izotope-rx4-denoise

It’s sometimes tough to write about audio tools precisely because they tend to bundle together a lot of features. So let’s step back and consider why they tend to do all of those things.

With audio repair, it’s a pretty easy explanation. From your perspective, your sound is $#*$#ed up. You want to get it un-$#*(&ed up.

Of course, in reality, there are tons of variables. The context can change: You might be repairing sound from a recording of instruments. You might be fixing dialog. You might know what you’re doing – even on big-budget TV and film, recordings can wind up with sound problems. Or, let’s be honest, you might kind of have no clue what you’re doing and wound up with $(&*ed up sound because you yourself $#(*&ed it up. (Uh… yeah, been there.)

The underlying problems can be varied, too – even in a single recording. Different takes didn’t match. There’s hum. There’s noise. There are unwanted sounds.

So, all of this is to say, over the years I’ve seen a number of general purpose repair toolkits, along with specialized toolkits. Right now, the one iZotope makes is special in that it bundles all the things you might ever need to fix audio into a single toolset that can work for more or less anyone. It doesn’t entirely eliminate the utility of more specific tools here and there – some of which may already be in some form in your DAW. But the tools are unusually advanced, unusually complete, and I think at the moment there simply isn’t anything else that does as much. If this is a First Aid Kit for sound, it’s kind of also a fully-staffed Emergency Room and Operating Room. Not like a field hospital. Like Mount Sinai.

I’m going to be talking a bit about iZotope this month partly because I’ve noticed that this year, they’ve shifted focus a bit from just reeling off features to talking about what they were doing in the first place. So I had a chat with them about RX and Ozone, in particular, two of their flagships, and it led to this.

RX4 is particularly useful in TV and film production because of the likelihood those users need to fix stuff – more on that in a moment. But it is worth considering in a production environment if you ever record anything for any reason.

Among other tutorials, iZotope have produced two videos that nicely illustrate what I mean by that.

First, this tutorial is great, because rather than the typical software demo sound, it’s — well, it’s awful. Buzz and cough and bangs — I’m sure this sounds familiar. To be honest, this is the point where iZotope RX becomes necessary, because otherwise you’re probably better off just going back and re-recording. What you can see is that the toolset of RX can be a musical one.

Even more in the musical direction, another interesting video from February released by iZotope deals with how to combine multiple takes that don’t match sonically. Here, the approach is potentially as creative as it is remedial; you get the opportunity to merge takes that otherwise wouldn’t fit.

I spoke to iZotope a bit about how they find users working with their tools, and they were willing to share exclusively with CDM a brief interview they did with Christian Beneventura, a re-recording mixer and engineer. That’s a fairly specific job – though also a reminder of the range of industrial gigs available to people with a solid sound/music background. (Use those ears, in other words.) Mr. Beneventura has an amazing resume, as you’ll see on IMDB, including The Vampire Diaries, Choke, Glee, and now Daredevil and The Following. (In fact, if you haven’t at some point heard his dialogue editing, you probably haven’t switched on your TV or Netflix lately, it appears.)

And he’s worked out how to deal with sounds in New York.

This is not some sort of advertisement for the product; to me, it’s interesting to hear this stuff and see how it works on production. Interview courtesy iZotope and Sean Greenhalgh:

christian

Why do you think you’ve been successful at your chosen craft while others have burned out or faded away?

I believe I’ve been successful because I’m constantly trying to get better at it everyday. Even though I have been doing this for quite some time, it’s important not to get complacent. I’m always trying to research and try new plug ins or try different techniques to get faster and more efficient. It’s important to keep learning and evolving because the technology does so.

Daredevil seems to be a very dynamic sounding show. Was this a conscious decision?

It was a conscious decision. From the very beginning, we knew that sound was going to play a huge part of this show. The character of Matt Murdock is blind and trying to convey how his other senses help him “see” especially sonically, is very critical to the show. I think every part of Daredevil’s soundscape was deliberate. Creating the backgrounds of Hell’s Kitchen, deciding what exactly Matt Murdock hears in the flurry of city chatter, having the rate of the heartbeats that Matt hears hit at exactly the right points. Every part of the editing and mixing process was very meticulous and we are very proud of how everything turned out.

