Need urgent solution to fix sound
 

Hello everyone,

The last post was incomplete and incorrect (when I wrote "this is not really what happend" I meant: "cant give all the details of all that happend", please forgive me for the error...this is all true I give my word...:-)

I'm in the need of DESPERATE HELP from knowledge people in sound to fix this sound (I've attached 2 sound clips one in this message the other in the next). I cant give all the details of all that happend, but I tried to recap:

1. Recorded using XL1S + MA200, 2 mics one directional (boom-Senheisser for public) and one wireless (AKG PR40 for spokeperson).

2. Recorded St1, 12bits (one channel for each mic).

3. We did our test and everything worked fine, but theres something with the spokeperson that brings some kind of strange noise that leave the sound kind of worthless.

I would like to know the reasons for this in the future, since the mic (wireless) works fine with other people except this person, but know I'm not interested in this, I'm interested in knowing a way to fix this problem in post. I doubt there is, but the last thing you loose is hope.

I'm no expert in sound in any way yet, and the software I use in sound for post is COOL EDIT PRO 2.0. If theres a solution to fix (and a detail step by step procedure please) this you will save me. I spent many hours to recorded this and if we can't fix sound we wont be able to get paid. We have to give an answer tomorrow!!! (need to really know if is possible to fix)

PLEASE I don't need (at least now) why this happend or reasons why we didn't use other material, no help with this right now...what I need is possible solutions in post please (later on if anyone is interested I'll give more explanations why we had no other choice).

Thanks!

izotope RX is a nice program able to fix strange problems, i will try on this sample.
but i need a longer sample (or the full length original sound) and only the faulty channel.
look at the car beep sample at http://www.izotope.com/products/audio/rx/
the feature you need is SPECTRAL REPAIRE

Well, at first the left channel is missing, but it's just at a really low level. Normalising that channel only brings up perfectly usable audio, although it's a little distant. The other channel just has typical rf noise from the radio mics, a little annoying, but the two blended together aren't anywhere near unusable.

Just a bit of RF fizzing.

You sound really worried - what exactly is it that you are hearing that I'm not?

As for the cause - well, distance is the usual culprit - but you could find that the aerial was bent over making it less effective - the subject could also be sweating - salty bodies mask RF very well - maybe the clothes were rather tight, pulling the aerial closer to the damp skin - maybe it was exactly on the other side of the subject, meaning the body absorbs lots of the RF leaving less for the receiver. With a receiver on the back of the camera, reliable range can be amazing, or really short. I'm assuming it's a UHF system? Many run 50mw or so, but some don't, and are designed to be lower powered - it's often not a problem - but RF mics are never totally reliable.

Second Longer Sound Clip
 

Hello this is the second clip a little longer, I have erase one of the channels (the boom mic and left only the spokeperson).

Thanks for al the help I really needed!!!

it sounds like a stange clipping (somebody shaking a box full of sand).
again , i am pretty sure that Izotope RX can clean this.
i need to go back to work and i will try... if you can wait 1 hour.
post the full length file so i can see if the cure is ok on all the file.

More Info
 

Hello again,

I really need info on how to fix this problem, and I don't sound worry, I am really worry tomorrow we have to give the answer if we are going to bill these recordings (6 days, making 10 hours of videos) and right now I trully don't think we can. You people say is not really that big, but we dont know how to clean this sound and make it professional.

Giroud, thanks for the info, please try to fix it yourself (if you dont mind) too and post the results (I have attached a second sound clip for testing), what I really need to know if is possible to fix.

Paul, the mics used were: PT40 pro + C417 L from AKG, the receiver has printed: Ultra hig frequency UHF pr 40 - 802.525 MHz for Europe). He spokeperson had his unit on his side, cliped on his belt. He was using a suit most of the time and some with only the shirt (the problem happened in both situations), he wasnt sweating at all, at least buy eye. In later test, with our people, we move around, touch the antenna, bend it, the cables of the mic and everything was okay.

Check this...we even did a test on the same spokeperson to check and everything was okay!!! as soon as we start recording the problem starts again!!!

