Skip to main content

help needed: Recording without hiss - how? + balance issue

I've been recording live recitals and performances for a while... I've asked a similar question to this before, and eventually fixed it. However, I am still bothered by this:

I eventually figured out that the reason I was getting so much hiss in my recordings was because of crappy equipment. Once I updated my equipment, the recordings come out better, but the background hiss is still there.

I have a simple setup: Laptop, Adobe Audition 2, PreSonus Firepod, 2 Samson C02 pencil condenser mics in X/Y, 2 identical cables. I set the levels, record, take it home, start editing, and just can't get over the ambient background hiss that's present. I hate using the noise-reduction and hiss-reduction plug-ins because they make the recording sound muddy and just plain bad. So I need to find a way to make the hiss as minimal as possible before recording. Somebody here must be able to point out what I'm doing wrong, or help me improve. The other problem is that I'm a perfectionist, so I have to have my recordings sound very good! Or maybe the hiss is just normal and I'm over-reacting? If there is equipment out that would help I would be willing to get it, I just need to know what it is first. Someone clarify this for me please.

Next issue:
Often, when I record these recitals, the applause are way louder than the performer. When you look at the wave-form you can see these huge spikes... when you click and listen to the "spikes" they are always the applause. This can be annoying, because when you want to normalize the track it normalizes the highest point of the piece... so it will hardly normalize because the spikes will hit the set % (Ie: normalize to 97%) before the rest. I usually reduce the volume of the applause manually, and then normalize and it usually turns out OK. If there are any suggestions on how to prevent this from occurring I would love to know. I'm trying to learn this stuff as best as I can. I know there must be something I'm missing, or doing wrong, or just not doing at all... Thank you.


sheet Sun, 06/03/2007 - 20:00
Everything in your system makes noise: Especially your mics, preamps and AC. If you can locate a portable balanced power conditioner, you will drop your noise level 10dB on average. Once you upgrade your mics and preamps, you will be better off.

You are always going to have some ambient room noise and noise from the air. When you mix, you can automate your high pass and EQ to compensate as needed. Your setttings should never be static.

zemlin Mon, 06/04/2007 - 03:17
What sample-rate/bit depth do you record? What format to do you save to when done? Saw a lengthy thread on this sort of thing once before and it turned out the fella' was saving 8-bit files! A very low record level on 16 bit files could also induce hiss when punched up to normal levels.

The NR tools in AA are quite good, but need to be used with a lot of care. There are a lot of setting you can adjust in the NR dialog, and it takes time to learn them. The main thing to remember is it's NOISE REDUCTION, not NOISE ELIMINATION, but eliminating/reducing the source of the noise is ALWAYS the best option.

Normalize (at least in Audition) does not ADD noise - it's simple amplification. In AA, a good way to balance the the applause level is to split the clips when the applause comes up and drag them to a different track. Then do a short crossfade between the clips on the two clips and lower the level on the applause track down to where you want it. You'll need to tweek the cross-fade timing and speed, but 97.8% of the time you can fade into quieter applause transparently.

HVAC systems even in some auditoriums make a surprising amount of noise - and I will use careful noise reduction with AA to bring the noise level down. AC systems usually record a LOT of low-end noise that can be cut with EQ or an FFT filter (depending on the recorded material) which sometimes eliminates the need for additional NR.

Simmosonic Mon, 06/04/2007 - 04:02
Re: help needed: Recording without hiss - how?! + balance i

Regarding the hiss: the first thing you need to do is determine whether a) it is actually noisier than what is generally considered 'acceptable', or b) you are becoming obsessive about it...

Regarding the applause: assuming you have positioned the microphones to get the best sound from the performers, the audience noise will be whatever the mics pick up from that position. Moving the mics to reduce audience noise would probably be detrimental to the sound you're trying to capture. Putting the cart before the horse, in fact.

The only way to 'fix' the problem at the source is to choose a microphone technique that uses very directional microphones and/or allows you to get very close to the performers, so that it increases the ratio of music to audience signals. But that can lead to a lot of futile tail-chasing.

I record a lot of live concerts, particularly small ensembles, and it is normal procedure for me to reduce the level of applause during mastering so that I can get the level of the music up for CD playback.

In fact, I recorded a string octet recital on Saturday night using a pair of Schoeps bidirectionals in MS Blumlein. Despite a relatively close microphone position, the applause is loud and clear and considerably higher than the average recorded level for the music. (There's no way that eight bowed strings on a stage can compete with 1200 or so people clapping and cheering in a concert hall with healthy reverberation!)


