Ozone 5 & Ozone 5 Advanced is Officially Released

[MattGray]

Originally Posted by bmanic

This I have to completely disagree with though. If you do a proper A/B of Ozone which has limited something moderately heavily (-3dB of gain reduction or more) and then compare it to the original source, at equal loudness, you will very quickly realize just how much Ozone is exaggerating the transients. They are very far from "natural" and don't resemble the original file at all.

I don't hear it as 'exaggerating' in as much as 'preserving'. Clipping an A/D 'preserves' transients also but adds peak distortion or saturation. Would you call this exaggeration also? This is how I'm hearing & would compare the sound of transients in IRC III (to high end A/D clipping).

It's reminiscent to how FG-X sounds.

With regards to transients a little maybe... I liked how FG-X treated drums (Slate is a drummer, so am I... he nailed that aspect). However, a BIG difference.. I don't hear IRC III generating any weird distortion artefacts or anywhere near the same level of pumping unpredictability as FG-X though. FG-X was very coloured though & based on 'intelligent saturation' which didn't work too well with artefacts & weird pumping at any setting imo. IRC III has a predictable release depending on where you set the character timing. FG-X was unpredictable & would change levels & short & long release timing independent of where the ITP control was set making it russian roulette with your audio.

Especially the first few samples are more brutal and "ticky" than the original file.

no more so than A/D clipping imo. I'd rather this that smeared mush (unless the mix needs it's transients tamed more of course).

In this case, when it comes to "natural" sound, I think Pro-L is superior.. but yeah, I am biased.

I still find the veiling which I assume is the oversampling &/or any dither it may be using the biggest downside to Pro-L. But I also thought the transients weren't as prominent as even IRC II in Ozone 4 when I compared them.

There is a lot of plug-ins that sound a bit plastic/veiled makes sense now hearing about the linear phase real-time oversampling that seems to be in a number of plugs. I just don't like the sound of it. Even inserting them without any limiting makes the signal sound worse. Fix this in Pro-L and I'm sure it would be a lot more 'natural'.

This is made worse by the tiny pumping/nervousness that the IRC3 algorithm has, especially at low Character values. It's really hard to describe it but "natural" is definitely not a word I would use for the new algorithm. Rather I would call it aggressive and upfront / in your face kind of sounding.

Looks like Alexey is onto this with a further revision, but honestly at around 1.7 it sounds pretty natural to me on a fairly wide range of material if it's not pushed too far & you keep ISP protection off. You can also see the actual peak reduction in the 'Gain Reduction Trace' display. I can't see that if you're actually limiting/clipping fast percussive transients and the release is set fast how you could perceive much in the way of nervous pumping because it's acting more like clipping. But I know what you're saying/hearing & that was more obvious to me with longer release settings not shorter ones. Try 1.7 for me with up to 2dB of GR & tell me it's not acceptable musically?

The older algorithms are much more natural in my opinion, at least when considering the whole sound, not just the initial transients (which are indeed a bit soft on the old Ozone 4 limiter).

It's still there & I still use IRC II on some jobs with the added stereo link feature. You choose the setting that works for the mix. I still highly rate IRC II & IRC III won't simply replace it for everything.

EDIT2: to further clarify; With the IRC1 and IRC2 algorithms when you set the character slider to maximum, you can keep cranking down the threshold almost all the way to the maximum -20dB value, without the music completely loosing it's inherent "flow" and feel. This is what happens when there is a well optimized and tuned 2nd "slow" release stage. Ozone always had this very nice 2nd release and a nice balance between the fast and slow envelopes. However, IRC3 algorithm works very differently.. try cranking that one down and the music gets completely destroyed, not just by distortion (that would be expected) but also in it's flow and rhythm.

I'm not interested in how any limiter sounds with a -20dB threshold. It's how it sounds as part of a normal "real world" work flow & chain that counts & speaks volumes to me (pardon the bad pun ).

These disruptions in the flow of the music are still noticeable even at moderate limiting and that's my main gripe with the new algorithm.

Fair argument, but I still would argue that at the right release setting & with a useable threshold amount it still has some significant benefits over other limiters including IRC II imo. Lets hope Alexey can improve the release algo even further so it becomes more flexible/natural at longer release settings.