Why do you use RX?

I use RX because it’s a life saver. There are many scenes that I have cut that would not have been possible unless I had RX Spectral Repair. RX has really changed the way I edit because I have integrated using the plug in within my editing routine. When I first began using it, I thought it was the future. “How can you not edit with this?” I said to myself and colleagues. RX is so reliable and I know what it’s going to do for me. Brake squeals or back up beeps, no problem. Lavalier mic cutting off or boom mic bump, got it covered. I use every single plug-in in RX because it’s reliable and I know it will get the job done.

What are some of the challenges of working on audio recorded in NYC?

Dealing with audio that’s been recorded in New York can be tough because of the pure fact the city is inherently noisy. Extraneous city sounds that you can hear when dialogue is being spoken is always a pain to take out. The traffic, brake squeals, people talking and shouting, music bumping from cars, New York will find a way to make a scene difficult to edit. RX spectral repair has always helped me in this bind. I could easily see brake squeals to take out over dialogue as well as people who talking who aren’t supposed to. I did work on a scene where it took place at Washington Square Park and there was a street performer drumming and singing over the dialogue and they didn’t want to ADR the actor. Sounds impossible to do but Spectral Repair and some fancy editing helped me achieve that.

Of course, to me, this is doubly interesting precisely because I’m not experienced at this stuff, nor many of the people I know. Sound we record for a video production is necessarily going to have problems because it’s not my area. And since I can’t afford someone like Christian, we have to DIY if I want to fix it. Ditto instrumental recordings. To me, the software doesn’t replace those skill sets – on the contrary, when you do have to learn this stuff yourself, you appreciate why those folks are so valuable. And, if you are willing to invest the time, you might even find a professional path you would otherwise not expect; there is huge need for people who are skilled to solve these problems.

I haven’t found anything coming close to what iZotope’s tools can do, but I would love to ask our readers – particularly any of those working in these industries – what you use. RX? Other tools? A combination? Let us know.

See also the excellent Designing Sound which covers these topics more regularly:
http://designingsound.org

While it won’t turn you into a TV sound editor overnight, if you want to take your first baby steps toward fixing the problems above, iZotope has some videos for that, too:

And for more on RX4 itself, our friends at Sonic State did a great video tour of what’s in this tool:

  • flplsx

    RX is fantastic and has saved me countless times. But like most tools, it’s not always the best for every job. Waves has a few offerings, like WNS or even the C4 which work in some scenarios. A more recent contender, Acon Digital, has a few products that rival RX, though without much in the way of an “auto” mode. At the top of the line, there’s Cedar, but that’s out of reach for most individuals.

    All of these have a learning curve and it takes some experimentation to get the best results. Often times, particularly in noise reduction, multiple passes with targeted settings in small increments works better than larger moves. Random tip: the Declicker is sometimes fantastic at getting rid of or reducing wet-sounding (not reverb wet, but actual moisture wet) voice recordings.

    • Yeah, the Waves stuff is nice, too, even if it doesn’t have a suite as here. Have you used Sony SpectralLayers? The editor-style approach there I find appealing.

      • Lee Faulkner

        In my experience the right tool in noise reduction and restoration is always a bit of a crap shoot. Sometimes the built in NR tools in FCP X ( in my case ) work great when RX 3 ( haven’t got 4 yet ) don’t work as well. sometimes RX beats SoundSoap or SoundSoap Pro … Sometimes not. Occasionally slicing, expanding, gating and adding atmosphere manually is required. I’ve tried Spectral Layers … again it works until it doesn’t!
        Frustrating if you want just one catch all tool…

        Nothing beats well recorded original material. Though I know it must be tough to get NY to shut up.

  • flplsx

    RX is fantastic and has saved me countless times. But like most tools, it’s not always the best for every job. Waves has a few offerings, like WNS or even the C4 which work in some scenarios. A more recent contender, Acon Digital, has a few products that rival RX, though without much in the way of an “auto” mode. At the top of the line, there’s Cedar, but that’s out of reach for most individuals.

    All of these have a learning curve and it takes some experimentation to get the best results. Often times, particularly in noise reduction, multiple passes with targeted settings in small increments works better than larger moves. Random tip: the Declicker is sometimes fantastic at getting rid of or reducing wet-sounding (not reverb wet, but actual moisture wet) voice recordings.

