Both the EQ and the reverberation aspects of acoustic spaces can be simulated using the convolution approach, for example with Sony's Acoustic Mirror plugin, bundled these days with Sound Forge.

When recording the location sound you can also record an "impulse" file which captures the acoustic character of the environment. You can later use this impulse file and the Acoustic Mirror software to impose the character of your original environment onto audio that was recorded under other acoustic conditions (ideally as dry and uncoloured as possible, for best results).

Mark

If you are using the on-camera mic, there is no way to match the audio. EQ is only one aspect.

Put a mic right on the talent, I even use cardioid in noisy or echoey environments, and get the mic as close to the mouth as possible without picking up breath noise.

There is a reason sportscasters use headset mics.

Yes, there are several techniques. Look up "wall of sound" and "adding air to audio".

MY technique is to first resample to a 48K uncompressed proxy in SF. Then I look for the WORST audio sections to work on first. Try to omit anything really bad if you can.

This is a multipass method, and you can save presets to help automate this for the future. DO NOT try to adjust all the bands on one pass. The idea is to "de-muddle" the flatness.

Start with the low frequencies and give a 2db boost and save. You'll see the waveform start to expand (get air). Keep an eye on the peak levels and adjust the gain down accordingly until it just barely wants to "clip".

Do it again and see if it helps or hurts. What you're listening for is a bass presence, just enough room for the lows to "be there", not so much that it's distorted at all. Do it again until you notice distortion then back off. This point is your limit for the lows.

The next 1/3 set of bands include the vocals. Baritones and sopranos land in different bands so adjust the band(s) that enhance the vocals 2db, leaving every other band flat, and do the same as for the bass above.

Same thing now for the highs.

At this point your waveform should look nice and fluffy with much better distribution, with clearly defined peaks, and you can "see" much more.

Go ahead and save what you have so far into a new WAV file. That will commit and flatten everything. Then reopen it and listen to it again.

What you're looking for is "harmonic balance". HB occurs as a sudden surge in efficiency the makes the audio "pop", and the volume will go up slightly. That's the best I can describe it.

Anyway, use the results of the worst audio to determine the standard for the rest and you'll get more even sound across the board.

As a final polish, you can run the whole track through the "smooth/enhance" filter in SF.

Good Luck

"Anyone got the magic touch out there for matching eq's?"

I've tried similar tasks and hit the same brick walls.
The problem is that you cannot use Eq to recreate what happens in the air before the sound hits the mic. Not only do you get shifts in the frequency domain you also get shifts in phase. The difficulty of matching sound probably gets more impossible the closer the mic is to the sound source. Cardiod mics make it even harder due to the proximiity effect.

Bob.

As usual, I'm giving advice for the wrong question.

There's some kind of free software floating around out there that is supposed to smooth and enhance audio for voice over or broadcast.

It has a funny name like "luminary" or some single word name like that.

Since you are talking about the same mic that makes the situation a little easier.

The first thing I would do is compress the vocals pretty heavily. I would hit them with a good compressor (Waves C1 or RenC come to mind). Attack of around 30ms and release about the same. Perhaps 15dB of compression. This will really even out the vocal and right there give both tracks a similar feel.

Next I would go for the EQ. Since the mics are the same you are really going to be dealing with proximity effect differences. Farther from the mic will attenuate the frequency extremes. You probably won't be able to add bass and treble to the vocal take that was farther from the mic. As others have noted in this forum the sound is just different.

But you can thin out the closer to the mic take. I would do that by first increasing the high pass filter until enough of the low end is removed to make them sound similar. Then I'd do the same thing with a low pass filter on the high end.

If you then need to simulate the acoustic space you'll just have to play around with a really good reverb that you know how to tweak well. Again I like the Waves Trueverb for this kind of thing.

And finally a lot of this comes down to experience and having a really good ear for it.

Good luck!

Mark

Mark,

What I wrote above is about trying to resurrect a too low level recording with -50db peaks or less or enhance good levels but poor bass.

What I usually want to do is de-normalize. Can you comment good or bad, and give me your advice? I'd appreciate it.

Harmonic Balance is an electrical engineering concept, but the only descriptive I could think of. Do you have a word for it? Do you hear it too?

I hope it's not my ears are no good.

"What I usually want to do is de-normalize"

Normalise refers to setting the peak or RMS value of audio to a defined level, typically 0dBFS for peaks. For RMS it depends on the content, -10dB is friggin LOUD for music, OK for speech.

Anything recorded at -50dB is in very serious trouble not least it will be very distorted once bought up to normal levels, it's effectively been recorded using something like only 7 bits, not good. If it's also missing bass then it mostly likely resulted from a faulty cable or plug not being plugged in. I've rescued audio where that has happened, you could just hear what was being said at the wedding but it sounded very harsh.
My advice, assuming it's speech is to forget about the bass and roll off the top end at around 6K to try to get the harmonics down so it sounds less distorted. If it's music, it's toast.
Bob.

I guess my terminology is wrong, because I know what you mean, but the technical verbiage is too arcane for me.

I see digital audio as a fixed canvas, with a spraygun mounted in front, and recording gain as a flowmeter, and clipped peaks as paint that flew off the edge of the canvas.

And using that analogy, I sometimes get audio that has a lot of paint but it's all clumped up in the middle and sounds flat. So I try to aerate it using the EQ.

Crazy person, eh? haha

Thanks, I'm going to study more.

