I just got this plugin for my MixPre 3. Absolutely love it. My office/studio doesn't have much sound treatment so there's a bit of natural reverb along with A/C noise. Running the NoiseAssist plugin at -5db is the perfect amount to lower the background noise and allows my voice to really punch through. $300 for the plugin is cheaper and far less work than treating my office. Love it.
@@curtisjudd Amazing bit of kit. Sometimes I wish I'd sprung for the 6 or 10 ii so that I'm future proofed. Really I'm using it primarily as an audio interface and to record talking head videos in my office. Even the 3 is probably a bit overkill. The vocal quality is noticeably cleaner than the scarlett I was running before. Thanks for responding Curtis.
I'm a documentary sound recordist and a lot of my work is recording interviews on location. I'm always fighting unwanted ambience and, given the nature of contemporary interior design trends, I'm almost always fighting reverberance. So I was happy to hear your kitchen scene, which seemed to exhibit a significant lessening of the the boominess. I'd love to hear the system deal with more reverberance. Many of my jobs have minimal resources for cleaning audio in post, so this would be helpful. But the biggest use case for me is in providing a fairly clean mix track for dailies. Many docs don't do any serious post on the sound until after they have locked picture, which can be a year down the road. I know I can convince a producer, when faced with a noisy or boomy space, that paying me $50 for the day will save them a year (and multiple rough cuts, work-in-progress screenings, show reels) of ugly audio. This may force me to abandon my 633 for interviews and go to a MixPre3 II just for the less expensive NoiseAssist.
Judging from this it sounds good up until 5dB most of the time. The asking price is simply FAR too high though (and correct me if I'm wrong... a bit crazy!). You can pick up the iZotope RX Elements bundle for like $50 on sale. But let's say you were on a job where you needed live noise reduction for dailies/director playback (bigger productions) you could rent/buy a Cedar unit and claim a rental cost for it, I doubt anyone would let you claim for using this plugin.
Thanks. I wouldn’t have any qualms about adding part of the cost of the plugin to a kit fee if the job required it. If they want live noise reduction, they have to pay for it.
Curtis, great video as always. I pretty much clicked away when I noticed that this NoiseAssist is specifically for this device. My hypothesis is that others would click away at the same time. I only mention this comment as a reference for your channel statistics and not as any sort of negative. Thanks for all that you do.
Curtis Judd - 👌 I meant this comment to be helpful and not critical. I do think the content is helpful for anyone considering these devices with Noise Assist. Thanks again for all of your videos. I really appreciate how much I’m able to learn from this channel. 👏
Thank you. As usual, helpful test and video... The price is out of my budget, but I think it will be useful for documentaries, where we don't have a full control of the environment/set. Sanken CS-3e/CS-m1 + NoiseAssist could be a Nice combination ^^
What a fantastic idea 💡 for the 8 series. I wish it was available for my 633. On a different note, I wish my RODECaster Pro had a similar option. In your lab test, I didn't notice significance until -10dB. Now your voice does seemed affected by the highway at -10dB. On the latency 1ms is incredible. Fantastic 👏 video testing on this noise assist!
This makes sense for fast turn-around "news" uses and live event streaming. But for more conventional productions with some time available in post, does it really make sense? One question I have is how does it interact with post processed noise reduction? Do weird things happen if you apply, say, 3db when recording, then realize it needs more noise reduction in post. Will odd effects result? As this is all happening late in the signal chain (after the pre-amp and the analog to digital converter) it can't help prevent clipping from loud noises or "shouting over" a noise source. Interesting that the processing power is arriving in such small units, but I'm not sure it would be helpful in most circumstances. (Though for some specific circumstances it will be fantastic.)
Interesting point, perhaps if you apply a low amount of NR and then you need to clean more in post, since the signal (in a perfect scenario) has a lower noise, might be simpler to remove, now i just need an 8 series, the plugin and a few minutes to try it out XD
I agree, not for every scenario. But I think it could be useful on the L or R bus for dailies on larger productions, so not just for live or fast turns. And good question about what happens when you apply more noise reduction in post. Will probably differ from one NR algorithm to another, but worth a test. Thanks!
I've tested NoiseAssist on MixPre II yesterday. A large diaphragm microphone gives the best results. I'm so happy that I can use a studio mic in almost any environment. Your voice sounds better on the Earthworks SR314 than any other mic you've used on this channel. It would de interesting to add the Earthworks SR314 and large diaphragm mics in a video about MixPre II NoiseAssist.
I think there are other AI background noise reductions which I have heard that appear to work a bit better but this is WORLDs better than “magic” cleanup usually sounds (ie final cut, audition), not enough for me to buy it. Wish it was an included firmware update, not paid. I think if it’s going to be paid, it better be a no brainer to be able to use it whenever needed without worrying it won’t sound great. Not feeling that here.
10:31 Ah, good to hear you have heard of the graphics card thing. Thank you for this demonstration, Curtis. Way out of my budget. Plus, I'd have to get those recorders to use it. ;D Have a great rest of the day!
I am belatedly learning about NoiseAssist and your tests indicate this is a practical investment for my type of work, Curtis. Much appreciated. If you happen to know or can guess: should I expect a slight drop in battery life on my MixPre-3 II if I use the plug-in more often than not? - Dave
Hi Dave, I haven't noticed a substantial difference but I use it on the 888, I don't have the plugin for my MixPre. And then again, I use big external batteries. Now that you mention it, yes, you should probably budget to use batteries more quickly.
I have the 833 here today - it features the CEDAR sdnx lowlatency noise reduction - and its unbelievable , nearly no sound artefacts even with -12dB , cant believe it. Worth it but very pricy... Perhaps you should do a review of it, too?
Thanks, Curtis. Excellent demo, as always! I can the value to this app for some applications. The outdoor attenuation is pretty rough, but the apps seems very useful for interior use. The 1ms latency with the SD Noise Assist is acceptable. I am curious to know how this would work with very soft voices (a common trend with overly dramatic actors these days) and sustained sounds from musical instruments and foley.
Thanks Dave. If there isn't a good signal to noise situation, it doesn't do well as you might imagine. It can also work decently outside - I just used ridiculous settings on that particular clip to show that you cannot push that far and expect good results. Works best with constant noise, not with music or inconsistent noise.
I think this is helpful for interviews in controlled environments, you might have a fan from a light source or a camera, some air-con, might work. Any chance you can get a Cedar DNS 2 and compare the results?
I think this could actually be pretty useful for certain types of sound effects recording in the field where noise is sort of a constant. Could have less of an effect on sound you are trying to record compared to recording voices. Overall pretty impressed with this and was enough of a motivator for me to upgrade from the MixPre 6 to the MixPre 6 II (along with 32bit recording).
If you have nvidia graphics card then you can try noise reduction for live streaming-or any desktop output called RTX Voice and it works in gtx card too with a minor tweak.
@@curtisjudd So you already tested it ? Nice, please do cover such topics as well in future videos.And also it doesn't seem to bother while livestreaming online as theres already Internet layency. Thanks !!!
For narrative pieces, I see this as useful for dailies by putting NoiseAssist on the L+R busses. Could also be useful for live to tape and documentaries as well. But for most scripted pieces, I would still want unprocessed isos for post.
Hi Curtis, is this plugin gonna take away the room reverb noise too? If yes, I can record in a non-treated room and far shorten the post production time! 😀
Thank you very much for the information that is very helpful. I hope you can compare the new plug-in CEDAR sdnx and noise assist to the sound device this time. There is a big difference in the amount between the two plug-ins, and I think CEDAR is better simply by looking at it, but if you look at this video alone, Noise Assist seems to be a plug-in with excellent cost-effectiveness.
