SpectraLayers 13 - A New Era in Unmixing and Spectral Editing. NOW AVAILABLE!

SpectraLayers
24

If you think this, I believe you should just look elsewhere, cause obviously the focus of Spectralayers is not in consonance with your needs, whether we take bugs into consideration or not. (And this is fine, the developer has no obligations to your particular needs or wants, nor are you without options.) Reading your posts it is clear to me that you don’t use SL for spectral editing in a post-production context, otherwise you would be able to appreciate the advances.

For spectral editing, which is undoubtedly the focus of SL (even more clear now eith v13), the upgrade is objectively immense and powerful and worth the price for any of the new main modules alone.

Mind you that I’m NOT a shill for Steinberg and I have been one of the more vocal users in this forum when it comes to expressing dissatisfaction towards bugs and what I consider bad practices.

But we must be fair. While Spectralayers indeed has bugs and problems (many of which are more related to Nuendo/ARA than SL itself), I use it daily in the past 3 and a half years, finishing more than 1500 minutes of runtime (my area is independent film and tv, mostly work alone doing everything from raw to finished mix) for various film and TV formats and genres, and it is a SPECTACULAR tool for spectral editing and audio restoration. I mean it is almost miraculous in what it allows you to do with audio. No other tool gives you so much power over manipulating audio and getting transparent results.

From this perspective, many of the bugs and issues I encounter while working with it I write them off under “being on the absolute cutting edge of the field and doing stuff that no one thought was possible a decade ago”.
So much so that you can get the old dialog editing manuals em largely throw them in the bin, cause it is a whole paradigm shift.

But I grant you: if you want to just unmix music stems, this is not your best bet.
Please do us all a favor and look for a better solution for you use case, there’s no short of alternatives mentioned in this forum.

26

Can you guys please stop this childish behavior like calling Apple “Crapple” and Microsoft “Microslop” and bashing operating systems like in a culture war? How old are you - five?

27

Maybe. Maybe not. Your expectations are yours alone. Wait for the trial demo of 13 next month and decide for yourself. Although if you like it, you’ll then be unhappy the intro price has expired :slight_smile:

29

I am not seeing this here:

image
image1920×785 316 KB

30

At no stage have I stated that’s my sole reason for purchasing and subsequently upgrading SL so feel free to mute or block me :face_blowing_a_kiss:

I am a paying customer, and that entitles me to have an independent opinion, I can make feature requests and and also air my praise or dissatisfaction.

From a workflow perspective, what would really make SpectraLayers invaluable for many would be the inclusion of dedicated VST3/AAX versions of the existing restoration modules (DeNoise/DeClick/DeClip etc). MAC and PC hardware is powerful enough to run these (CEDAR, Sonnox and iZotope show this). Being able to run these processes in real time inside a DAW mastering chain (and non-realtime rendering/bouncing) would be ideal. If they introduced a realtime alternative to DNS that would be a game changer! ARA just doesn’t work. I’m sure, in addition to music, those who edit videos and work in FCPX, DaVinci Resolve & Adobe Premiere would love to have these tools too.

33

Yepp, you’re right. Might be a graphic artefact. The waveform looks fine. So it’s no error here.

34

I have to voice my opinion, and I’m sure it is shared with others: this should have been an update. It hasn’t even been a year. Complete upgrades that fix bugs, or make things easier that should have been easier to begin with are inappropriate for people that use this and do not make enough money to pay for it. And it hasn’t even been a year!

35

This is incorrect. It has been exactly one year, almost to the day and this is even a long time in software cycles these days. The development of 25+ features is cost intense and since it isn’t a subscription, we don’t force anyone to buy. SpectraLayers 12 has received 4 maintenance updates in the last 12 months, of course, free of cost.

36

But what about all the bugs and glitches all over the place?
Why not fix them first and make the programm reliable and trustworthy, before introducing new features?

37

My mistake. Thank you for correcting me.
For some reason I thought it came out in November. Probably because I have several Steinberg products.

38

The anti Mac feeling sometimes can be strong here. As far as I’m concerned, the developer had years to overcome his problems, others do and their apps work as well on Mac as on PC. All I read here are excuses and anti Mac rhetoric voiced against “Crapple”.