We couldn't not stop the recording and we had to keep going.

Later on we did another test (with our people) and recorded on the camera and no problem, this is really strange!!!

But as I told I first need to know if we can fix it and later worry about the real problem, but of course all the info you are giving me is priceless and I'm thankful you are telling me.

Trully desperate...keep the help and insight coming please!!!

Is It Long Enough??
 

Hello I've posted a longer clip...about 14 second long. I wanted to posted a longer one but seems like it was a problem with the upload and had to lightend the file. I can post a .mp3 but think the resolution is much worse to test.

Thanks.

p.d. Of course I can wait, I'm going to be checking all night for help on this forum.

p.p.d. Thanks in advance

I have to leave for an hour too, please keep posting, as soon as I get back I'll check everything so please dont think I'm gone.

you can send me a long sample of the original sound to my mail
francois(at)giroud(dot)com , my mail is not limited in size so even 200Meg are ok.
it must be the original sound, not a mp3 conversion or anything that has been modified. I will be at work in 30 min and will check this.

In the future, you really need to do a 16 bit recording.

Well I can tell you for sure the problem (even though you dont want to know just yet ) is your wireless. After the fact all you can do is to see if you can extract the sound from the boom mic of your presenter. solo the sound from the shotgun mic and see if you can use that track. if you can then you could possibly add in the wireless track to re-enforce the sound and reduce the effect of the wireless spikes. sound is like baking bread, once the ingredients are mixed and baked they are extremely hard if not impossible to separate.

As far as what caused it, it is definitely your wireless. It is not likely caused by that particular person but it could be caused perhaps by his cell phone or other metallic objects upon his person. but frequency spikes like that are usually the cause of other RF problems in your area. whether you want to hear this or not.

So your best bet is to see how salvageable your shotgun track is. (it shouldn't have the noise in it).
then use this for your main track, and bring the wireless track up a little at a time to re-enforce the shotgun track find a happy medium and if you can automate the volume on the wireless track automate a lowering of say 3 db or so to minimize further over the short spikes.

If the izotope plug in some one else suggested works then its all good. if it only minimized it a bit then you can use both methods together to achieve a usable result.

ok , it seems really hard to get rid of it, but setting a filter a 8500Hz (-70Db) is lowering a lot the problem.
it seems a simple set of filter should the trick

another correction with different filter

Hello again,

Gerry, is not that I dont want to hear about what caused the problem, quite the contrary, the thing is that I have a deadline (tomorrow. Here is already 20.30 and tomorrow morning we have know if this is "fixable" or not) to answer weather we are going to be able to get paid for this proyect or not and thats way I keep saying I would like to get ways to fix the sound...it doesn't mean I dont want to hear about the problem, of course not, so I'm greatful to get the insight you people are giving me.

The sound from the boom mic is not usable, because it was pointed to the people sitting in the hall and it only pick sound from the spokeperson sometimes not throughout the recording so this is not an option, the only solution is being able to fix the sound from the wireless mic.

Francois (I already emailed you), I have a 40 meg sound file, do you perhaps have a ftp server so I can send you the file???

Thanks.

replied to your mail

Excellent , sounds like Giroud got rid of most of it and it still retained a good amount of presence . one other solution you might want to try is utilizing a de-esser set at those frequencies, so the cut only happens when the noise crosses the threshold and leaving the rest of the recording alone.

Hello,

It got better, but we need to get rid of the background noise to be able to sell the proyect.

Other thing is that I really dont know much about audio postproduction, I only now how to volume things up or down, noise reduction (getting a clean sample) and not much more.

All the little filters, parameters and other methods I do not know how to aply, thats why I asked a step by step solution better with Cool Edit Pro if possible or in the software mention before, but with a step by step procedure because this is a total new world for me.

But I can wait for this a little...I would like to hear a cleaner sample so hear from myself to see if is possible to eliminate the remaining background noise (is it???)...of course only if you are willing to do so, I don't have any more hope.

I'm trying to send a longer clip at this moment.