Interestingly, the artifacting present by the highway with -10Db sounds pretty similar to the quality of bandwidth-limited sources like traditional POTS call-ins to a radio show. Agree with the other comment that this seems most useful for live, on the scene kind of scenarios. Seems pretty pricey, for what it is, but probably fits the market.
Is there a device out there that can help remove overall hiss and self noise while recording? And, what is the best way to determine noise frequencies your are trying to eliminate while recording? Trying so hard to reduce basic noise in my F6 and various mics like my Sennheiser 416, Octava, etc. I'm in a room with a lot of hard drives and can't really avoid that scenario.
I know the primary use case is different, but I wonder how this compares to the Waves NS1 Noise Suppressor plugin. The latency of that plugin is stated as zero additional samples, which in a live situation would be on top of a typical 0.8 ms system latency for a SuperRack or LV1 host. I assume the major value of having this on the recording device is that it allows for an easy way to get confidence immediately whether the take will be usable.
Hi Ronald, I think confidence is one use case. Dailies might be another by placing it on the L or R bus. Fast turn or live to tape pieces would be another.
Didn't know you'd purchased an 888! What made you go for the 888 over the 833? (or ha, at the other extreme.... a Scorpio!) Have you kept your 633? . (edit: ah, I went and rewatched your "Year in Review" video, guess you choose the 888 as you wanted more channels on your cart than the 10 channels an 833 could handle) . As for NoiseAssist, I'm currently not going to get this for my 833. Have got quite mixed feelings about this, as on one hand it is great to see advances in the development of the 8 Series platform, on the other hand it is disappointing to see this change in tack by Sound Devices, with them putting out highly priced plugins. Makes you worry about what this means for the future now that Sound Devices seemingly is a different company from how they acted in the past. . If I was to relive 2019 (ha, and who currently wouldn't want to swap the glorious idyllic days of 2019 for the 2020 nightmare we're living right now!) I'm no longer sure it would be such a slam dunk decision to choose the Sound Devices 833 over the Zaxcom Nova.
Yes, track counts and potential to use Dante in the future. The way I look at NoiseAssist is that it is a lot less expensive than Cedar DNS and far more convenient. I’m happy to add this to my kit fee for jobs where they need noise reduction. As for 833 vs Nova, that’s a bigger discussion. No doubt Nova is a compelling option but still showing as pre-order at Gotham Sound and to get all the benefits, a major re-investment of all my wireless. That’s be a big price tag.
@@curtisjudd I think it might just be showing as "pre-order" because they're out of stock? I think they're selling out very fast due to popularity, although my local store in NZ has it in stock. And I know for months now Nova has been shipping out to customers, many have their Nova. You're right that getting Zaxcom wireless as well means it is a big big total cost to fully get the most out of a Nova. However the new Zaxcom RX8 looks interesting, could be a way to help ease the transition over to the Zaxcom ecosystem before buying a Nova.
Isn't it better to just suggest people buy a standard Isotope RX7 version? I have never had the need for other 'sound repair' plugins after RX. Super high quality and - in case of sales - pretty affordable.
Hello from France, thank you for your work. I'm recording nature ambience sounds exclusively with a MKH30/MKH40 configuration (I do not record human voices or vocals). Often I need to raise the gain when I record in a forest in order to record birds and insects, which leads my microphones to produce more white noises. Would NoiseAssist fits this kind of use? Would it erase or ruin the nature sounds (birds, insects, water.. any kind of environmental sounds) the same way it erase the white noises?
Hi Izarra, Hmmm, is the "white noise" in the recordings or only in your headphones while recording? The Sennheiser mics generally do NOT produce a lot of self noise. And I don't think that using NoiseAssist will help with this case.
Hi Curtis, I’d love to see you do an episode on how to eliminate breath noises from voice over recordings. At present, I am manually cutting out each obnoxious breath for each hour or longer podcast episode....hours of work! I’m trying to understand whether it is better to try and eliminate noisy breathing using a noise gate as its being recorded, or whether to try and treat the noise later using a plugin? My breathing is quite noisy and so any help would be greatly appreciated. Thanks mate. Steve.
Ah, yes, I use the Izotope RX de-breath plugin which is sort of like a noise gate/expander but is substantially more transparent because it targets breaths, not just a threshold.
Also, the goal is not to eliminate breaths, just pull them down a bit so they aren’t distracting. When completely eliminated, the VO sounds a bit artificial.
Ditto with what Curtis said about reducing the breaths. Cut the heavy breaths but otherwise reduce them. It sounds strange when all the breath's are removed. Also, when you cut breaths, tighten the gap a smidge to help with flow - something no de-breath plugin will do.
Thank you for this. So valuable. My impression is that 10db noise attenuation sounded much more benign on your voice indoors than outdoors. I wonder if the adaptive noise cancelling algorithm also moves the threshold for coloring voices. My driving use case at the moment is to suppress summer air conditioning noise while presenting at meetings. The expander strategy works reasonably well -- feedback is I sound fine and noise the air conditioning is not an issue. Nevertheless, use of NoiseAssist on my MixPre II would be an improvement. Here is a marketing tip for SoundDevices - I would buy NoiseAssist for $99, but for $300 I am giving it a pass.
@@curtisjudd Okay cool thank you. I think Izotope more suits my needs then. Although it would be nice to have it in the field. I do not think the price reflects the use case for me. Maybe if they have a sale one day.
Hi, thanks for your smart review. I'd have a question... I have a MixPre-6 II and I see that NoiseAssist is available for my device too. Anyway, before investing 300$ on it, I'd like to understand if it's useful for my needs: I usually use MixPre-6 II for recording ambient sounds and doing a few foley. No dialogue, sometimes voices to be used as background sounds. Often the sounds I capture are relatively far from me (birds, crickets) and “tiny” sounds outdoors. Unfortunately, when I record these sounds I can also hear traffic noise (even if the street is relatively far from me). Up to now I tried to reduce traffic noise in post with a dedicated plug in a DAW. The question is: do you think NoiseAssist could match my needs, or since I don't do dialogue things I should countinue with classic NR plug in post? Thanks in advance.
I would only use NoiseAssist for situations where I cannot do post production or very little post production. For Foley and other effects recordings there will almost always be a post workflow and in that case I would not risk adding artefacts to the audio. You generally also have more control during effects recordings to turn off or address noise you don't want - wait, add sound blankets, turn off compressors/HVAC.
@@curtisjudd Thanks for your impressively quick reply. Yes, during foley all can be more under control. Honestly in ambient is a little more difficult because actually I can't avoid traffic background. But perhaps, as you said, it's better to go for post (having time). Thanks again.
I haven't used it since I did these original tests. But if I only applied 2 or 3 dB of attenuation, I didn't notice a lot of coloration. But with more than that, yes, it starts to affect the desired sound.
I listened with interest, but slightly disappointed. I listened on my nokia phone over AKG K271, so far from perfect (phone won't drive those headphones very loud) but still. Even at the lower settings I think I can hear noise pumping at voice onset and end. Sounds a bit like a softly setup noise gate....
I can still hear the noise pumping, and I'm pretty sure the noise is not very well suppressed behind your voice when you're talking. Haven't got instruments to actually measure noise background behind your monologue.
They just released this for the mixpre generation 2’s. And for 300 bucks as opposed to 600. Hey, does it take out any of the bass in your voice? I’m wondering if this has anything to do with low cut filters...
I think they need to move the decimal place over on that price... ouch... I understand that it is all about what the market will bare, but still... ouch!