It’s on the developer to make their software work well in a particular environment.

39

I’m with Timo on this one.

IMO, this is a VERY worthy upgrade. And absolutely the best version of Spectralayers so far.

I think many people here are not grasping how powerful these new modules are. Perhaps I’m particularly happy because they fit my workflow so well? I dunno… SL is very tightly integrated into my way of working.

43

Reading a manual and actually using the product are 2 very different things. You cannot genuinely pass comment on any software if you have never actually used it. Furthermore, when it comes to something as complex as SL or ReTouch, even a 14 day demo is not really enough time in my opinion.

I have used both extensively. I have paid for both SL and ReTouch, buying the latter twice when it launched on SADiE and again on the Cambridge system.

It is clear you work primarily with dialogue, and I honestly wish I had that luxury. Music restoration and TV/film are very different fields, and I find that music needs far better interpolation. I am happy for you that SL does all that you need of it. That is a wonderful position to be in, but please do not pass judgement or make wild assertions about me or my knowledge of any particular product that I own or use.

44

Ok dude whatever you say.
Best of luck in your search!

ba-bye

45

Mod note:

I have tried to remove the vitriolic posts from this topic, and to leave in place the posts specifically about Spectralayers. I apologize to anyone whose post was removed.

49

I am genuinely not sure what search you are referring to, but that is on you.

As a polite suggestion, perhaps in the future you could refrain from telling other professionals what to do.

We work in very different fields and are clearly never going to need the same tools or have the same experiences (given that the Windows version has had GPU rendering for years), so your suggestions aimed at me should really end here.

Take care.

50

interpolation (resynthesis) from left, right, up, down, horizontal and vertical is not new in spectral editors. I’ve been using those options for over 20 years in a competing product.

51

Dear Steinberg: If indeed SpectraLayers 13 HAS improved, then please PROVE this by creating demos that we can all watch on You Tube. Not just the short 5 minute demos for selected features. But real tutorials that show us, in depth, how a feature has been improved.

I have SpectraLayers 12, which I updated to without question when it became available last year. After one year, I have to say, I don’t use it at all. I do all my stem separation in Logic Pro, because IT WORKS, and SP12 does not work. Literally. I’m not exaggerating (I wish I was). Case in point, I recently used stem separation in a project. Project 1 had vocals, guitar, electric piano, bass, and saxophone and/or flute. Everything split into stems in Logic Pro perfectly. This didn’t work in SP12: even the sax did not split out as one would expect; only a few notes got extracted. But even the other instruments were very spotty, with some sections missing. Sure, I could spend hours cutting a pasting from one voce to another, but why?

(I also tried creating an Unmix Flute module by following the instructions within SP12 — but it did not work either.)

For Sound Repair, I still use RX9 Advanced from Izotope. Personally, I hate the upgrade prices that Izotope wants: $600 to upgrade is ridiculous! Now, perhaps SP12 (and SP13) can do the same repairs as RX (?) But it is hard to know whether this is true, because you can find many, many RX tutorials on You Tube; but almost nothing for SP.

Really, I am not a complainer normally, But I am simply asking: if SP13 is really an improvement, please make video demos of each feature and show me what has improved. And I mean, real tutorials, explaining how to do it.

Thanks for listening.

52

@ErnieMuse Apparently at the end of June/the beginning of August the demo version of SpectraLayers 13 will be available. I assume that the best idea is, to test it then by yourself.

53

I agree completely with what you said about the lack of training material on Spectralayers - and on Steinberg stuff in general. They really lack in this department compared to other companies.

Absolutely it can, and much more. But with a few caveats:

DeClick, Voice Declick and Unmix Multiple Voices modules are essentially broken, and have been since at least v9.
I find de DeReverb pretty weak, so I don’t use it.
Some FFT sizes still cause the GUI to get laggy when enough spectrum is visible on the X-axis.
ARA can be finnicky, but you can learn how to work with it.
There are bugs present, as people have reported here many times (myself included).

Now mind you, I only work with it inside Nuendo in ARA2 mode.

