the signal level deliveded out its S/PDIF jack. This latter means that the gain control should effect clipping by the DUO's DAC but it should not effect the signal level that is recorded in the computer.

I need to test using the line out jacks on the DUO. This bypasses its DAC. This analogue output will need to go into the analogue input of one of the PCI soundcards. Since I currently only have cables with 1/4 inch phone jacks on one end and RCA jacks on the other, I will only be able to input to the Audiophile. Both the DUO's output and the Mia's input are 1/4 inch phone. The output +4/-10 level switch on the DUO should come into play here, but not its input +4/-10 switch.

SteveG · Location: United Kingdom Posts: 6695

Obviously my comments about the levels were intended to relate only to the analog ins and outs. The Duo A-D converter should automatically be fed with the correct level.

AndyH · Posts: 1425

SteveG, your statements about level vs gain leaves me confused. Possibly the problem is that I don't now the definition of "level." I understand that a control's function could be to reduce the voltage of the input signal (the level?) while leaving the preamp's gain constant, or could be to vary the preamp's gain without effecting the voltage (level?) of the input signal. Don't both approaches vary the voltage (the level?) of the output signal (in the analogue domain, anyway)?

You say "If the Duo has a gain control rather than a level control, then it will always be possible to get a full output from the preamp whatever position it is in" Does the "possible" mean that it is also possible to get less than full output? If so, what does one normally do to achieve that possibility of full output? Possibly your "full output" refers to something other than the signal voltage at the DUO's line out jacks, or possibly I am more confused than I ever believed about how things work.

The DUO's manual makes various references to gain, as in "This potentiometer controls the amount of gain applied to the microphone preamp" so I guess it isn't a level control. Maximum gain is specified as 60dB. Is this normal for a microphone preamp?

However the ins and outs of analogue signals work, I believed that the output of the S/PDIF and the USB were simply digital signals such as those that run around inside this computer. Their voltage level would thus be constant, or rather an as-instantaneous-as-possible alteration (i.e. square wave) between zero and some constant such as 5V, 12V, .... whatever the standard says it should be. If this were so, however, I don't see how the VU meters in CE or the soundcard control panels could ever report clipping (as they were). Clipping could only occur in the analogue domain, before the DUO's DAC, no?

REGARDLESS, today I placed the microphone about six inches from one speaker, center of microphone vertically about half way between midrange and tweeter, pointing directly at the speaker. HiFi volume level was again at my normal listening level.

This produced a signal that was much easier to handle. The preamp gain control had to be more than 3/4 of the way up to get a useful signal to the computer. This is still more than I would have thought reasonable, but much better than yesterday. It was also easier to control the signal level (that word again). I did not need to move the gain potentiometer in thousandths of a degree to maintain control, as I had yesterday.

Perhaps this microphone, an AT3035, is not particularly sensitive and is intended to always be placed right next to the sound source, or perhaps it produces only a small signal and this preamp is not the most suitable one for it. The specs say open circuit sensitivity -32 dB (25.1mV) re 1V at 1 Pascal but that tells me nothing about what to expect. It claims to have a 136 dB dynamic range and to handle 148 dB SPL maximum input at 1kHz, which I should think means that it picks up fairly low level sounds without difficulty.

With this arrangement I was able to record Sara K at reasonable CE levels via the preamp line out to the Audiophile line in (no cables that will work with the Mia), via the USB, and via the S/PDIF to both the Audiophile and the Mia.

SteveG · Location: United Kingdom Posts: 6695

It's really quite simple - a 'level' control is a volume control. You have a signal, and you can either have all of it by turning the control up to the top, or none of it by turning it right down to the bottom.

A 'gain' control doesn't do this at all. It controls only the amount of amplification that a preamp stage has. So it alters the gain from a maximum of 60dB down to a minimum of probably about 10dB, or thereabouts. In other words, if you put a large signal in, which peaks at the -10dB level, then you will still get the full output from the preamp - because it is correctly set to give the correct amount of gain for this input. To get no output, you would have to have a gain control that reduced the gain to minus infinity, and you can't do that easily - but you don't need to, because that's what a level control is for!