If I received a track with this on it I would ask... Why? The trick mixing Dialog is to do a little amount of processing with a few tools instead of 1 with a lot of processing. Also remember that in Post Sound we also are interacting with ADR , Group , Music, SFX, Foley , Sound Design etc. When you listen to a Dialog track by itself on the set on a set of headphones you might think you are being "Helpful". On "The Dub Stage" we are listening on extremely high quality Meyer Acheron speakers that reveal everything , even over processing on a Production track. Please don't ever use this without sending a sample of a clip that you used this on an wait for your Dialog Mixer to give you their observations and a go ahead. Personally don't ever go past 5 (better yet 3) . If this is a situation where there is little or no Post Sound then follow my advice for Noise Reduction.. ........Listen to it and when you think you have it right back off 25%. Then let me do the rest in the mix!
Pretty useful for live or no-post workflows. No video in the works at present, but hoping to get a copy to confirm it works the same as on the 8xx series.
I think it is fantastic for non-production work and potentially for post work. But for on set or in the field, it isn't practical due to the power requirements and 50+ms latency.
I wish you had done a test where you'd first do the things one would normally do before recording stuff at home - i.e. turn the kitchen fan off instead of on, and keep the kitchen window closed rather than open. And putting a highway behind you is not the best possible place to test how much rumble there will remain on the track on most recording situations. It is a bit like you went to an airfield to test a hearing aid....
@@curtisjudd Hi Curtis, I did not mean to say should there shouldn't be *any* noise. That'd be an insane thing to do, and I have this principle of trying to avoid doing insane stuff. I was merely pointing out that putting the fan *on* instead of off (etc.) is not a good move, if you try to find out how much the noise reduction system would help in a *typical* home environment interview. Take an interview or recording I would make in my home environment: I would have the kitchen fan off. I'd turn a noisy PC, and room fans off. I'd close the windows. But I still would have the ubiquitous background hum / low rumble of lorries, and the annoying higher-pitched sound of mopeds, the rumble & shaking and clinking of glasses in the cupboards when a train passes as teh house, I live just a hundred yards away from the tracks. And if I'm really lucky the rumble of my stomach and that of my freezer (which I often forget to unplug) would be there to sabotage the recording . Outdoors, then. An outdoors recording I'd do in a park, not on a field near a motorway. There would be car bird noises, an occasional plane, but it would not sound like being insine a turbine engine.. In my case - real, not theoretical - the dealbreaker on NoiseAssist would is the the amount of help it would give me while recording at somewhat acoustically softened home environment where the most obvious and easily removable noise sources would have been dealt with. -And FWIW, please don't regard my comment as a put-down. I find your videos extremely helpful, well done, and enjoyable. Very high quality stuff, your content. Thank you for it! MK
@@curtisjudd You two! And a new finding in my world of recording: My SD Mixpre-6 II seems to be allergic to non-Neutrik XLR plugs. It gets them stuck. Had to have a pro to take it out. it seems cheaper XLR cabler do not have up-to-standard plugs - and/or Sound Devices' Mixpre-6 II is too fastidious for the *real* prosumer world.
@@markuskajo903 Yes, I decided quite some time ago to spend the extra money for cables with Neutrik connectors. It isn't just Sound Devices gear where cheaper connectors get stuck. There are seemingly hundreds of people complaining about the same thing with the Tascam recorders.
Hello Chris. Can you compare new Noiseassist for Mixpre generations recorders vs 8xx series? Why is half price ? Of course, if you have it available. Thanks 👍
Near the highway, when u went to 10db your voice started to sound like on the telephone. Was wondering if you could affect the voice back to normal while keeping 10db?
@@curtisjudd which brings to mind another question - is it sometimes okay to process the voice in real time and with mixpre and the effects app, have processed vocals on a separate track.?
Curtis maybe do a demo of the Absentia DX plugin made by Todd AO it can be used on your files and keeps the metadata or also be used as a plugin and it’s cheap
Yes, but that's for post. This is for production cases where you cannot do post - fast turn-around news - or live. Definitely some great options in the post world. Will check out Absentia. Thanks!
I could hear difference on my ISK HD9999 headphones at 8dB. At 12dB it started to be unacceptable for me. But since you are not talking the same thing all the time without the slightest brake, it is possible that the difference could came sooner.
@@curtisjudd OK. I must have been distracted. I am not used to hearing your voice through lav. I can definately hear something strange at 6dB in the first trial. In phrase "difference in terms". Some metalic nasality. Like highs and lows were missing. But it was noticeable the moment you moved. So... Outdoors it was comical. Not only it coloured your voice as if you were talking through phone but it also attenuated it. It clearly treated your voice as noise. Traffic noise seems to be to unstable for such an algorythm. I must notice that ISK HD9999 has very smiley tone (boosting highs an lows). Very musical. Not necessarily accurate.
@@przybylskipawel Ah yes, smiley voicing. I think for the traffic sample, the setting needs to be a lot closer to -2 or -3dB to avoid the issues we encountered.
It sounds horrible starting at -5 dB. I’ll stick to using the Noise Reduction effect in Audacity. It allows me to take a sample of the noise first, which it then uses to reduce just the noise frequencies. And since it does not affect the original recording, I can adjust it as needed until I get it right.
Unless the Audacity NR effect can work in real time, it's apples to pears. And it definitely was horrible around -5dB. I was listening through very modest JBL LSR2300 monitors, and the sound became irritating at that point for the Sennheiser in the kitchen. My reaction was to jerk my head somewhat to the side and downward to redirect my ears.
Understood. I think that I'd keep unprocessed isos and just use this on one of the busses for fast turn-around or live work. Could also be useful on the mix bus for dailies on bigger productions. But yes, post will still want clean, unprocessed isos in most cases.
At -9 dB your voice sounded muffled for sure... And at -15 dB, you were talking through a dark chocolate muffin :D If SoundDevices gave you a truly fabulous EQ, it could go further!
Hahaha! I agree, unless one is recording on a dynamic microphone about 1-2 inches from the mouth, about 4dB of attenuation is roughly the limit. And now I'm hungry for a dark chocolate muffin.
Yes, if I happen upon a DNS2 I can use for the demonstration. I don't have any rental houses here in my city so that'll be tough, but it would make an interesting comparison some day. DNS2 would need to be substantially better for me to consider adding it to a kit with the extra bulk and powering it would require.
tomorrow need to record one actor unfortunately the location is unsuccessful historical film but a highway runs nearby :( I thought cheers there is a solution, but unfortunately after 3 decibels the sound goes bad :(( thanks for the interesting video :)
Interesting. I'm sure you probably know all about software based noise reduction developments and I don't know how they compare in practical use but I was impressed by this BBC clip of software being used to reduce other noise...quite impressive as to where things are going. I wonder what you think of this, ongoing, development. BBC clip: czcams.com/video/BRg7LHAJqeU/video.html It's pushed to the limit but if not over taxed with ambient noise? Interesting to see the progress being made.
Definitely interesting to see the advancements. I think they have different use cases - the graphics card and computer software based have a lot of promise for post production but not practical for production like NoiseAssist. Great to hear all the options.
I’m guessing this NoiseAssist “plugin” would have been far more effective if it employed a LEARN function. Seems like a missed opportunity. Maybe SD can release and update before the competition does.
Hey @@curtisjudd!, thanks for the reply. Big fan here. Just want to add that SoundDevices are not going to secure any patents with that level of detail, hehe. I’m sure they’re using an algorithm similar to the dynamic noise detection employed in popular DAW plugins, which can be quite effective. And they do use FPGAs, which means they can update the code in firmware if they wanted to employ a user-defined noise print ( “learning” ) feature, which would generally be more effective than using an algorithm to derive a noise print. Combining both methods would be even better!
If your workflow includes post, yes, this is not meant to replace that. But if mixing live or something with fast turn around, then this becomes useful.