Thanks

Jose, it may be possible to make it 'better' but it might not be possible to eliminate it altogether. Only you can judge how good it has to be to be acceptable. Have you considered looping the presenter's lines - re-recording them in lip sync in studio? If perfect sound is required that may be your only option. As far as step-by-step directions, it might not be that simple. Because of the variable nature of the noise you're going to have to go through the entire clip bit by bit and different parts of the material might require different settings of the equalization, noise reduction, etc. Folks here may be able to offer suggestions on things you can try but other than cleaning up the entire track for you (and an offer of compensation to them would be appropriate if they did) it's going to be very hard for anyone to give you a step-by-step cookbook solution for you to follow from beginning to end.

OK, is it me or do I hear a delay of the audio track?? It almost sounds like the speaker is speaking, and then there is a delayed "echo"... or is someone else speaking over him?

If it's an "echo", perhaps using phase cancellation of the original track with an offset? Not sure how much damage this woudl do to the primary track, but thought this observation might help suss out the problem... it's not just noise, but almost like a second delayed track to my ear.

Hello again,

Steve: For us usable sound is one without the background noise, the sssh sound.

We can not take the spokeperson to dub over in studio, this was a course with many people and he is not a professional dubber, no way his going to do 10 hours of dubbing. Thats way our only chance is to make the sound as good as possible and see if we can sell the product.

Dave: I cant hear the echo you talking about, but once again I'm no expert at this. Can you give steps to use the filter (phase cancellation of the original track with an offset???) in Cool Edit?

Thanks.

To make things clear,

I dont expect anyone to fix the clips (its about 10 hours!!!), I just want it to know if is possible to fix, hear a clip that has been cleaned and know is possible and hopefully get some guidence on the steps or things to look for on how to fix it, Francois is doing some samples and that is helping alot, hope he keeps doing it if he feels like it. Once I know this I'll do it myself of course, my intentions is not to rip people off by no means, I'm talking with desperation view! If it was video I have more tools and a little bit more of knowlegde but with sound problem and as big as this, I dont know what to do since all the tools available in Cool Edit seems very complicated and intimidating.

If I know it can be fixed, and hear it myself we can say we would sell the proyect, but if theres no way tomorrow we'll have to say we cant sell anything leaving 10 hours of wasted material.

We are going to say (at least from this poing) that we wont sell anything because sound is not good.

If in the following hours I get more help and maybe a sample that is possible we could change our response and spend the hours needed to fix it ourselves and learn what it has to be learn, but like I say all the time we need to know is possible and at least some guide on how to do it in Cool Edit or on some other software.

Thanks

ok i do my quick very best by applying filters to the file.
I do not try to fix local problem since you got 10 hours of recording and only global filter applied to all the files can be used in a realistic way.
the original
http://www.giroud.com/divers/Corrupted_Sound_long.wav
the corrected version
http://www.giroud.com/divers/Corrected_Sound_long.wav

basically i first remove some noise around 1000Hz, then removed cliks and pops, then setting a filter (around 8Khz) to remove the wireless fadings, then increased volume, then removed again clicks and pops.
it is far from perfect but not too disturbing.

Dot not play from the link because it cause sound to break with pops and click that are not in the file.
save it to your disk and play from there.

My thoughts...

It's the compander circuit in the radio mic.

You probably were not getting enough level and that has made the companding problem worse than normal. Sony have a reputation for this as does Trantec.

Compander circuit is a compressor on the TX and a expander on the RX to give double the dynamic range due to FM 's limited dynamic range.

Sometimes this sounds alot more useable when broadcast due to the average TV set having poor speakers. Your clip has it bad so do your best with eq/noise reduction and then listen on a tv set. Burn it to DVD and judge the results there. God luck and try recording with a higher level at the TX next time.

Hello Jimmy,

Thanks for the reply, but can you explain a little bit more, I don't know what you mean by TX???

I can increase the level in the camera (XL1s), but that increase the background noise a lot. What I always do is do a test, first silence, then I decrease the levels until I dont hear any more background (sssshhh sound) noise, make the person talk and see if the volume is okay, if not increase the level a little. I trully thought that increasing the levels would worse the problem even more (so I though lowering a little bit more was better).