    • Yeah, the Waves stuff is nice, too, even if it doesn’t have a suite as here. Have you used Sony SpectralLayers? The editor-style approach there I find appealing.

      • Lee Faulkner

        In my experience the right tool in noise reduction and restoration is always a bit of a crap shoot. Sometimes the built in NR tools in FCP X ( in my case ) work great when RX 3 ( haven’t got 4 yet ) don’t work as well. sometimes RX beats SoundSoap or SoundSoap Pro … Sometimes not. Occasionally slicing, expanding, gating and adding atmosphere manually is required. I’ve tried Spectral Layers … again it works until it doesn’t!
        Frustrating if you want just one catch all tool…

        Nothing beats well recorded original material. Though I know it must be tough to get NY to shut up.

  • cola verde

    It won’t “save my bacon” ….I am vegan…..

    • Actually – no you’re not. Not in the specific case here, because “save *my* bacon” implies that you’re the meat. 😉

  • cola verde

    It won’t “save my bacon” ….I am vegan…..

    • Actually – no you’re not. Not in the specific case here, because “save *my* bacon” implies that you’re the meat. 😉

  • dustinw

    Anyone with feedback on how RX4 compares to SpectraLayers? I am interested in getting one of the two for cleaning up background noise etc.

    I am a Soundforge user so I have an upgrade path that saves a bit of money (making it slightly less the RX4 if I buy it at the same time as I upgrade Soundforge).

    Retail Prices:

    SpectraLayers: $399.95
    RX4 $349.00
    RX4 Advanced $1,199.00 (out of my budget range)

  • dustinw

    Anyone with feedback on how RX4 compares to SpectraLayers? I am interested in getting one of the two for cleaning up background noise etc.

    I am a Soundforge user so I have an upgrade path that saves a bit of money (making it slightly less the RX4 if I buy it at the same time as I upgrade Soundforge).

    Retail Prices:

    SpectraLayers: $399.95
    RX4 $349.00
    RX4 Advanced $1,199.00 (out of my budget range)

  • James

    I’ll 2nd flplsx’ comments and expand on a point of contention:
    I’m concerned as a sound person (production dialog) that the representation on the Izotope site is not really helping film-makers to appreciate the price of a technician who is capable of salvaging audio for a project.
    Each interviewee swears that it is a time saver and project saver. What all those folks have in common is a regular, contracted means of employment in-house, with a show, or on A-list films. In other words, they can all absorb or account for their editing work in the asking price and are by default appreciated for their work. Albeit they are also probably subject to the same recession-pinching strategies that those lucky enough with work are facing, namely bosses who expect us to make up for the work of the 2 colleagues that were fired or face the notion that we too can be replaced.

    Full-budget films and broadcasted shows can then lean into ADR, and subsequently employ another set of professionals. But what of the 1st-time film maker, ignoring all signs of good measure who then realizes that the 10K that they raised on kickstarter was squandered because every scene they shot would be otherwise thrown out due to the corners they cut on sound. Isn’t it bad enough that sound people are probably not being employed by these well-intended, potentially ambitious and misguided projects?

    So what then is the cost of this mistake? Well it used to be ADR across the board or dumping your production HD’s at Tekserve’s recycling program. But now we are encouraging if not incentivizing a new wave of film makers who think these mistakes can be “fixed in post”, a phrase that used to be synonymous with “famous last words.” Essentially repeating the same mistakes that we’ve been trying to reform in the film-making process.

    So what’s the real story with audio restoring software? If it’s going to save your ass, what’s it’s worth in weight? Are we telling these same film makers that these mistakes are also less costly? Having worked extensively with RX, I can do wonderous things with the outcome. As far as I can tell, the methods in RX4 can be down and dirty and apparently solve things with a preset. For me, I haven’t noticed the out-of-box recipes to fix better than 70% of the problem, let’s quantify this with say 4db of restored headroom, or a baseline of removed crackles and thuds. More typically running the ambient noise sampler and pressing “process” will result in a lesser performance. Mileage will vary.