There are significant differences. NoiseAssist is made for field recorders which are often battery powered and designed for use on film sets and locations. It'd be a different game to record a film with a PC. Also, this operates with 1ms of latency vs. 50+ms of latency on the graphics card. With film, 50ms is far too much latency. So while this will not eliminate or even reduce all noise, it does have some advantages to graphics card AI noise reduction. Also, graphics cards are not free whether they're RTX or not. 🙂
@@curtisjudd but for film isn't it better to modify the sound afterwards ? Then you can take samples of noises and apply a reverse Fourrier transform to remove noise frequencies. You can do that with any too like Audacity and it works very well for hiss and most cyclic noises. Is there any professional tool to graphically pinpoint noise peaks and allow a sound designer to clean them (like Magix Audio Cleanic for Vinyl discs, but on steroids) ? I don't understand the utility of NoiseAssist out of the monitoring purpose (like you said in the outro of this video it can helps to do fine synchronization).
@@sitedel A Dailies workflow is one case where I might want this just on the mix bus - I'd still want unprocessed isolated channels of each microphone for post. Corporate and new fast-turnaround is another - any case where there isn't time for post is where this has value. I've even done a livestream job recently where this would have been helpful.
Curtis' sophisticated, methodical style is a HILARIOUS contrast with the recent videos about NVidia RTX Voice. See it: czcams.com/video/HA6R9NmZO80/video.html
Hahaha! Agreed, I am definitely not producing videos for the audience which found that to be a useful test. And this product is also not for the same folks or use cases. 🤓
Bad timing. Nvidia RTX Voice does this way better and it's free. I've been using it to clean up existing audio and it's almost magical how well it works.
RTX is cool but it has more than 50ms latency and is power hungry - well suited to gaming. NoiseAssist is 1ms latency and requires low power - much better suited to live professional production and on location.
Agreed, not bad timing. Completely different use case. RTX voice burns power which is usually sourced from batteries on film sets and on location. That 50+ms latency is also problematic in film production. And I don't have room for a PC on my sound cart. 🤓
@@curtisjudd I've been using RTX Voice in post production to clean up audio I recorded on set. It's not the intended use case, but it's been working amazingly well so far. Doing it that way negates the power consumption and latency issues. Well, the latency isn't completely negated as it does and in the step of re-syncing the audio but most good editing software seems to make that pretty easy. This NoiseAssist does seem like a much simpler solution, but given it's high cost and lower effectiveness than RTX Voice I still can't see myself considering it.
@@WrongThemeSong Thanks for the additional info. Yes, definitely looks like an interesting new option for post. I'm not trying to sell NoiseAssist, just explaining the intended use case and how it differs from graphics card NR. When I'm hired for a freelance location sound job, NoiseAssist is a better fit over graphics card or other hardware based solutions for me since I already own the 888. The NoiseAssist plugin is way cheaper than other hardware solutions typically used in location and on-set for noise reduction.
Hi Petr, I haven't used Cedar yet, but it uses a different algorithm to accomplish the same general goal. If the comparison is anything like the difference between MixAssist and Dugan Automixing, they each have their strengths and weaknesses - different situations which benefit from each. MixAssist, in my experience is a little more aggressive about reducing noise whereas Dugan is more transparent.
@@curtisjudd i was trying demo of cedar and noise assist indoor with mic on close distance and in noise assist i hear some kind of dullness in voice while in cedar i dont hear it affect the voice so much and even the residual ambient noise sounds very clean.
This is the only channel that I pause to get my headphones on
👍
I just got this plugin for my MixPre 3. Absolutely love it. My office/studio doesn't have much sound treatment so there's a bit of natural reverb along with A/C noise. Running the NoiseAssist plugin at -5db is the perfect amount to lower the background noise and allows my voice to really punch through. $300 for the plugin is cheaper and far less work than treating my office. Love it.
Glad to hear it Ron!
@@curtisjudd Amazing bit of kit. Sometimes I wish I'd sprung for the 6 or 10 ii so that I'm future proofed. Really I'm using it primarily as an audio interface and to record talking head videos in my office. Even the 3 is probably a bit overkill. The vocal quality is noticeably cleaner than the scarlett I was running before. Thanks for responding Curtis.
I'm a documentary sound recordist and a lot of my work is recording interviews on location. I'm always fighting unwanted ambience and, given the nature of contemporary interior design trends, I'm almost always fighting reverberance. So I was happy to hear your kitchen scene, which seemed to exhibit a significant lessening of the the boominess. I'd love to hear the system deal with more reverberance. Many of my jobs have minimal resources for cleaning audio in post, so this would be helpful. But the biggest use case for me is in providing a fairly clean mix track for dailies. Many docs don't do any serious post on the sound until after they have locked picture, which can be a year down the road. I know I can convince a producer, when faced with a noisy or boomy space, that paying me $50 for the day will save them a year (and multiple rough cuts, work-in-progress screenings, show reels) of ugly audio. This may force me to abandon my 633 for interviews and go to a MixPre3 II just for the less expensive NoiseAssist.
Thanks Doug, sounds like a good use case for NoiseAssist.
Honestly pretty incredible that we're at a point where we can apply noise reduction during location recording. Very neat feature
👍
Judging from this it sounds good up until 5dB most of the time. The asking price is simply FAR too high though (and correct me if I'm wrong... a bit crazy!). You can pick up the iZotope RX Elements bundle for like $50 on sale. But let's say you were on a job where you needed live noise reduction for dailies/director playback (bigger productions) you could rent/buy a Cedar unit and claim a rental cost for it, I doubt anyone would let you claim for using this plugin.
Thanks. I wouldn’t have any qualms about adding part of the cost of the plugin to a kit fee if the job required it. If they want live noise reduction, they have to pay for it.
Yes, the price is too high for the results you get.
Curtis, great video as always. I pretty much clicked away when I noticed that this NoiseAssist is specifically for this device. My hypothesis is that others would click away at the same time. I only mention this comment as a reference for your channel statistics and not as any sort of negative. Thanks for all that you do.
Thanks Matt. Understood that this is not for everyone. Have a great weekend!
Curtis Judd - 👌 I meant this comment to be helpful and not critical. I do think the content is helpful for anyone considering these devices with Noise Assist. Thanks again for all of your videos. I really appreciate how much I’m able to learn from this channel. 👏
@@mlegrand Not taken as critical at all. Thanks Matt.
Nice job on the video, Curtis. Thorough demonstration 👍
Thanks Jim.
Thank you. As usual, helpful test and video... The price is out of my budget, but I think it will be useful for documentaries, where we don't have a full control of the environment/set.
Sanken CS-3e/CS-m1 + NoiseAssist could be a Nice combination ^^
Good point!
What a fantastic idea 💡 for the 8 series. I wish it was available for my 633. On a different note, I wish my RODECaster Pro had a similar option.
In your lab test, I didn't notice significance until -10dB. Now your voice does seemed affected by the highway at -10dB. On the latency 1ms is incredible.
Fantastic 👏 video testing on this noise assist!
Thanks Geek Homeworld!
This makes sense for fast turn-around "news" uses and live event streaming. But for more conventional productions with some time available in post, does it really make sense? One question I have is how does it interact with post processed noise reduction? Do weird things happen if you apply, say, 3db when recording, then realize it needs more noise reduction in post. Will odd effects result? As this is all happening late in the signal chain (after the pre-amp and the analog to digital converter) it can't help prevent clipping from loud noises or "shouting over" a noise source. Interesting that the processing power is arriving in such small units, but I'm not sure it would be helpful in most circumstances. (Though for some specific circumstances it will be fantastic.)