So in a mixer, the sliders are level controls, and the knob at the top of each channel is a mic gain control. If you like, it 'normalises' the channel for the level of signal coming from the microphone. And on a really loud signal, you would set it somewhere near its minimum setting.

AndyH · Posts: 1425

Thanks. I guess the summary for me is that the equipment is probably working correctly. The fault was in thinking I could put the microphone in my normal listening position and pick up a decent signal with the HiFi running at normal listening levels. Actually, I don't have a reference point to give me any insight about how the math quoted above works out in this situation.

The responses to my earlier post about recording my room contained many objections to my idea, but having a problem getting a strong enough signal to record properly wasn't one of them. Thus I had no preconception that there might be any difficulty making a recording, only that there might be issues using the recording in the way I was considering.

SteveG · Location: United Kingdom Posts: 6695

Levels aren't really the issue with that project - all of the original objections stand. All that's really happened here is that you've discovered the nuts and bolts of 'gain staging' - the art of making sure that the correct signal levels are present at each stage in the process of recording.

Where levels do become an issue is with the self-noise figure of the mic - and this is what will limit recordings of really quiet sources. Unfortunately condensor mics don't score as highly in this area as most other sorts - this is all to do with capsule impedance and the thermal noise generated by it. Ribbon mics do best, followed by dynamics, with condensors of all sorts coming a poor third, but I think that piezoelectric ones are actually the worst. It's the only bit of the mic equation that condensors lose out on, though, and can be compensated for to a large degree by sensible mic placement.

AndyH · Posts: 1425

"noise" is listed as 12 dB SPL, but you probably already knew that, more or less. I again have no references, but as this is not a very expensive microphone, as they go, its figures are probably nothing to write home about, or to the forum about, anyway.

While no expert, I think I understand the basics of gain staging. However, I don't see that is quite the issue here. There is one and only one stage of gain accessable. The issue really is that this microphone doesn't put out enough signal from "ordinary" sound levels to work well with as much (or as little) gain as the preamp has.

I guess one might say gain staging as in I need another stage of gain ... or a different microphone designed to be used further from the sound source than this one. That is, if I really needed to record in that particular manner. I was actually just doing it that way because I did not know any better, never having tried before.

SteveG · Location: United Kingdom Posts: 6695

AndyH · Posts: 1425

SteveG · Location: United Kingdom Posts: 6695

AndyH · Posts: 1425

Yes, that is my whole point -- that I should have a higher signal level, that is. Or at least that my expectation was to be easily able to record from the back of the room without maxing out the preamp gain. Since that expectation was actually based only upon some impressions formed by reading other people's comments, I am still up in the air. Perhaps I could haul things down to the music shop and ask them to run a few test. That ought to thrill someone.

SteveG · Location: United Kingdom Posts: 6695

VoodooRadio · Location: USA Posts: 3971

SteveG · Location: United Kingdom Posts: 6695

AndyH · Posts: 1425

I also am less than thrilled by the prospect. They are friendly enough folk, but it always seems to be a mad house in there. Not beng a musician, I do not give them very much business; while they have all sorts of fascinating looking things, they are just not things that fit into my life. I suspose any largish store is bound to have a few goofy customers that repeatedly ask strange questions. The last few days, and quite possible the next few, as just too filled with other activities for me to find time for a microphone before 3:00 AM, but I guess I better do something sooner rather than later.

VoodooRadio · Location: USA Posts: 3971

I must profess.... several years back, I played in a band with a guy who (once they opened the first GC in Houston) went to work there. I established a good relationship with them and over the years, he has gone from salesman, Pro Audio Dept Manager, Asst Store Manager and now has relocated to work in Regional Sales. While they don't make any "special" deals for me, they usually get about at low as they can go and are..... very accommodating! Approve