With careful adjustment of the volume envelope during mastering, I was able to pull the applause down by 9dB in a way that no-one will notice (unless, of course, they are specifically listening for it). Now the highest peaks in the recording come from the music, not the applause.

I was helped in this endeavour by the fact that these particular artists know how to 'hold' an audience at the end of a piece, keeping them quiet until all the reverberation in the room dies down (and thereby giving me plenty of time to drop the volume envelope with subtlety). On one piece there is almost 10 seconds of 'silence' between the perceptible end of the last note and the first clap from the audience - nice!

The ability to 'hold the audience' is something you may want to discuss with the conductor or musicians you are recording - it can really help to have a bit of a gap between the end of the piece and the start of the applause.

Also, on the topic of filters and following from Karl's post:

After editing, I like to run the material through a spectrum analyser to determine where the LF rumble of the room is (from air conditioning, etc.). By switching between music and the 'silence' between pieces, it becomes easy to see where the music rolls off, and so I set a HPF to that point and remove everything below it. On the string octet recording mentioned above, I used a very steep linear phase HPF set to 63Hz, although I probably could've gone higher than that. Nonetheless, it cleaned things up nicely.

BobRogers Mon, 06/04/2007 - 04:58
As Sheet said, everything has a noise floor, but in your case we have only a few candidates.
1. The room
2. The mics
3. The cables
4. The firepod

Here is a test you can do with no new equipment. If you turn the levels up on the firepod (even to the point of clipping during the applause) does the hiss increase or decrease? If it increases, the hiss is before the pres (room, mic, cables). If it decreases, you are hearing the noise floor of the firepod.

My guess is that the noise is before the preamps. I would put any investment into the basic signal chain - not into other equipment. Assuming you have reasonably good quality cables, I'd upgrade the mics before the pres.

DavidSpearritt Mon, 06/04/2007 - 05:08
Re: help needed: Recording without hiss - how?! + balance i

Simmosonic wrote: With careful adjustment of the volume envelope during mastering, I was able to pull the applause down by 9dB in a way that no-one will notice (unless, of course, they are specifically listening for it).

Yep, I do this routinely, as I am sure others do, with a little custom fade out profile in Wavelab as shown here.

Pro Audio Guest Mon, 06/04/2007 - 05:25
Most of the essential things have already been said. I just wanted to point out that I feel that your mics will rather be the weak link in the chain. The Samsons have a noise figure of 22-23dBA. This is 6-10dB higher than most high quality SDC mics. Taking into account that companies like Schoeps, DPA, Sennheiser or Neumann tend to quote worst case figures shows that there will be great differences in mic noise.
Best regards

Cucco Mon, 06/04/2007 - 06:41
Where others are pointing to your equipment here, I'd like to chime in and say that, even given your budget gear, you should not be getting excessively high noise (hiss).

It's FAR more likely that it is your room or a problem with gain staging.

1 - The room - I get to record in a VARIETY of spaces, from glorious concert halls (Kennedy Center, Strathmore, Meyerhoff) to High School Auditoriums, to cafeterias, gymnasiums and warehouses! I've found that the most consistent offender to the noise issue is.......
The high school auditorium. For some horrible reason, people (school systems) invest over a million dollars in a beautiful *looking* concert hall, but throw the loudest damn AC known to man in the hall to cool it down. What's worse (at least in this area) is that the AC system is controlled by computers located 50 miles away and can't simply be "turned off" with a switch.

I can say with some comfort that in a high-school auditorium, I'm usually quite pleased with noise floor figures that are around -50dBFS (with no filtering applied). However, I'm often confornted with noise floor figures closer to -40dBFS or higher. (Figured by "manually normalizing" or finding the loudest point in the music and ensuring that it peaks at around -1 dBFS or higher.)

Usually, I can put a high pass filter on and affect this figure by as much as 10dB or more. However, I won't (under most circumstances) place a global HPF on. Instead, I'll go through and put it on where there is absolutely no LF content (flute solos with violins, etc.) and then fade it out (back to all-pass) during louder/fuller sections. This can be time consuming, but hey, it's what I get paid to do.

A good dynamic eq (or careful use of a multi-band compressor acting as a frequency conscious expander in the low frequencies) can also help with this, but setting it correctly can actually be even more time consuming than manually applying HPF's where appropriate.

2 - Gain Staging -

This is an important aspect to recording. Are you getting hot enough levels going in? Are you getting too hot of levels?

Are you clipping at any one of the inputs during any part of the concerts? If yes, then your gain is too high. (I know...."DUH.")

Do you have 20 dB of headroom at all times on any of the channels? If yes, your gain is too low.