Interesting point, perhaps if you apply a low amount of NR and then you need to clean more in post, since the signal (in a perfect scenario) has a lower noise, might be simpler to remove, now i just need an 8 series, the plugin and a few minutes to try it out XD
I agree, not for every scenario. But I think it could be useful on the L or R bus for dailies on larger productions, so not just for live or fast turns. And good question about what happens when you apply more noise reduction in post. Will probably differ from one NR algorithm to another, but worth a test. Thanks!
👍
I've tested NoiseAssist on MixPre II yesterday. A large diaphragm microphone gives the best results. I'm so happy that I can use a studio mic in almost any environment. Your voice sounds better on the Earthworks SR314 than any other mic you've used on this channel. It would de interesting to add the Earthworks SR314 and large diaphragm mics in a video about MixPre II NoiseAssist.
Thanks for the update. Interesting that it works best with LDCs...
Noise assist is available on mixpre models as well?
-3dB is enough for me but I think I can go up to -6dB with no problem. Thank you so much for this.
👍
It started noticeably affecting the voice at -6db to my ears.
Agreed. And I believe your ears. 👍
I think there are other AI background noise reductions which I have heard that appear to work a bit better but this is WORLDs better than “magic” cleanup usually sounds (ie final cut, audition), not enough for me to buy it. Wish it was an included firmware update, not paid. I think if it’s going to be paid, it better be a no brainer to be able to use it whenever needed without worrying it won’t sound great. Not feeling that here.
10:31 Ah, good to hear you have heard of the graphics card thing.
Thank you for this demonstration, Curtis.
Way out of my budget. Plus, I'd have to get those recorders to use it. ;D
Have a great rest of the day!
You too, huyked!
I am belatedly learning about NoiseAssist and your tests indicate this is a practical investment for my type of work, Curtis. Much appreciated. If you happen to know or can guess: should I expect a slight drop in battery life on my MixPre-3 II if I use the plug-in more often than not? - Dave
Hi Dave, I haven't noticed a substantial difference but I use it on the 888, I don't have the plugin for my MixPre. And then again, I use big external batteries. Now that you mention it, yes, you should probably budget to use batteries more quickly.
Thanks, that was really well explained and demonstrated.
Thanks Sean.
Mix Pre Support Now 🙌🏼. It’s not cheap , well all win :)
👍
In the kitchen 5-6dB was obvious difference. Unpleasant at moments. I guess that depend on your movement relative to the mic.
Yes, definitely. I'd probably go with -3dB or so in that situation.
I have the 833 here today - it features the CEDAR sdnx lowlatency noise reduction - and its unbelievable , nearly no sound artefacts even with -12dB , cant believe it. Worth it but very pricy...
Perhaps you should do a review of it, too?
👍
Thanks, Curtis. Excellent demo, as always! I can the value to this app for some applications. The outdoor attenuation is pretty rough, but the apps seems very useful for interior use. The 1ms latency with the SD Noise Assist is acceptable. I am curious to know how this would work with very soft voices (a common trend with overly dramatic actors these days) and sustained sounds from musical instruments and foley.
Thanks Dave. If there isn't a good signal to noise situation, it doesn't do well as you might imagine. It can also work decently outside - I just used ridiculous settings on that particular clip to show that you cannot push that far and expect good results. Works best with constant noise, not with music or inconsistent noise.
I think this is helpful for interviews in controlled environments, you might have a fan from a light source or a camera, some air-con, might work. Any chance you can get a Cedar DNS 2 and compare the results?
Thanks Carlos. Will see what we can do. they're quite different in price.
Yes, I was thinking the same. Maybe Noise Assist can be compared to DNS2??
I think this could actually be pretty useful for certain types of sound effects recording in the field where noise is sort of a constant. Could have less of an effect on sound you are trying to record compared to recording voices. Overall pretty impressed with this and was enough of a motivator for me to upgrade from the MixPre 6 to the MixPre 6 II (along with 32bit recording).
Congrats on the new recorder! Let us know how it works out for efx recording!
If you have nvidia graphics card then you can try noise reduction for live streaming-or any desktop output called RTX Voice and it works in gtx card too with a minor tweak.
Thank you, yes. Just be prepared for a 50ms latency which will affect sound to video sync.
@@curtisjudd So you already tested it ? Nice, please do cover such topics as well in future videos.And also it doesn't seem to bother while livestreaming online as theres already Internet layency. Thanks !!!
Not for the MixPre3? What the heck!! (joking). Really cool feature if dialed in. I imagine it's voice dependent.
Hahaha! Nice to have another tool in the toolbox for when you need it.
There is one for MixPre: store.sounddevices.com/product/noiseassist-for-mixpre/
Is this preferable to post processing, in your opinion?
For narrative pieces, I see this as useful for dailies by putting NoiseAssist on the L+R busses. Could also be useful for live to tape and documentaries as well. But for most scripted pieces, I would still want unprocessed isos for post.
Hi Curtis, is this plugin gonna take away the room reverb noise too? If yes, I can record in a non-treated room and far shorten the post production time! 😀
It will not remove reverb, but it will help. Temporary blankets will do a better job of that.
Works well. Shame I can’t get this on my 633. Thanks!
👍
Thank you very much for the information that is very helpful.
I hope you can compare the new plug-in CEDAR sdnx and noise assist to the sound device this time. There is a big difference in the amount between the two plug-ins, and I think CEDAR is better simply by looking at it, but if you look at this video alone, Noise Assist seems to be a plug-in with excellent cost-effectiveness.
I hope so too. Can't justify the expense at the moment, but hopefully soon.
@@curtisjudd It's okay if it doesn't work quickly. I can trust and enjoy your CZcams videos.😄😄😄
@@curtisjudd There is a cedar sdnx demo mode for 8-Series running v7.40 or higher, have you try it ?
@@alanwoo6960 I haven't, good to know. Thanks.
I feel like it reduced some of the reverb in the kitchen. Is it me?
I think it did as well.
From 0 to -4db everything sounds pretty much the same, but -4 to -5 db has noticeable difference in coloring your voice.
👍
Interestingly, the artifacting present by the highway with -10Db sounds pretty similar to the quality of bandwidth-limited sources like traditional POTS call-ins to a radio show. Agree with the other comment that this seems most useful for live, on the scene kind of scenarios. Seems pretty pricey, for what it is, but probably fits the market.
Thanks for the insights Jason.
Thanks, very useful demo.
Thanks Michael.
Is there a device out there that can help remove overall hiss and self noise while recording? And, what is the best way to determine noise frequencies your are trying to eliminate while recording? Trying so hard to reduce basic noise in my F6 and various mics like my Sennheiser 416, Octava, etc. I'm in a room with a lot of hard drives and can't really avoid that scenario.
Yes, Sound Devices MixPre with NoiseAssist plugin or Cedar DNS (pricey).
Dear Curtis, what kind of carrying strap do you use (the one near the highway)? Thank you for your answer in advance. Thank you for your nice videos.
Hi Andrius, that is the Orca harness for the Orca sound bags: www.bhphotovideo.com/c/product/1082117-REG/orca_or_40_audio_bag_harness.html
I know the primary use case is different, but I wonder how this compares to the Waves NS1 Noise Suppressor plugin. The latency of that plugin is stated as zero additional samples, which in a live situation would be on top of a typical 0.8 ms system latency for a SuperRack or LV1 host.
I assume the major value of having this on the recording device is that it allows for an easy way to get confidence immediately whether the take will be usable.
Hi Ronald, I think confidence is one use case. Dailies might be another by placing it on the L or R bus. Fast turn or live to tape pieces would be another.
Didn't know you'd purchased an 888! What made you go for the 888 over the 833? (or ha, at the other extreme.... a Scorpio!) Have you kept your 633?