[lowland]

Thanks for posting, Alexey. This is a good opportunity to say I greatly appreciate your calm and measured approach, and usually learn something when you post - I expect there are others here who would agree with that.

WRT the intersample detection, very possibly what I'm hearing is the slight increase in limiting with ID engaged as you say. That said, simply backing off the output a few tenths seems enough to avoid obvious intersample peak problems here - for example, I don't get mp3 conversions collapsing or clipping unacceptably, even on very loud material.

[Mike Couzzi]

I am demoing the Advanced version.In WAVELAB 6 it CRASHES in the Montage track insert.In the Montage Master section it is fine.
I like to use this plugin sometimes before external hardware processing.
I wouldn't consider buying this till this problem is resolved.
Anyone else having this problem?

[mljung]

Originally Posted by Mike Couzzi

I am demoing the Advanced version.In WAVELAB 6 it CRASHES in the Montage track insert.In the Montage Master section it is fine.
I like to use this plugin sometimes before external hardware processing.
I wouldn't consider buying this till this problem is resolved.
Anyone else having this problem?

I had, it's actually the authorization pop-up that's making trouble. Contact iZotope support for a workaround...

::
Mads

[bmanic]

Originally Posted by Alexey Lukin

This is strange. I've heard this opinion many times, but algorithmically the only difference should be the slightly deeper limiting in places where intersample peaks are present. It's only a fraction of a decibel though, should be verifiable by a null test.

I too am one of those nutters that think Ozone sounds a bit "dull" and loses some impact when you turn on the 'prevent intersample peaks' option.

After a lot of testing a few years ago I came to the conclusion that even fractions of a SAMPLE position (ridiculous levels like 0.05dB and minimal time changes like 0.02ms) can and WILL be heard. It seems that for whatever reason (which I don't understand and thus speculate) the actual analogue waveform from the DA converter changes slightly even with these fractional sample positions. The place where these changes are heard are at the very start of a transient. That's where it is detected.

I have yet to get a good enough analogue oscilloscope (apparently those monsters are ridiculously expensive!) to see how much the waveform actually changes but if I may speculate, I'd say the human hearing is sensitive to transient changes. That tiny difference of a "tick" that happens on each impact, on some sounds more than others, can be detected. I've done quite a few A/B tests to confirm this but alas, it's not as clear cut as I had hoped.. but the results are leaning in favor for my theory. I don't remember how many tests I did but for a few sample position changes within transients I must have done at least 50 ABX tests and got over 80% correct. Statistically I'm not sure how valid that is but it might be worth investigating.. or then all I've written is pure placebo. That is also entirely possible but I doubt it.

Is it relevant and important in the context of what clients and end users may think? Perhaps not.. but it never hurts going for the ultimate goal. "The Bestest!!"

Interestingly, Ozone does not oversample the limited signal at all! It only detects levels from the oversampled signal, but limits the original signal: this saves us from any artifacts of upsampling and downsampling.

This is indeed interesting! I also get a weird feeling when stuff is up/down sampled. I don't exactly "hear" it but somehow a sound is more tight and slightly constrained.. kind of like it was subtly compressed. It physically manifests on me as a slight pressure feeling on my forehead! I kid you not!

Anyhow, Kjaerhus MPL-pro did the same thing, and I'm sure a few others do it as well (oversampling the sidechain/detector path) and I firmly believe this is the correct way of doing it, so massive thumbs up for Ozone on this front.

When Transient recovery is off, Ozone does not apply any "special treatment" to transients except for changing its attack/release time to almost instantaneous (becoming a soft clipper). Perhaps this change in the attack/release duration is too strong and causes too significant level variations.

This is very possible. I do hear a change when compared to the original content. This is to be expected of course, all limiters and clippers change the sound, of course, but in the case of Ozone and FG-X, it goes in the opposite direction. Usually transients are a bit softened and dulled but in these two limiters they get a bit harder. Kind of like just the initial part of the transient got a massive high frequency boost.

Are you sure Ozone doesn't suffer from a short period of some rather high intermodulation distortion, perhaps during this very change? What I hear does remind me of distortion of somekind. Modulating the attack and release during the transient does indeed make it a waveshaper so perhaps it can be smoothed/tamed a bit?