I realize that with 24 bit recording, you have a little flexibility with extra head room, but why not use all of your available bits? Get your input gain to a nice healthy range. This will keep you from having to raise your gain once inside your DAW and thus raising the digital noise floor along with your analog noise.


As for noise reduction....I can't really bring myself to use it - ever! No matter how little it's used, I can hear it. Even when used quite sparingly, I find it creates an "off-ness" to the sound which hurts my ears after only a few seconds of listening.

In addition, it's one of the most abused tools by amatuer acoustic recordists. The only way it should be used is the way that Karl mentions - very sparingly, cautiously and done little bits at a time with the intent of reducing the noise, not eliminating it.

My advice - avoid it if you can, use it only if you must.

Never try applying it "globally" - work with it one track at a time. It's very time consuming to do correctly.



I'm with David and Simmo - a gentle adjustment is more than enough to level the two out.


Normalizing -

A tool which I never use. When normalizing, I do this manually (as mentioned above.)

That being said, I don't "normalize" each track. The entire event should be treated as one "song." If string quartet is normalized to the same average amplitude as an orchestra, it's gonna sound odd.

I do disagree with one statement made....

Someone above said "Settings should never be static..."

I whole heartedly disagree. Within one piece of music on a band/orchestra live recording, settings should be as static as possible. Fader movements, momentary EQ settings, etc will be audible through a wind/string ensemble piece. Notable exceptions - gentle HPF ins-and-outs can be acceptable as can gentle (long) gain changes applied globally (such as adding a long crescendo that the ensemble didn't quite pull off).

I won't even adjust spot mics during an individual track as nothing occurs within a vacuum. Think of an orchestra with spot mics on the clarinet- you're going to pick up some viola (or cello depending upon seating style) and second violin. If you pull up and down the clarinet spot mic at various times, you're now directly affecting the balance between the mid strings and the upper-and-lower strings. Not good. It can be pretty hard to find one "global setting" that works for the entire piece, but IMO, it's our jobs to do so.

Sorry for the lengthy post.....

Cheers -


BobRogers Mon, 06/04/2007 - 06:47
I meant to mention this previously, and it has been touched on before, but if the signal to noise problem is coming out of the mics, then the first thing to try is to move the mics closer to the sound source. Signal from the source is increased; noise in unaffected.

On a more general note, when working with budget equipment I'd worry about things like frequency response, clarity, image, and dynamic range before signal to noise ratio. And I'd accept a little noise to improve any one of these. You have to make compromises with budget equipment and I'd rather pay a penalty in noise than in any other factor if I have a choice.
UPDATE: Cross posted with Jeremy. This is an area where our posts disagree. He is saying (in his gain structure comments) to get the maximum gain without clipping out of your pres. I'm saying to give them a lot of headroom. (Maybe not 20dB, but a lot.) To answer his (maybe rhetorical) question - the reason not to use the last few bits is that that last few dB of gain in a cheap pre sounds much worse than its nice linear middle gain stage. Since you aren't clipping on the applause, you probably have sufficient headroom on the music.

Pro Audio Guest Mon, 06/04/2007 - 07:02
Thanks for all the responses. This should be helpful - I will try them out.

Also, I have the ability to record at a sample rate of 96,000... however, I can't figure out how to make Audition 2 match the sample rate of my firepod! Does anyone know if it is possible to do so? I also have Cubase SE, but I haven't used it much... but I believe cubase has the ability to record at higher sample rates.

JoeH Mon, 06/04/2007 - 07:34
Not much to add to the great advice you've already been given, Ardroth, except to counter what someone said earlier. DO NOT ride the faders during a performance. (You probably already knew that!) In classical/acoustic music, this is a big no-no. (Live PA is a completely different story, of course.)

As for recording at 96k; I doubt you'll gain much, at least with your present gear. You'll be eating up bandwidth (and computer performance) where it's not going to do you much good. Perhaps get better mics instead?

Set up a good gain stage with your gear, use a 24/44 bit depth/sample rate, and you should be good to go. Trust me on this, this is NOT where your hiss problems are happening.

I have to chuckle in agreement with Jeremy on the HVAC/High School problems. I too get to record in some of the best places in town, a few are downright spectacular, knock-yer-socks-off, breath-taking acoustics, where EVERYTHING works as it should. In these wonderful places, you really do literally work with just the artist, the space around them, and THEN you get to hear what your equipment is capable of doing.