After years of using RX advanced in music, when I moved to Post I just couldn’t handle the hassle of working with it, for two reasons: RX Connect is terrible; and even if I used all of RX’s editor power, there was just so much stuff that I couldn’t clean up in an acceptable time frame. So I looked into Spectralayers, learned it, and never went back.

For me it works better than RX for mainly two reasons: one is the automated separation of audio in layers (mainly Unmix Noisy Speech) and the ability to work those layers in parallel. The other is the ability do adjust the FFT size with a key command. This makes its ability to restore audio (at least in a post-production environment) simply unmatched IMO.

For example, using RX there is no easy way to separate the voice from the background, for starters. It would be an extremely time-consuming, herculean endeavor, and with poor results. So while many modules there work very well, you are stuck with the “whole” audio the whole time. Any unwanted sounds that overlaps with what you want easily become an issue, cause there is no way you can work around everything manually.

Spectral and/or dynamic de-noisers have severe limitations, cause artifacts, lack control and deal poorly with random or periodic “events”.

Now with Spectralayers you can just separate dialog from background, quickly send any mistakes to their correct layer, and remove or attenuate whatever bothers you while keeping your rich, unique backgrounds (if that’s what you want), cleaning your dialogue perfectly (since you can constantly adjust the FFT size as your work, you can very quickly tackle even the most minute details, be it short or long, broad or narrow).

And I have seen this tested in real time a while back, when I supervised a TV series and the dialogue editors working with RX and real-time plugins simply couldn’t get rid of a LOT of stuff that I could clean up with SL very easily. I’d say “hey, you missed that at xx:xx:xx:xx:xx” (knowing that I absolutely could clean that with SL) and I’d get “oh, but that overlaps with word, is too broad, can’t clean it”. And these were editors with college degrees in cinema/audio and years of experience.

Why I’m not bothered by the caveats I pointed out:

Honestly, I have de-click modules already, I don’t really need another one. I have RX 10 declicker and Acon Digital declicker, and they take care of simple, generalized declicking jobs. Just I give it a pass after SL when needed.
They are not enough anyways, right? They work well with proper “clicks”, but if it is a ting, a scratch, a crumple, a crack, a ploc, a crash, or anything else that might excite the room, then you are toast. So they do their job, the spectral editor does its job. For detailed work the Transient selection + Reconstruct is the most transparent Declicker I’ve ever used.

De-reverb is the same. I already have de-reverbing tools that I prefer, so I use those for general de-reverbing.
Any particular more problematic passage will need special attention anyways, so when needed the other tools in SL are used.

The lag problem has improved tremendously to where it is almost a non issue, but I still hope Robin will fix it soon.

ARA is problematic, but all I need to do is: apply it, work on stuff, make permanent. That’s it. And that is MILES better than the old RX Connect days.

I think that is it.
For me the bottomline is: Even with all the problems, I find that Spectralayers allow me to work faster than the alternatives while delivering a superior result. It can do things in a way that no other software I know can, because of the SUM of its capabilities. Many times during work I just start laughing cause I was able to clean up something that I just cannot believe is possible. I very rarely have to comp room tone nowadays, or worry about audible loops, or whatever (I always hated comping room tone).

Speaking on a direct comparison between similar functions of RX and SL, I don’t think the results from the processing in RX are any better than in SL. And surely any marginal advantage RX could possibly have on specific processes quality is far surpassed by the undeliable workflow advantages SL has over it.

But that is for the SPECTRAL EDITING SIDE!!! (and my bias is towards Post-production)

I don’t use stem separation much, but from what people report here, there are (much) better alternatives. And I get it. We are in the AI age already, things like Stem Separation will always develop much faster than the lifecycle of Spectralayers is capable of keeping up with.

And to be honest I hope it keeps that direction. Stem separation in SL can save you in a pinch, but I hope Robin keeps focused on the Spectral Editing side, specially for post-production.
Man, these new modules are amazing. Reconstruct, Ambience Heal, Attenuate… are just such time-savers!

And let me add that no software is without bugs. A guy in this thread recommended Cedar Retouch, and I went to try it and it just refused to work at all. So…