So any insight if this microphones are any good, or how to be able to use them?? We are going to be making more important projects and can't be having these types of problems any more.

We are cleaning the sound clips and getting okay results, if you people are interested I'll post some samples in the future. It's going to be a lot of work to clean them all.

I would like to thanks Francois for his knowledge and the time he took to provide some samples and his thoughts on all these problem and the motivation he gave me to continue and of course the rest of you people for the tips and insights, if you dont mind keep them coming.

Thanks and a happy new year!

TX = Transmitter
RX = Receiver

Upon listening to the samples, I think Giroud's version sounds much better and may be very suitable for your needs.

I do have one question though. One channel, and this is probably where Dave hears the 'echo', is an open mic, my guess is this is the boom mic, yes? If that's the case, you may want to spend some more time cutting this thing apart. When you've got the lav mic going, kill the boom mic. Place the audio from the lav into both channels. When the panel is talking on the boom, kill the lav mic and put the boom on both channels. This may take you a while to do in post, but based on how it's recorded now, it may be the only way to get it up to the level you're expecting to hear.

Wayne

Hello Wayne,

You're right the open mic is the boom mic.

As for you're tips thats what we're trying to do, mainly from the lav mic but we should use the boom channel also to restore even more.

Thanks again.

There are two problems here. You have a poor recording (ignor the reason) and you need to fix it, but you are also lacking the technical audio skills to do it. Why not find someone near you who can make it the best it can be.

Broadband noise like this - and it's signal strength that is the problem, can be filtered, expanded, compressed or combinations of these - but what isn't recorded is lost - you may be able to mask the noise, but the words spoken at the time of the noise might simply never be recovered. So what you will have to do is treat all the problem areas, then find the same words on the boom track, even if they are distant, sync them up and attempt to replace the lost words with the other track. You'll need to alter the eq to match, maybe filtering out noise above or below the speech frequencies and then see if this can 'lift' you out of trouble during the worst bits.

All this is disaster recovery.

It's too late to tell you all the things you should have done at the recording stage, but having somebody on sound who knows good from bad is pretty essential. A poor sound op is worse than no sound op at all. What they should have been doing is hearing this noise and stopping you recording until fixed. The best sound ops spend a lot of time listening for things they don't want to hear, rather than things they do!


If you have a proper audio editor - not the one in the video editor, you will be able to make it better - how much, depends on the person doing the editing as well as source material.

Sorry
P

Hello Paul,

We are already making progress and we are making the sound useable.

We try to do test before recording and everything. System (mics) work previously and no problems at all, except for this project (very, very strange). All the pre-test for these recordings worked but in recorded time the sound appear (very, very strange. Later test were recorded--video+sound--and everything was fine).

We could not stop the recordings or the event, so that was not an option.

Like I said I'm not expert in sound, but I'm learning a lot in the proccess, so not everything is bad (trying to be optimistic and see the bright side...:-).

Thanks for the thoughts.

Glad to hear that you're making progress.

Couple of points. It's not enough to do tests before recording. You must monitor critically DURING recording as well. A burst of RF noise from someone's Blackberry, for example, may not happen during the test but jump up during the shoot and bite you on the backside - when something like that might happen is a roll of the dice.

You may well have not been able to stop during the event but are you sure? I've been at a number of concert tapings for network broadcast production where they did, in fact, stop tape and restart segments a number of times when somone flubbed a cue, or dropped a line or a technical glitch occured or any number of other things happened. If your event was held specifically in order to tape the presentation, it's not any different from any other TV production. The fact an audience is there during the shoot is irrelevant.

Hi
when i was talking about levels I meant that the belt pack transmitter needed it's gain pot setting just as critically as your cameras level pot/knob.

Each stage of the chain has to be set and optimized. If the gain is too high on the belt pack it will distort before being transmitted and if too low it will mean you hear back ground noise and with certain systems, companding, which sounds like breathy noises that come on with the speech.