One technique that I'm surprised hasn't been mentioned yet in this thread is the recording of what's called "room tone" at the time of the original recording.
In case you don't know what I mean, it's recording at least 30 sec. (1 min. or more is even better) of location sound.
I try to do this any time I'm recording someone outside of my studio as I can use it to cover a multitude of errors that need to be corrected afterward (breath pops, etc.).
I've also used it if I have to bring the person back in again to re-do one or two sentences.
Record them dry, drop it over a track of "room tone" and it's very close to the location recording.

Mike

One technique that I'm surprised hasn't been mentioned

MarkWWW mentioned it. But it's too late now.

Recording "room tone" is what people do when they understand the benefits and develop the discipline to actually do it.
Sound Forge comes with Acoustic Mirror, which can do the job just fine.

I think the OP is stuck with d*cking around with EQ and maybe some 'verb.
... or maybe an impulse file somebody else made and which happens to work in this instance.

i always thought it called recording 'atmos' (as in atmosphere). always worked well in my day....

"the technical verbiage is too arcane for me."

BudWzr, if you haven't bothered to learn the terminology, then what makes you think people are going to take you seriously? We're all talking English and you're talking Martian. Why don't you just lurk in the background a little more until you get an education and can contribute something useful to the discussion.

As to the original posters question; the viewer is going to fully expect the acoustic character of the audio to change from room to room or from indoors to outdoors. Assuming that there were no serious problems during the recording, I think mostly that the issue is a matter of matching audio levels with perhaps a little EQ and a touch of compression here and there. The EQ is a matter of having an ear, there is really no piece of software that's going to make the decisions for you.

A lot of people will use the "Normalize" function in Vegas thinking that this will match levels. It won't, it just adjusts the peak level to a certain value. Unfortunately, peak levels are meaningless when it comes to determining how "loud" your final product is. The human ear doesn't determine loudness by the peak level, it determines it by the average (or RMS) level. Like I said, The "normalize" function in Vegas is useless because it only makes adjustments based on peak levels and that's not they way we hear things.

Audio compression and limiting is an art form and it takes a lot of experience to do it "correctly." There are no hard and fast rules to determine the appropriate average level, you'll just have to play it by ear.

You could use the Normalize function in Sound Forge, it can be set to normalize for average levels, which is what you want. Render out the audio from the entire timeline and pull it into Sound Forge, then mark each different clip or section that needs to be level matched. Use the Normalize function in Sound Forge and set it to normalize using "Average RMS power" and normalize to -20db and select "If clipping occurs apply Dymanic Compression." -20db is a good starting point, you will have to play with this value for your project. Once you have determined the appropriate RMS level, use that same value for everything and the loudness of each segment will be matched throughout your entire project.

You could also try using the free Levelator.

I've only used it once, only because there was hours of audio with levels all over the place and it was only some corporate thing that simply needed to be heard. You need to render the audio out as a separate wave file and then just drag and drop the file onto the application. For a quick and dirty fix it's great.

Bob.

Whew...I can smell your stinky breath.

==========================================================
BudWzr, if you haven't bothered to learn the terminology, then what makes you think people are going to take you seriously? We're all talking English and you're talking Martian. Why don't you just lurk in the background a little more until you get an education and can contribute something useful to the discussion

Woops, Bud,

This is not appropriate. Coming in here and throwing garbage, offending professionals, and spouting misinformation to newbies who are uninformed and trusting just doesn't cut it. Once again, by posting on these forums, responsibility is implicit.

I am going to suggest you use your mouse more, your keyboard less, start actually reading the posts, run your own tests and post the results for the rest of us to see.

Just my impression.

"Whew...I can smell your stinky breath."

That's the best you can do?

and it smells like a Heineken fart.

How old are you? Seriously.

John thanks for that useful direction

Just one question How does the Gain effect the levels?

Some folks crunch the Gain = flatten the audio especially if the clip is going to be played back on el cheapo speakers

Budwzr on the forum we often see a good scrap but that’s facts vs. facts that’s how we learn, stick to the facts
I have been getting a free education from this forum for some time that’s why I am here

Don’t throw away a good opportunity

Bud,

As Bob noted a signal down 50dB is seriously hurting for bits. If we assume 7dB as Bob estimated then you are looking at 128 discrete levels for each sample. Of course you will have a very squared off wave upon playback and it will sound harsh and garbled.

Compounding the problem is the fact that signal processing, and EQ in particular in my experience, is increasingly less effect with lower resolution signals. Basically you have a signal with very little information available. It's like trying to sharpen a blurry image. Despite what we see on CSI there is only so much that can be done.

As Bob said go easy with a good EQ to get as much sound quality as you can, focusing on intelligibility and not actual fidelity.

Sorry I couldn't be of more assistance but only Star Trek technology could save such a signal!

- Mark

Well.... the method I described above just sets a specific length of audio to some average loudness and setting a series of segments to the same average loudness will make them "flow" better. While the processing does affect the dynamic range, it does not measure or set it to a specific value. Generally, as you increase gain (volume) and limit the peaks, the average level (loudness) goes up and dynamic range (ratio between loudest and softest sounds) becomes narrower. Dynamic range is one of those things that's exclusively dictated by taste and experience.

As far as I'm concerned, audio is every bit as important as the video and it's even more of a "black art" than video. It's also difficult to teach someone anything audio related via a text-based forum since the English language is primarily visual and has very few words which effectively describe sounds.

We are (mostly) all born with the ability to hear but listening is something that is learned. One must spend a lot of time listening to a variety of commercial releases on your particular "reference" speakers so you can judge your audio versus the "norm."

Of course, BudWzr isn't going to agree with this at all since he is such a visionary production rebel.