Noise Reduction 2.0 50% off

Jay Gladwell wrote on 5/7/2003, 10:00 AM
I didn't see a forum for Noise Reduction 2.0, so I'm forced to post here, since I would use it with Vegas.

Got an e-mail from them about a "special" on the product. Went to the web site and listened to the samples--before and after. With the exception of the "pops" and "crackles" files, I couldn't hear any difference in the other files.

Am I the only one with tin ears?

Jay

Comments

TheHappyFriar wrote on 5/7/2003, 10:39 AM
On the weekend I work for a place that makes weekly radio show. They take music from old records, cassete, etc. They say it works great. It doesn't make the sound as good as if the origional recording were on CD, but it's make it better.

If you know how to use it.
Chanimal wrote on 5/7/2003, 10:58 AM
I have used Noise Reduction extensively to clean up sound filmed in offices (where there is a constant hum of florecent lights, the air conditioner, the noise from my previous camera, etc. Works very well--I wouldn't work without it.

***************
Ted Finch
Chanimal.com

Windows 11 Pro, i9 (10850k - 20 logical cores), Corsair water-cooled, MSI Gaming Plus motherboard, 64 GB Corsair RAM, 4 Samsung Pro SSD drives (1 GB, 2 GB, 2 GB and 4 GB), AMD video Radeo RX 580, 4 Dell HD monitors.Canon 80d DSL camera with Rhode mic, Zoom H4 mic. Vegas Pro 21 Edit (user since Vegas 2.0), Camtasia (latest), JumpBacks, etc.

mikkie wrote on 5/7/2003, 11:19 AM
I've got it, it works, is it better then other stuff on the market? I'm not sure. A LOT depends on what kinds of noise you're trying to eliminate.

"With the exception of the "pops" and "crackles" files, I couldn't hear any difference in the other files."

The difference after using the regular noise reduction is audible, where you're subtracting basic noise captured from a silent selection of your recording, but a lot depends on the amount of noise obviously, and hearing the noise beforehand depends in part on your equipment.

I picked it up in a bundle from SOFO for about $100 if memory serves, and I considered it a good deal. On the other hand, I used to use Dart Pro (liked it a lot), and there are all sorts of competeing products you could look at, including stuff like Goldwave and Cool Edit that include some noise reduction capability. EasyCD Creator 6 plat. includes it, as does the stuff that comes with some creative products.

If it helps... any noise reduction software is kind of like compressing video - you have to choose a compromise between the data you lose and the end result. I don't know if I'd use any noise reduction stuff outside of an audio editor like Soundforge. While I often use the noise reduction plugin as a part of normal processing to remove tape noise etc., the plugin earns its keep with problems, and those I normally isolate in soundforge, often enough doing some manual editing as well... Not sure how I'd handle them outside of an audio editor, or if I could.
BrianStanding wrote on 5/7/2003, 11:37 AM
I have no experience with competing products, but have been very happy with SoFo's Noise Reduction. Does a great job at reducing background noise from traffic, air conditioners, etc. when you are forced to shoot video in a less than ideal environment. I've also used it successfully to restore audio quality to old LPs.

Takes a little practice to use effectively, but once you get the idea, it's a cinch. Maybe there's snazzier products out there for high-end audio recording, but for my purposes (documentary video), Noise Reduction can't be beat.
Jay Gladwell wrote on 5/7/2003, 1:51 PM
Does it act as a plug-in for Vegas, or is it a stand alone?
atedee wrote on 5/7/2003, 2:31 PM
I got the same e-mail but I guess SOFO has to straighten it out. The e-mail says the regular price is $399 so fifty percent off makes the price $199 when using the link from their email. Now, if you go down the SOFO site and buy from there, it will give you a $279 regular price. What the??? I'm planning to get NR2.0 but would only buy if I can get 50% off the $279 regular price. That would come out to $140 only. Now, SOFO what can you say about it?
mikkie wrote on 5/7/2003, 2:40 PM
"Does it act as a plug-in for Vegas, or is it a stand alone? "

It's a plug in, or rather a set of plugins that can be used by anything DX - shows up as a DX plugin in any app that supports DX audio plugins, as VV4b.
atedee wrote on 5/7/2003, 2:42 PM
"It is a directX plug-in so NR2.0 AFAIK can be used as plug-in to Vegas and SoundForge as well as other audio programs like Cool Edit Pro, etc."

opps..Mikkie beat me to it..
BillyBoy wrote on 5/7/2003, 2:53 PM
To use the Vegas Noise Reduction filter or others that employ capturing a noise print by far the most important thing is being able to shape the filter so it removes as much 'noise' as possible without damaging sounds you want to keep.