What wireless were you using??

the sony uwp-c1 never let me down.
use uhf frequency, lower vhf is always fading in some way.

The Sony's are well known for breathy and noisy companding issues. They sound bad. I have a UTX system with an ECM77 and it lacks the top end clarity which this mic should exhibit. This is the wireless' fault. The actual mic is good but the radio and electronic circuits sound bad. Frankly I prefer the Sennheiser without diversity (EW100 MKE 2.4) system I have for sound quality.

I have gone off the Sony since listening to it's compander circuit even when fed with decent mic level. Having said that it does not tend to translate or come across when listening to it on a tv set due to the average crapness of tv set speakers which hide a multitude of sins.

Try playing with yours and others radio's a bit more and don't expect to get it right first time. I would recommend you got a bit more experience with your kit before doing paid work as you will forge the wrong reputation for yourself by not knowing your kit.

Hello Jimmy and everyone else,

I think I'm going to start a new thread about the equipment we have to know if the are good for work, but here's the info:

The TX: AKG PT 40 Pro, the RX: UHF PR 40

Mic is C417

The RX only has one dial which is for volume.

Are they (tx's, rx's and mic's) up for professional sound and work? Is it strange what happend or this units are not too good?

The transmitter audio gain adjustment is a screwdriver-slotted pot inside the battery compartment. Did you remember to adjust it up to the point you got an occasional weak flicker of the transmitter's 'clip' indicator LED?

Hello Steve,

I did not touch it up, I left it as it came, could you explain a little more the steps and how to do it, I did not understand completely the method you mention.

I did not touch the dial since the mic's were functioning fine inside the studio and in the test (outside test were fine too...was I lucky?) but then happened what I already told.

Mic your subject, open the transmitter's battery cover, and with a screwdriver turn up the level pot until the transmitter's clip LED just barely flickers on voice peaks. That gives you optimum levels at the transmitter. Then turn to the receiver and set its output level to match the camera or mixer inputs.

Some of the noise in the first examples you posted sounded to me like cell-phone bursts, perhaps from people in the audience, and there's no adjusting that'll prevent that. Phones "phone home" every few moments to let the system let them know where they are, even if their ringer is turned off and the phone on silent. The only prevention is to make sure everyone in the room has their phones completely powered off.

Okay got that.

You mean like in the TX with a screwdriver? and how to match it with the camera? you mean by ear or trying to put the same values as the ones in the camera? Sorry to sound stupid, but I want it to get this as clear as possible.

Yes I knew some were cause by cell phones, but the ones that worries me are the breathing type ones that weren't suppose to be there.

Thank you for giving me all these details!!!

Note - I'm just reading the online manuals here, I don't actually have one of these units.

I mis-typed - you adjust the level in the transmitter battery box until the red clip LED on the RECEIVER just flickers, I think I said before that it was on the xmtr. The receiver output is fixed - the volume control is only for its headphone monitoring. The specs say it's at -6dBm which kind of straddles the middle ground between consumer -10dBv line level and professional +4dBu line level. You could set your camera input to line or insert an inline pad and drop it to mic level. I'd first try setting the camera to line level and adjust the camera's input gain so that the camera meter hovers somewhere around the -12 dB mark during normal speech.

I know I'm a broken record but pre-shoot tests, although vital to do, are no substitute for monitoring the camera audio with a good set of cans during the actual shoot.

Hello Steve,

This is worth gold for me! Very hard to find info on how to set up these types of mics.

Whats kind of shocking is that you are telling me to sep up the camera to LINE instead of MIC, I'll have to do some test another day, I'm leaving the studio.

But I'll check dvinfo at home for after post. Any more settings you have I'll take 'em and now that we are with it, what lav do you recomend for professional work that are good (but not so expensive that you have to be a network/tv studio to buy them)?

Many thanks.

There are a number of good lavs to consider but I'd suggest making sure you've got everything set up and working right with your stock mic before investing. Switching mics before then will just introduce more variables into the matrix and it will be hard to evaluate just what's going on.