    How about for a performance that would keep ADR at bay or genuinely salvage a scene/film? Well that requires an expert. And if you want them to put their discerning ears to the test while working out the parameters of each tool while paying them the afterthought budget of your film project, then you might as well have hired child labor. Because the cost on the many hours of listening, A/Bing, and the like is exhausting work. You might as well go into the diamond business instead and hire little children to cut your gemstones what with their nimble little hands and unhampered 20/20 vision. I’ve finished sessions where I literally thought a truck was backing up into my kitchen while I’m drooling over my dinner (enough to make me duck), soiled from days straight of file management, listening, batch processing and the like. That’s how disruptive iscolating low-end, comb filtered, or phased noise (and bad boundaries with an employer admittedly) can be. But if you need to rescue 6-12db for your mix that’s the real cost. Or you pay real people their real worth up front and make real mitigated decisions about how you are going to shoot.

    So why then has izotope sold this as another opportunity to dupe post-production workers into devaluing their contribution? That’s the message I get at least. “See, it’s easy!” You show a few successful cats who get a living wage based on their editing skills, and for the remainder of us who might consider audio restoration a specialized discipline we need to clarify for people that we’re singularly saving your ass and that’s worth something. Consider the ratio between your production schedule and the amount of time you intended to edit this thing, and then multiply what you should have paid your location sound guy by said factor.

    I actually think this area of craft should really form a guild.

    • Right, although this same argument can be used with mastering and mixing tools as well as post-production… or, really, *anything*, for that matter. My sense is that people either understand the value of skilled professionals or they don’t. Anyone leaning on the tool and saying “the devil made me do it,” as it were, likely was missing some experience and understanding in the first place.

      I don’t disagree with anything you’re saying, but you also observe that a low-to-no-budget DIY project is going to need to DIY its post needs, too – just like everything else. That’s not putting anyone out of work, because they didn’t have the resources to begin with. Sometimes those people might wind up building a career later.

      It’s tough – in software, you’re trying to market to as many people as you can. And in some sense, they’re just trying to show what the tool itself can do. I know that can include marketing to people who actually know what they’re doing.

      Anyway, I was a little surprised when iZotope pointed me to the post example at all, because this is primarily a music production site – hence the name. 😉 But it’s also a really important use case, as you say.

      If you look at their other examples, a lot of them explicitly show things like independent TV productions and the lot. There are a ton of potential users who simply can’t afford a dedicated person for the role; it’s not all Hollywood.

      But I hear what you’re saying. To me, the most interesting part of this for our audience is really people using this to be more creative or productive in their music work.

  • James

    I’ll 2nd flplsx’ comments and expand on a point of contention:
    I’m concerned as a sound person (production dialog) that the representation on the Izotope site is not really helping film-makers to appreciate the price of a technician who is capable of salvaging audio for a project.
    Each interviewee on the site swears that it is a time saver and project saver. What all those folks have in common is a regular, contracted means of employment in-house, with a show, or on A-list films. In other words, they can all absorb or account for their editing work in the asking price and are by default appreciated for their work. Albeit they are also probably subject to the same recession-pinching strategies that those lucky enough with work are facing, namely bosses who expect us to make up for the work of the 2 colleagues that were fired or face the notion that we too can be replaced. (Thankfully, I think making shows is a more recession-proof industry than most. Arguably a reverse trend even.)

    Full-budget films and broadcasted shows can then lean into ADR, and subsequently employ another set of professionals. But what of the 1st-time film maker, ignoring all signs of good measure who then realizes that the 10K that they raised on kickstarter was squandered because every scene they shot would be otherwise thrown out due to the corners they cut on sound. Isn’t it bad enough that sound people are probably not being employed by these well-intended, potentially ambitious and misguided projects?

    So what then is the cost of this mistake? Well it used to be ADR across the board or dumping your production HD’s at Tekserve’s recycling day. But now we are encouraging if not incentivizing a new wave of film makers who think these mistakes can be “fixed in post”, a phrase that used to be synonymous with “famous last words.” Essentially repeating the same mistakes that we’ve been trying to reform in the film-making process.