The "secret" is the default setting while pretty good, and automatically generated normally doesn't compare to setting points along the waveform manually, just like you can with color curves on the video side. The more difficult the noise is to remove, the more important customing the filter is to get best results. Again, like with the color curve just a very minor adjustment sometimes can make a big difference.

First capture a print of the noise obtained from a second or so of "silence" which is a portion of your project where there is ONLY noise you want to get rid of that is more or less constant, like a buzz or hum or some annoying sound from a camera, air conditioner, whatever.

Now look at the waveform generated (click on noiseprint tab in Vegas) and you'll see a series of points all along the waveform. The scale at the bottom of the graph represents frequency, while the vertical scale represent how much noise filtering is being applied. At the default setting the points closely hug the waveform or in other words the same amount of filtering is applied across the entire sprectum. The points that control this don't have to stay in their default position.

With a typical noise there is one or more 'hot' spots along the waveform that falls within a narrow frequency that's causing much of the noise. By applying MORE filtering at these points you'll get a better result. Generally higher frequencies can be filtered more without introducing artifacts. Once you get good noise reduction you can use several of the other Vegas built-in audio filters to further enhance your audio.

Just keep in mind like with the video filters it is easy to get carried away and go overboard. If you apply too much noise filtering you'll probably start removing too much of the audio you want to keep. That's because most 'noise' at least partially also falls within the range of those sounds you wish to keep. So remember you're also effecting 'good' sounds, so trying to remove all noise isn't a good idea.

Jay Gladwell wrote on 5/7/2003, 3:44 PM
Thanks, BB, for the info. Fine tuning sound is not my strong suit!
MixNut wrote on 5/7/2003, 7:17 PM
Hello,

Sound Forge's NR 1 package has been a mainstay for corrective processing in my audio post department since SF version 3. We upgraded to NR v2 as we also upgraded to SoundForge 5 and then 6 to [finally] work in 24bit on the PC side.

All 3 engineers [with 3 different systems/installs] have since noted that NR2 does not seem to work nearly as well as NR1 [which nearly equaled our CEDAR NR package in many cases]. In spite of hours of settings tweaks and noise print resampling, I've not yielded *one* useable result from NR2 in the last year, now that I stop and think...my other guys agree...no dice for them either!?

All previews indicate useful noise reduction...toggling to "keep residue" shows some reduction result, but in the end, the NR process is not effective without significant distress on content.

Is this package *really* 24 bit capable? Has anyone else yearned for NR1's relative superiority?

Are we the only ones who think this package is not working?

What gives?

TIA ~ MixNut
MixNut wrote on 5/8/2003, 10:33 PM
Anyone?
Chanimal wrote on 5/9/2003, 12:18 AM
I have had very good success with the latest version. I often play with a lof of the settings and perform the reduction in different phases. First, I normalize the sound (usually voice). Only if it is really weak do I use noise hammer. Second, I use the equalizer to try and isolate and drop the exact frequency. Then, if I still have much trouble (about 40%), I go into noise reduction. I accept the default sound file capture settings and then go to the slide bars and tweak as neccisary. Sometimes I've eliminated about 95% of the remaining noise. Other times the sound starts to turn metalic and I back off (and increase the volume of the background music). It has saved a few "wind" blown scenes (filmed by committee that didn't turn against the wind for their shot) that would have been trashed. There could be better "tips" that could have reduced my learning curve, but it works well for me. P.S. I never use it on the Vegas timeline--I always pull it into Sound Forge 6.0 (I got it as part of a SF 6.0 bundle).

***************
Ted Finch
Chanimal.com

Windows 11 Pro, i9 (10850k - 20 logical cores), Corsair water-cooled, MSI Gaming Plus motherboard, 64 GB Corsair RAM, 4 Samsung Pro SSD drives (1 GB, 2 GB, 2 GB and 4 GB), AMD video Radeo RX 580, 4 Dell HD monitors.Canon 80d DSL camera with Rhode mic, Zoom H4 mic. Vegas Pro 21 Edit (user since Vegas 2.0), Camtasia (latest), JumpBacks, etc.

farss wrote on 5/9/2003, 7:13 AM
I only play around with audio as a passtime but have a fair understanding of electrical engineering. I was tempted to buy this package but from what I'm reading its a fairly crude approach. It seems to me it works something like this:
Do a spectral analysis of the offending noise and then subract that from the polluted signal using notch filters set to match the derived spectrum of the noise.

A good many years ago when audio was amplified through little glass things with lights in them hum was a big problem, one way to get rid of it was to add more in but out of phase. Signals cancelled out, no more hum.