.
(edit: ah, I went and rewatched your "Year in Review" video, guess you choose the 888 as you wanted more channels on your cart than the 10 channels an 833 could handle)
.
As for NoiseAssist, I'm currently not going to get this for my 833. Have got quite mixed feelings about this, as on one hand it is great to see advances in the development of the 8 Series platform, on the other hand it is disappointing to see this change in tack by Sound Devices, with them putting out highly priced plugins. Makes you worry about what this means for the future now that Sound Devices seemingly is a different company from how they acted in the past.
.
If I was to relive 2019 (ha, and who currently wouldn't want to swap the glorious idyllic days of 2019 for the 2020 nightmare we're living right now!) I'm no longer sure it would be such a slam dunk decision to choose the Sound Devices 833 over the Zaxcom Nova.
Yes, track counts and potential to use Dante in the future. The way I look at NoiseAssist is that it is a lot less expensive than Cedar DNS and far more convenient. I’m happy to add this to my kit fee for jobs where they need noise reduction. As for 833 vs Nova, that’s a bigger discussion. No doubt Nova is a compelling option but still showing as pre-order at Gotham Sound and to get all the benefits, a major re-investment of all my wireless. That’s be a big price tag.
@@curtisjudd I think it might just be showing as "pre-order" because they're out of stock? I think they're selling out very fast due to popularity, although my local store in NZ has it in stock. And I know for months now Nova has been shipping out to customers, many have their Nova.
You're right that getting Zaxcom wireless as well means it is a big big total cost to fully get the most out of a Nova. However the new Zaxcom RX8 looks interesting, could be a way to help ease the transition over to the Zaxcom ecosystem before buying a Nova.
Isn't it better to just suggest people buy a standard Isotope RX7 version? I have never had the need for other 'sound repair' plugins after RX. Super high quality and - in case of sales - pretty affordable.
Hi Nic, yes, for post. But for live, this makes more sense.
Hello from France, thank you for your work. I'm recording nature ambience sounds exclusively with a MKH30/MKH40 configuration (I do not record human voices or vocals). Often I need to raise the gain when I record in a forest in order to record birds and insects, which leads my microphones to produce more white noises. Would NoiseAssist fits this kind of use? Would it erase or ruin the nature sounds (birds, insects, water.. any kind of environmental sounds) the same way it erase the white noises?
Hi Izarra, Hmmm, is the "white noise" in the recordings or only in your headphones while recording? The Sennheiser mics generally do NOT produce a lot of self noise. And I don't think that using NoiseAssist will help with this case.
Hi Curtis, I’d love to see you do an episode on how to eliminate breath noises from voice over recordings. At present, I am manually cutting out each obnoxious breath for each hour or longer podcast episode....hours of work! I’m trying to understand whether it is better to try and eliminate noisy breathing using a noise gate as its being recorded, or whether to try and treat the noise later using a plugin? My breathing is quite noisy and so any help would be greatly appreciated. Thanks mate. Steve.
Ah, yes, I use the Izotope RX de-breath plugin which is sort of like a noise gate/expander but is substantially more transparent because it targets breaths, not just a threshold.
Also, the goal is not to eliminate breaths, just pull them down a bit so they aren’t distracting. When completely eliminated, the VO sounds a bit artificial.
Curtis Judd thanks so much for that. I’ve seen reviews on that plugin so it’s reassuring that you are using it. Thanks mate.
Ditto with what Curtis said about reducing the breaths. Cut the heavy breaths but otherwise reduce them. It sounds strange when all the breath's are removed. Also, when you cut breaths, tighten the gap a smidge to help with flow - something no de-breath plugin will do.
Thank you for this. So valuable. My impression is that 10db noise attenuation sounded much more benign on your voice indoors than outdoors. I wonder if the adaptive noise cancelling algorithm also moves the threshold for coloring voices.
My driving use case at the moment is to suppress summer air conditioning noise while presenting at meetings. The expander strategy works reasonably well -- feedback is I sound fine and noise the air conditioning is not an issue. Nevertheless, use of NoiseAssist on my MixPre II would be an improvement. Here is a marketing tip for SoundDevices - I would buy NoiseAssist for $99, but for $300 I am giving it a pass.
If the expander is working, then sounds like that's the tool for this particular job.
SOLD!
It is quite useful in the right circumstances.
3-5db is fine, even more voice is more distort and coloring.
Agreed, for these samples.
Curtis would you recommend this or Izotope RX? They are the same price
Depends. Izotope RX is more versatile, but isn’t built in to a recorder to do the noise reduction in real-time such as when live-streaming.
@@curtisjudd Okay cool thank you. I think Izotope more suits my needs then. Although it would be nice to have it in the field. I do not think the price reflects the use case for me. Maybe if they have a sale one day.
From 6db it starts sounding too much for me . I do post audio
Agreed on the -6dB setting. NoiseAssist is for cases where you cannot do post production. Otherwise, yes, I agree. I'd do it in post as well.
Hi, thanks for your smart review. I'd have a question... I have a MixPre-6 II and I see that NoiseAssist is available for my device too. Anyway, before investing 300$ on it, I'd like to understand if it's useful for my needs: I usually use MixPre-6 II for recording ambient sounds and doing a few foley. No dialogue, sometimes voices to be used as background sounds. Often the sounds I capture are relatively far from me (birds, crickets) and “tiny” sounds outdoors. Unfortunately, when I record these sounds I can also hear traffic noise (even if the street is relatively far from me). Up to now I tried to reduce traffic noise in post with a dedicated plug in a DAW. The question is: do you think NoiseAssist could match my needs, or since I don't do dialogue things I should countinue with classic NR plug in post? Thanks in advance.
I would only use NoiseAssist for situations where I cannot do post production or very little post production. For Foley and other effects recordings there will almost always be a post workflow and in that case I would not risk adding artefacts to the audio. You generally also have more control during effects recordings to turn off or address noise you don't want - wait, add sound blankets, turn off compressors/HVAC.
@@curtisjudd Thanks for your impressively quick reply. Yes, during foley all can be more under control. Honestly in ambient is a little more difficult because actually I can't avoid traffic background. But perhaps, as you said, it's better to go for post (having time). Thanks again.
Curtis, do you notice that the NoiseAssist makes your S's sharper - or adds sibilance?
I haven't used it since I did these original tests. But if I only applied 2 or 3 dB of attenuation, I didn't notice a lot of coloration. But with more than that, yes, it starts to affect the desired sound.
I listened with interest, but slightly disappointed.
I listened on my nokia phone over AKG K271, so far from perfect (phone won't drive those headphones very loud) but still.
Even at the lower settings I think I can hear noise pumping at voice onset and end. Sounds a bit like a softly setup noise gate....
Thanks for the feedback Hans. Interested to know your thoughts when using a proper headphone amp.
I can still hear the noise pumping, and I'm pretty sure the noise is not very well suppressed behind your voice when you're talking.
Haven't got instruments to actually measure noise background behind your monologue.
They just released this for the mixpre generation 2’s. And for 300 bucks as opposed to 600. Hey, does it take out any of the bass in your voice? I’m wondering if this has anything to do with low cut filters...
It sounds to me like it takes out various frequencies, not just bass.
Is it worth the money to get this for my MixPre 6ii? Have you found the 8-series version helpful?
I don't use it a lot for my work. I think it will be most useful for live situations.
I think they need to move the decimal place over on that price... ouch... I understand that it is all about what the market will bare, but still... ouch!
Understood. It is also about the time invested in developing this and relative to the other main options on the market, this is a lot less expensive.
Thank you for doing this Curtis. I think this sounds horrible and noticeable from 1dB. Saved me some money!