    So what’s the real story with audio restoring software? If it’s going to save your ass, what’s it’s worth in weight? Are we telling these same film makers that these mistakes have become less costly? I see the same thing happen with folks who don’t know how to slate. You sure your editor owns plural eyes? And have you really made less work for everyone or just “squeezed the tube of toothpaste in the middle” and dumped on your editor?

    Having worked extensively with RX, I can do wonderous things with the outcome. As far as I can tell, the methods in RX4 can be down and dirty and apparently solve things with a preset. For me, I haven’t noticed the out-of-box recipes to fix better than 70% of the problem, let’s quantify this with say 4db of restored headroom, or a baseline of 1000+ removed clipped peaks, crackles or thuds (which is a really silly measurement but indulge me). More typically running the Learn noise profile and pressing “process” will result in a lesser performance than actually knowing what every parameter does and knowing not just to use your ears but also knowing what to listen for. Mileage will vary.

    How about for a performance that would keep ADR at bay or genuinely salvage a scene/film? Well that requires an expert. (Not saying that anyone couldn’t become one, but that it takes work. I’m not trying to shroud this work in further mystique.) And if you want them to put their discerning ears to the test while working out the parameters of each tool while paying them the afterthought budget of your film project, then you might as well have hired child labor. Because the cost on the many hours of listening, A/Bing, and the like is exhausting work. You might as well go into the diamond business instead and hire little children to cut your gemstones what with their nimble little hands and unhampered 20/20 vision.

    I’ve finished sessions, cans off my head and away from the monitors-(ie. not listening anymore) where I literally thought a truck was backing up into my kitchen (enough to make me duck) while I’m drooling over my dinner , soiled from days straight of file management, listening, batch processing and the like. That’s how disruptive iscolating low-end, comb filtered, poorly grounded, or phased noise (and bad boundaries with an employer admittedly) can be. But if you need to rescue 6-12db for your mix that’s the real cost. Or you pay real people their real worth up front and make real mitigated decisions about how you are going to shoot.

    So why then has izotope sold this as another opportunity to dupe post-production workers into devaluing their contribution, or as additional confirmation that we can ignore the advice of the on-set mixer? That’s the message I get at least. “See, it’s easy!” “When I’m doing dialog, I can’t have anything slow me down. I need to get in there and get out.” Yadda yadda. You show a few successful cats who get a living wage based on their editing skills, and for the remainder of us who might consider audio restoration a specialized discipline we need to clarify for people that we’re singularly saving their ass and that’s worth something.

    Consider the ratio between a production schedule and the amount of time you intended to edit this thing, and then multiply what you should have paid your location sound guy by said factor. Then double this figure for the fact that a proper marketing schedule has already kicked in and you’ve already promised a parade of shareholders and contributors, not to mention an audience, that you’d have something to show for it.

    Audio restorers are still a life saver if you compare this to the cost of bringing in a named actor to the studio (who should be pissed at you for asking but is nonetheless a consummate professional about it. Day players on the other hand with thank you for your mistake.) or reshooting the scene/film again. But to think these problems go away for the price of software is a really an everyone-for-themselves marketing campaign.

    I actually think this area of craft should really form a guild.

    • Right, although this same argument can be used with mastering and mixing tools as well as post-production… or, really, *anything*, for that matter. My sense is that people either understand the value of skilled professionals or they don’t. Anyone leaning on the tool and saying “the devil made me do it,” as it were, likely was missing some experience and understanding in the first place.

      I don’t disagree with anything you’re saying, but you also observe that a low-to-no-budget DIY project is going to need to DIY its post needs, too – just like everything else. That’s not putting anyone out of work, because they didn’t have the resources to begin with. Sometimes those people might wind up building a career later.

      It’s tough – in software, you’re trying to market to as many people as you can. And in some sense, they’re just trying to show what the tool itself can do. I know that can include marketing to people who actually know what they’re doing.

      Anyway, I was a little surprised when iZotope pointed me to the post example at all, because this is primarily a music production site – hence the name. 😉 But it’s also a really important use case, as you say.

      If you look at their other examples, a lot of them explicitly show things like independent TV productions and the lot. There are a ton of potential users who simply can’t afford a dedicated person for the role; it’s not all Hollywood.

      But I hear what you’re saying. To me, the most interesting part of this for our audience is really people using this to be more creative or productive in their music work.