Why can't this trick be used again? You have a sample of what you don't want, loop it, invert it and add it to what you want. Now I know this is going to take a lot of smarts to keep the looped signal in sync with whats hidden amongst what you want but thats what all that computing power Intel gave us is for.

Maybe thats what the more expensive packages do?
mikkie wrote on 5/9/2003, 10:30 AM
In all fairness, the noise reduction plugin actually does a lot more then just remove constant or semi-constant noise - that's just one part of it. Also, like a lot of software, there's a place for higher end stuff, and one for lower end products which function more or less on autopilot. Just as we use notepad on many occassions instead of Word, there are simpler programs that might solve one's immediate problem 80 or 90 % of the time.

"It seems to me it works something like this:
Do a spectral analysis of the offending noise and then subract that from the polluted signal using notch filters set to match the derived spectrum of the noise........"

It is possible to take a sample of the offending noise, & then subtract it. It gets involved though, and you have to consider harmonics. I read an excerpt from one of Jay Rose's books (dplay.com) that went into this, possible one of his articles for DV (DV.com). You also have the problem that noise isn't always stuck all by itself in a particular frequency, but blends in all to often with those needed/used by speech and so on. And while it is possible I imagine to do a fair job manually, lot of folks don't want to bother, and not all (especially music folks?) are say engineering inclined.

Stuff like the noise reduction in Goldwave tries to use some fairly std algo methods to remove mostly just the noise and not any of the stuff you want to keep. Sometimes it works well enough. The SOFO plugin uses, does a more involved analysis, so in my experience at least, it does a better job of leaving non-noise alone.

If it helps... For basic tape & system noise, once I open the audio in Sound Forge, I capture a noiseprint from the intro usually, then I normally use the defaults for fast system, drop the Attack Speed down to around 76, boost the reduce noise by to 13 - 15, and let 'er rip. If I have any work to do on the track, I'll do it now. Then I'll normalize the track, and run the noise reduction again using mode 3 and reducing noise by about 10. In a very few cases with critical audio in the noise effected range, I've found it handy to go to the graphic EQ in Vegas and mimick the envelope shape from the noise capture window of the plugin - this way I can add a compromise boost, tinkering as much as needed, & it is still quieter then before I ran the plugin.

IMO, if you've got a problem audio file, one with some serious noise, it's still best to isolate the problems sections as possible in something like Sound Forge, and individually use your tools (including this plugin). If you've got bad background noise that's in the frequency range of the dialog for example, might not hear it too badly except during quiet passages without any or much dialog. So you can crank up the filtering where it's important, where you hear the noise worse, and either apply minimal or no filtering where it's less audible. Point is, there is no miracle cure I'm aware of, and noise reduction/removal can take some time if you've got a real problem audio file.
BillyBoy wrote on 5/9/2003, 12:21 PM
I'm by no means claiming any expertise... my experience seems to suggest best results are obtained tuning the noise print thus getting better reduction of the 'noise' assuming the noise is constant, like a hum or motor noises from the camera.

The 'technical' term (I think) is attenuation. Just to give a somewhat silly non technical example, if you place your hand over your mouth when you're talking, you'll muffle your voice. The harder you press your hand over your mouth the more muffled your voice will sound. So while you'll quite your voice, you also are distorting it to the extend it gets harder and hander to understand what you're saying.

That's why the better noise reduction packages allow more control over attenuation and you can more finely adjust WHERE along the captured noise print the extra filtering is needed, so in effect you only muffle the worst parts the most leaving the rest of it alone or only apply less attenuation.

There are many other excellent audio filters already build into the regular Vegas package. You may consider FIRST applying either a high pass or low pass filter before attempting to do sound reduction. Also tweaking the noise gate, by carefully adjusting the attack and release times can help alot depending on the noise. The noise gate is what its name suggests, a gate in that you are controlling when the filter opens and shuts, more or less. I like to start with attack times around 20-35 milliseconds and release times two to three times the attack value. As far as attenuation, you want to avoid too much otherwise you take away too much ambiance.

So start with a fairly low valule for attenuation, maybe 15-20 db, then build it up if there is still too much noise. If you apply too much then you'll get digital aliasing artifacts, and distort the audio much like applying too much/little hue kills the overall appearance on the video side of things. Like with other things its a trade off. If you try to remove all the noise you'll take away the life of the audio and maybe end up with artifacts.
MixNut wrote on 5/11/2003, 10:36 PM
All of these responses are helpful in describing how to apply noise reduction...I have long used SF's NR v1 with great success and can vouch for its effectiveness.

What I now encounter, however, is the failure of these same approaches to yield nearly so good a result since "upgrading" to version 2...

Can anyone verify that v2 NR is capable of correctly processing 24bit files? Does noone else find v2 *less* effective than 1?