Ok, best wishes!
SD page says only one instance per serial number. Do I have to tell them I’m Curtis Judd to get two or did SD change that since you made this?
One instance on MixPre, two instances on 8xx series. 😀 You just need an 8xx series.
will u be updating this video with Noise Assist being used on a MixPre II?
I hope so! Gotta save up $300 first. :)
@@curtisjudd or just reach out to them and get it free? :)
@@Seanonyoutube I am told it is the same on the MixPre except applies to just one channel so probably not going to do a separate video on it.
It sounds to me like there is some gating taking place in the NR algorithm, especially noticeable during the highway traffic test.
They have an article covering how it works: www.sounddevices.com/noiseassist-simplified/
@@curtisjudd you don’t hear gating/compression? It sounds like it to me.
@@aerofart Oh, yes, in that sample by the highway I heard some seriously bad flapping.
If I received a track with this on it I would ask... Why? The trick mixing Dialog is to do a little amount of processing with a few tools instead of 1 with a lot of processing. Also remember that in Post Sound we also are interacting with ADR , Group , Music, SFX, Foley , Sound Design etc. When you listen to a Dialog track by itself on the set on a set of headphones you might think you are being "Helpful". On "The Dub Stage" we are listening on extremely high quality Meyer Acheron speakers that reveal everything , even over processing on a Production track. Please don't ever use this without sending a sample of a clip that you used this on an wait for your Dialog Mixer to give you their observations and a go ahead. Personally don't ever go past 5 (better yet 3) . If this is a situation where there is little or no Post Sound then follow my advice for Noise Reduction.. ........Listen to it and when you think you have it right back off 25%. Then let me do the rest in the mix!
Thanks Marti, point taken.
Available for all MixPre II recorders at $300
Yes, and for a single channel. Thanks.
Whoa now available for MixPre II series for $300. What are your thoughts on that Curtis? Or will there be a new video in the works?
Pretty useful for live or no-post workflows. No video in the works at present, but hoping to get a copy to confirm it works the same as on the 8xx series.
What do you think about noise reduction with artificial intelligence in the new graphics cards?
I think it is fantastic for non-production work and potentially for post work. But for on set or in the field, it isn't practical due to the power requirements and 50+ms latency.
I wish you had done a test where you'd first do the things one would normally do before recording stuff at home - i.e. turn the kitchen fan off instead of on, and keep the kitchen window closed rather than open.
And putting a highway behind you is not the best possible place to test how much rumble there will remain on the track on most recording situations. It is a bit like you went to an airfield to test a hearing aid....
Huh? The purpose of the test is to see how it handles noise. If there isn't any noise to assist in removing, how is it a NoiseAssist test?
@@curtisjudd Hi Curtis,
I did not mean to say should there shouldn't be *any* noise. That'd be an insane thing to do, and I have this principle of trying to avoid doing insane stuff.
I was merely pointing out that putting the fan *on* instead of off (etc.) is not a good move, if you try to find out how much the noise reduction system would help in a *typical* home environment interview.
Take an interview or recording I would make in my home environment:
I would have the kitchen fan off.
I'd turn a noisy PC, and room fans off. I'd close the windows.
But I still would have the ubiquitous background hum / low rumble of lorries, and the annoying higher-pitched sound of mopeds, the rumble & shaking and clinking of glasses in the cupboards when a train passes as teh house, I live just a hundred yards away from the tracks.
And if I'm really lucky the rumble of my stomach and that of my freezer (which I often forget to unplug) would be there to sabotage the recording .
Outdoors, then.
An outdoors recording I'd do in a park, not on a field near a motorway. There would be car bird noises, an occasional plane, but it would not sound like being insine a turbine engine..
In my case - real, not theoretical - the dealbreaker on NoiseAssist would is the the amount of help it would give me while recording at somewhat acoustically softened home environment where the most obvious and easily removable noise sources would have been dealt with.
-And FWIW, please don't regard my comment as a put-down.
I find your videos extremely helpful, well done, and enjoyable. Very high quality stuff, your content. Thank you for it!
MK
@@markuskajo903 Thanks for the detailed thoughts, Markus. That gives me some things to think about. Happy recording!
@@curtisjudd You two! And a new finding in my world of recording: My SD Mixpre-6 II seems to be allergic to non-Neutrik XLR plugs. It gets them stuck. Had to have a pro to take it out. it seems cheaper XLR cabler do not have up-to-standard plugs - and/or Sound Devices' Mixpre-6 II is too fastidious for the *real* prosumer world.
@@markuskajo903 Yes, I decided quite some time ago to spend the extra money for cables with Neutrik connectors. It isn't just Sound Devices gear where cheaper connectors get stuck. There are seemingly hundreds of people complaining about the same thing with the Tascam recorders.
Hello Chris. Can you compare new Noiseassist for Mixpre generations recorders vs 8xx series? Why is half price ? Of course, if you have it available. Thanks 👍
Hi Jan, yes, just need to save up for the MixPre version. 😀
Did SD licence the NR algo from another company or develop it themselves?
They developed it.
Near the highway, when u went to 10db your voice started to sound like on the telephone. Was wondering if you could affect the voice back to normal while keeping 10db?
In post? Not likely. But dialing the setting back to -3dB while there probably would have worked much more nicely.
@@curtisjudd which brings to mind another question - is it sometimes okay to process the voice in real time and with mixpre and the effects app, have processed vocals on a separate track.?
god job!
Thanks Moabe Filho!
Does it work for cicadas? They emerged this summer.
I haven't specifically tried but if it is like I've experienced, fairly constant sound, this should get you a few dB less cicada.
Curtis maybe do a demo of the Absentia DX plugin made by Todd AO it can be used on your files and keeps the metadata or also be used as a plugin and it’s cheap
Yes, but that's for post. This is for production cases where you cannot do post - fast turn-around news - or live. Definitely some great options in the post world. Will check out Absentia. Thanks!
Curtis Judd for 633 users , if had a 833 I’d use it ion the mix track for sure
@@MrBigbangbuzz Ah, I understand. Thanks!
I could hear difference on my ISK HD9999 headphones at 8dB. At 12dB it started to be unacceptable for me. But since you are not talking the same thing all the time without the slightest brake, it is possible that the difference could came sooner.
👍 started sounding a bit processed to me at -6dB, maybe -5dB.
@@curtisjudd OK. I must have been distracted. I am not used to hearing your voice through lav. I can definately hear something strange at 6dB in the first trial. In phrase "difference in terms". Some metalic nasality. Like highs and lows were missing. But it was noticeable the moment you moved. So...
Outdoors it was comical. Not only it coloured your voice as if you were talking through phone but it also attenuated it. It clearly treated your voice as noise. Traffic noise seems to be to unstable for such an algorythm.
I must notice that ISK HD9999 has very smiley tone (boosting highs an lows). Very musical. Not necessarily accurate.
@@przybylskipawel Ah yes, smiley voicing. I think for the traffic sample, the setting needs to be a lot closer to -2 or -3dB to avoid the issues we encountered.
@@curtisjudd Voicing! That was the english word, I was missing! :)
And I like the description “smiley”. 👍
Nice
👍
It sounds horrible starting at -5 dB. I’ll stick to using the Noise Reduction effect in Audacity. It allows me to take a sample of the noise first, which it then uses to reduce just the noise frequencies. And since it does not affect the original recording, I can adjust it as needed until I get it right.
Unless the Audacity NR effect can work in real time, it's apples to pears. And it definitely was horrible around -5dB. I was listening through very modest JBL LSR2300 monitors, and the sound became irritating at that point for the Sennheiser in the kitchen. My reaction was to jerk my head somewhat to the side and downward to redirect my ears.