Not so with about 60% of the rest of the "bad" spaces we record in. Even a lot of modern churches have HVAC systems that rumble and rattle on, all through performances. In a big urban city such as Philadelphia, we also have a "Steam-loop" - in which customers can "Buy" steam heat from PECO. In the wintertime, this is always "on," always clanging and banging away in some of the older churches where recitals are held. One of the church's caretakers refers to it as the "Dead guy in the basement, banging the pipes."

I've had a busy weekend just end (seven events!) and two of the recordings we did were "youth" choir recordings. One of them was a performance in a church, the other was a CD recording session. Interestingly, neither had AC available, although the humidity was high, and the temps reached 82 degrees both days. (The church had no AC at all, and the AC in the building where the choir did their recording had died! ) In the past, I have always had to work long and hard on removing the ever-present Air conditioning fan and hiss in these recordings with this group. This year, I'm going to have a MUCH better recording, even though we literally SWEATED through the session.

Just remember, for "live" recordings, some of the hiss and background noise is always going to be there, you just don't want TOO much. But getting rid of too much may be counterproductive. Many folks want to present "live" performances of their work, (esp to prove it's really THEM, with no edits or tricks!), and the concert-hall ambience (not to be mistaken for hiss) is an important part of the listening experience.

Cucco Mon, 06/04/2007 - 08:05
BobRogers wrote: UPDATE: Cross posted with Jeremy. This is an area where our posts disagree. He is saying (in his gain structure comments) to get the maximum gain without clipping out of your pres. I'm saying to give them a lot of headroom. (Maybe not 20dB, but a lot.) To answer his (maybe rhetorical) question - the reason not to use the last few bits is that that last few dB of gain in a cheap pre sounds much worse than its nice linear middle gain stage. Since you aren't clipping on the applause, you probably have sufficient headroom on the music.

Bob -

Go back and re-read my post. I didn't say to let the music peak but let the applause clip. I said:
"Are you clipping at any one of the inputs during any part of the concerts? If yes, then your gain is too high. (I know...."DUH.")

Do you have 20 dB of headroom at all times on any of the channels? If yes, your gain is too low. "

This doesn't imply that I'm saying "peak the music"

In fact, you're right about modest pres not being very linear in their performance. However, I could hardly imagine a performance (other than using ribbons on a harp recording) where the pres would have to be cranked! In fact, for most live concerts, to obtain the kind of gain staging I'm referring to, the pres would be well within their "good" middle range.



Boswell Mon, 06/04/2007 - 08:25
Firstly, the hiss problem.

The hiss is coming from the front end of the recording chain, i.e. microphones or FirePod pre-amp inputs. A power conditioner cannot produce hiss of this type.

Using the equipment you have, you have to strike a balance with the setting of the gain trims. Firstly, though, do an experiment to see which is the dominant hiss source. You will need an in-line fully-balanced XLR attenuator to do this. If you are not able to borrow/beg/buy one, you can make one easily.*

In as quiet a room as possible, record two sections of "silence", one without the attenuator and one with the attenuator inserted in the line between the microphone and the pre-amp. This can be a mono (single-channel) signal for this test. On playback, adjust the faders for the inverse of the attenuator setting, e.g. if you have a 10 dB attenuator, set the fader to 0dB for the sections recorded with the attenuator, and to -10dB for the un-attenuated sections, so you have the same overall gain from microphone to replay monitor. Note what happens to the hiss level. If it stays much the same, the hiss is predominantly from the microphones. If it varies in proportion fader setting, it's predominantly from the pre-amps. It may be between the two, which means both sources are contributing the the hiss. Once you have identified the microphones or the pre-amps in the FirePod as the hiss culprit, you are armed with the knowledge and can decide which to upgrade. I would personally tolerate some hiss rather than use noise-reduction techniques on that account, as they do take a toll on the audio quality.

* Making a fully-balanced 9.5dB attenuator yourself: Take three 150 Ohm 1% resistors and an XLR cable plug and cable socket. Wire the three resistors in series and take the outer ends to pins 2 and 3 of the XLR socket. Take the ends of the middle resistor to pins 2 and 3 of the plug. Wire pin 1 of the plug to pin 1 of the socket. Use the attenuator inline between the microphone and the FirePod input, either directly at the microphone or at the FirePod (it doesn't matter for this type of test).

Now the applause problem:
If you treat the applause as a necessary evil, in that it has to be there, then I let it clip at recording and reduce the level of the applause by 12dB or so in post-production. The clipping will usually be inaudible. The only difficulty is managing the fade between the end of the performance and the start of the applause, the dynamics of which will vary from piece to piece. It's something you can get quite skilled at with a little practice, but I take the view that I'm not going to compromise my recording of the piece by pandering to the applause in the realms of microphone positioning, level setting etc.