In Pro-L you CAN actually get almost the same behavior, by opening up attack completely. This makes it virtually a program dependent clipper.

IRC III mode is currently under improvement in order to reduce these level variations.

Cheers!
bManic

[bmanic]

Originally Posted by MattGray

I don't hear it as 'exaggerating' in as much as 'preserving'. Clipping an A/D 'preserves' transients also but adds peak distortion or saturation. Would you call this exaggeration also? This is how I'm hearing & would compare the sound of transients in IRC III (to high end A/D clipping).

Well at least on the prism converters I have it doesn't sound like this.. but yeah, converters all clip very differently. For instance the Lavry Gold mk3 clips really beautifully in my opinion. It's actually a bit "soft" somehow.

The first "tick" of a transient is NOT equal to the punch or impact of said transient. It's just the very initial impact that is changed in Ozone. I agree that the overall transient is extremely well preserved in Ozone (also in FG-X) but something is a bit weird about the very first tick of it. It feels forced on certain sounds (mainly sharp kicks, hihats, rimshots, claps and sometimes plosives on vocals like 'p' 'b' 't' etc).

Anyway, this is splitting hairs.. it's not at all a serious issue.

With regards to transients a little maybe... I liked how FG-X treated drums (Slate is a drummer, so am I... he nailed that aspect). However, a BIG difference.. I don't hear IRC III generating any weird distortion artefacts or anywhere near the same level of pumping unpredictability as FG-X though.

Agree and disagree. IRC III is much, MUCH cleaner than FG-X (which can get very constrained sounding on material which contains little to no distortion to being with) but that pumping "unpredictability" is exactly what I hear in Ozone IRC III. It's on a tinier scale than the FG-X (which only has that problem when you have to turn down the ITP slider) but it is still enough to disturb the "flow" of the music.

no more so than A/D clipping imo. I'd rather this that smeared mush (unless the mix needs it's transients tamed more of course).

Yeah, that's what I thought as well until we did that crazy double blind test a few years ago. The king of smeared mush was actually the runner up in this test (waves L2, winner was Voxengo Elephant). I now know the reason for it.. it preserved the actual musical flow better than the others we tried, sometimes even enhancing it. Sure, once you looked at the "details" like individual transients and such it was obviously inferior to some of the competition but very few tracks needed that "attention to details" but rather suffered much more due to the interruption of the flow and rhythm. I'm very sensitive to these things now and try to stop looking at details too much.

I still find the veiling which I assume is the oversampling &/or any dither it may be using the biggest downside to Pro-L. But I also thought the transients weren't as prominent as even IRC II in Ozone 4 when I compared them.

Perhaps you didn't try it properly? Who knows.. the transients are definitely there if you set it up the way you want it. At least compared to IRC 2.. then again, all the limiters are so different. None can work best in all situations, not even Pro-L, though it contains 4 entirely unique algorithms (each algorithm has absolutely nothing in common with the other.. they are not merely a few changes to envelope timings, they are completely independent algorithms using different techniques to achieve the limiting. Kind of like the major modes of Elephant).

There is a lot of plug-ins that sound a bit plastic/veiled makes sense now hearing about the linear phase real-time oversampling that seems to be in a number of plugs. I just don't like the sound of it. Even inserting them without any limiting makes the signal sound worse.

I agree! My description of "the sound" is different but I wholeheartedly agree. Something is "off" and slightly unpleasant when it comes to linear phase oversampling. I do still use Pro-L almost constantly in x4 oversampling mode because the good far outweighs the bad in this case.

Fix this in Pro-L and I'm sure it would be a lot more 'natural'.

Nothing I can do about it but I hope the FabFilter guys will at least consider it. Aleksey Vaneev of Voxengo has already begun implementing minimum phase oversampling for his plugins and that is absolutely awesome of him!

Looks like Alexey is onto this with a further revision, but honestly at around 1.7 it sounds pretty natural to me on a fairly wide range of material if it's not pushed too far & you keep ISP protection off. You can also see the actual peak reduction in the 'Gain Reduction Trace' display. I can't see that if you're actually limiting/clipping fast percussive transients and the release is set fast how you could perceive much in the way of nervous pumping because it's acting more like clipping. But I know what you're saying/hearing & that was more obvious to me with longer release settings not shorter ones. Try 1.7 for me with up to 2dB of GR & tell me it's not acceptable musically?