Understood. I think that I'd keep unprocessed isos and just use this on one of the busses for fast turn-around or live work. Could also be useful on the mix bus for dailies on bigger productions. But yes, post will still want clean, unprocessed isos in most cases.
Agreed, apples to pears. Definitely a tool to use in very small doses, just enough to take the edge off.
At -9 dB your voice sounded muffled for sure...
And at -15 dB, you were talking through a dark chocolate muffin :D
If SoundDevices gave you a truly fabulous EQ, it could go further!
Hahaha! I agree, unless one is recording on a dynamic microphone about 1-2 inches from the mouth, about 4dB of attenuation is roughly the limit. And now I'm hungry for a dark chocolate muffin.
Can you do a comparison with cedar dns2?
Yes, if I happen upon a DNS2 I can use for the demonstration. I don't have any rental houses here in my city so that'll be tough, but it would make an interesting comparison some day. DNS2 would need to be substantially better for me to consider adding it to a kit with the extra bulk and powering it would require.
tomorrow need to record one actor unfortunately the location is unsuccessful historical film but a highway runs nearby :( I thought cheers there is a solution, but unfortunately after 3 decibels the sound goes bad :(( thanks for the interesting video :)
You're welcome. 3dB is pretty substantial in terms of reducing distraction. Best of wishes on your shoot!
Interesting. I'm sure you probably know all about software based noise reduction developments and I don't know how they compare in practical use but I was impressed by this BBC clip of software being used to reduce other noise...quite impressive as to where things are going. I wonder what you think of this, ongoing, development. BBC clip: czcams.com/video/BRg7LHAJqeU/video.html It's pushed to the limit but if not over taxed with ambient noise? Interesting to see the progress being made.
Definitely interesting to see the advancements. I think they have different use cases - the graphics card and computer software based have a lot of promise for post production but not practical for production like NoiseAssist. Great to hear all the options.
I’m guessing this NoiseAssist “plugin” would have been far more effective if it employed a LEARN function. Seems like a missed opportunity. Maybe SD can release and update before the competition does.
This is a high-level description of how it works: www.sounddevices.com/noiseassist-simplified/
Hey @@curtisjudd!, thanks for the reply. Big fan here. Just want to add that SoundDevices are not going to secure any patents with that level of detail, hehe. I’m sure they’re using an algorithm similar to the dynamic noise detection employed in popular DAW plugins, which can be quite effective. And they do use FPGAs, which means they can update the code in firmware if they wanted to employ a user-defined noise print ( “learning” ) feature, which would generally be more effective than using an algorithm to derive a noise print. Combining both methods would be even better!
Is this working on pre or post fade?
Depends on whether you assign it to an input channel or the L or R bus. Bus is post fade, of course.
@@curtisjudd Thanks! Considering this 4 mixpre6ii
To my ears, -9dB was the point where your voice noticeably was colored.
👍
Curtis, I can still hear the fans in my Condo.
Yes, just for noise reduction, not elimination.
It works well but but surely I could do this in post with a very simple piece of software and save 300 dollars?
If your workflow includes post, yes, this is not meant to replace that. But if mixing live or something with fast turn around, then this becomes useful.
@@curtisjudd THANKYOU - I See what you mean. Great channel by the way, I have learnt so much from you.
RTX voice seems a lot better because it also reduce lots of random background noises like barkings, baby cries, keyboard strokes, and for free :-)
Besides it's name RTX Voice also works very well with non-RTX NVIDIA cards with a slight change in the setup file.
There are significant differences. NoiseAssist is made for field recorders which are often battery powered and designed for use on film sets and locations. It'd be a different game to record a film with a PC. Also, this operates with 1ms of latency vs. 50+ms of latency on the graphics card. With film, 50ms is far too much latency. So while this will not eliminate or even reduce all noise, it does have some advantages to graphics card AI noise reduction. Also, graphics cards are not free whether they're RTX or not. 🙂
@@curtisjudd but for film isn't it better to modify the sound afterwards ? Then you can take samples of noises and apply a reverse Fourrier transform to remove noise frequencies. You can do that with any too like Audacity and it works very well for hiss and most cyclic noises. Is there any professional tool to graphically pinpoint noise peaks and allow a sound designer to clean them (like Magix Audio Cleanic for Vinyl discs, but on steroids) ? I don't understand the utility of NoiseAssist out of the monitoring purpose (like you said in the outro of this video it can helps to do fine synchronization).
@@sitedel A Dailies workflow is one case where I might want this just on the mix bus - I'd still want unprocessed isolated channels of each microphone for post. Corporate and new fast-turnaround is another - any case where there isn't time for post is where this has value. I've even done a livestream job recently where this would have been helpful.
Whoa I wonder how this would do at a convention hall or sporting event when they reopen in 2222?
Hahaha. I'll test it out as soon as I can.
Sounds as if it’s better than adobe premier noise reduction
👍
I wouldn’t use it more than 3 db
👍
I would not go lower that -2db -3db honestly
Fair call.
🙏🙏
👍
Curtis' sophisticated, methodical style is a HILARIOUS contrast with the recent videos about NVidia RTX Voice. See it: czcams.com/video/HA6R9NmZO80/video.html
Hahaha! Agreed, I am definitely not producing videos for the audience which found that to be a useful test. And this product is also not for the same folks or use cases. 🤓
Bad timing. Nvidia RTX Voice does this way better and it's free. I've been using it to clean up existing audio and it's almost magical how well it works.
RTX is cool but it has more than 50ms latency and is power hungry - well suited to gaming. NoiseAssist is 1ms latency and requires low power - much better suited to live professional production and on location.
Agreed, not bad timing. Completely different use case. RTX voice burns power which is usually sourced from batteries on film sets and on location. That 50+ms latency is also problematic in film production. And I don't have room for a PC on my sound cart. 🤓
@@curtisjudd I've been using RTX Voice in post production to clean up audio I recorded on set. It's not the intended use case, but it's been working amazingly well so far. Doing it that way negates the power consumption and latency issues. Well, the latency isn't completely negated as it does and in the step of re-syncing the audio but most good editing software seems to make that pretty easy. This NoiseAssist does seem like a much simpler solution, but given it's high cost and lower effectiveness than RTX Voice I still can't see myself considering it.
@@WrongThemeSong Thanks for the additional info. Yes, definitely looks like an interesting new option for post. I'm not trying to sell NoiseAssist, just explaining the intended use case and how it differs from graphics card NR. When I'm hired for a freelance location sound job, NoiseAssist is a better fit over graphics card or other hardware based solutions for me since I already own the 888. The NoiseAssist plugin is way cheaper than other hardware solutions typically used in location and on-set for noise reduction.
Your face is waay too dark on one side.
Sorry you didn't like the lighting.
Hi Curtis, what is the difference between noise assist and cedar sdnx plugin from your experience?
Hi Petr, I haven't used Cedar yet, but it uses a different algorithm to accomplish the same general goal. If the comparison is anything like the difference between MixAssist and Dugan Automixing, they each have their strengths and weaknesses - different situations which benefit from each. MixAssist, in my experience is a little more aggressive about reducing noise whereas Dugan is more transparent.
@@curtisjudd i was trying demo of cedar and noise assist indoor with mic on close distance and in noise assist i hear some kind of dullness in voice while in cedar i dont hear it affect the voice so much and even the residual ambient noise sounds very clean.
@@curtisjudd does the distance from the mic influence in which attentuation the voice will be distorted by the plugin?
@@sulfurhomeo Great - sounds like Cedar is better suited for your situation.
@@sulfurhomeo with NoiseAssist, I find that it works best when the microphone is close. Otherwise, there tends to be quite a lot of artefacting.