I actually found values under 2 rather nervous sounding. Perhaps we are listening to different things? I'm not worried about "traditional pumping" as in obvious, uh, pumping!

Rather, what I hear is a kind of "micro stuttering", where the transient takes precedence over important sustained elements in a mix, like a vocal. Kind of the opposite to what good AD clipping does. If you have a vocal heavy mix with sharp transients and you clip it, nothing changes. Vocals still sound exactly as smooth and controlled as in the original mix, just with some added distortion and the transients have lost a tiny bit (or a huge amount as it is sadly with today's levels!) of impact.

However, in the current IRC3 algorithm I actually hear the vocals "ducking" at a very rapid rate, sometimes significantly so, during each and every limiting peak. It's kind of like a program dependent hold was a bit delayed. Very hard to explain but the overall feeling is that the actual transient gets precedence over vocals, making the vocals feel slightly nervous and disjointed. This means it loses it's flow and depending on the music this is catastrophic.

Once you hear the added option Aleksey has planned you'll hear what I mean. It's going to be awesome.

It's still there & I still use IRC II on some jobs with the added stereo link feature. You choose the setting that works for the mix. I still highly rate IRC II & IRC III won't simply replace it for everything.

Agreed! IRC1 and IRC2 seem to be very useful as well. I just noticed this while playing with the demo. It makes the limiter section of Ozone extremely versatile and desirable.

Unfortunately I don't yet own Ozone (have been on the fence a very long time, since Ozone 3) but with every new revision the limiter part seems more interesting. At the moment the pricing scheme is putting me off quite a bit as all I want from the thing is the limiter and I'm definitely not paying 999$ for it.

However, I'm liking the new IRC3 mode a lot. It has tremendous potential and I'm pretty sure that with the new addition it might be extremely good. 999$ good? We'll see. Hopefully the anti-pumping measures can somehow be activated in the more basic version, then I could purchase that instead.

I'm not interested in how any limiter sounds with a -20dB threshold. It's how it sounds as part of a normal "real world" work flow & chain that counts & speaks volumes to me (pardon the bad pun ).

.. but testing the limits () will show you quickly where possible problems might arise. I test all new dynamics processor this way. Slam them to bits and then listen to what makes them tick. Ruben Tilgner of Elysia and SPL fame told me a few years back that it's a really good way of testing dynamics processors.. so having a genius pretty much agree with my testing methods was a huge ego boost!

Fair argument, but I still would argue that at the right release setting & with a useable threshold amount it still has some significant benefits over other limiters including IRC II imo. Lets hope Alexey can improve the release algo even further so it becomes more flexible/natural at longer release settings.

Absolutely! I didn't mean to come out all negative about the IRC3 algorithm. It's exactly the opposite!! I like it very much and can clearly hear the potential. There are only a few issues with it as far as I can tell and one is the micro-pumping stuff I've discussed here. I've already heard some extremely promising results from Alexey. The future is very bright for you lucky Ozone 5 users!

Cheers!
bManic

[Laurend]

Originally Posted by Red Mastering

I also noticed that guy who is behind flux company (mastering plugs)
opened a webmastering service for 5 Eur per song if I am correct - so he is hitting his own potential clients
I dunno is it greed or blindness, or what...
sorry for offtopic - checking this ozone thing

I'm no more involved in this company for almost 3 years now.
Ozone is my main challenger, not the mastering studios

[bcgood]

So far I'm liking it better than 4. Good stuff, good job Mark and team! I guess your education at MIT and strict adherence to good programming and science is paying off!

[thermos]

I sure hope those limiter improvements make it into the standard version, even we don't have as much control as advanced users.

[Lights]

This thread has been educational. Ozone 5 is so deep that it's interesting to see 5 pages or largely constructive dialog on the limiter stage alone. We haven't even touched te surface of the multi-band dynamics, new saturation and reverb algorithms...

Phenomenal work Alexy, BTW. To my ears Ozone 5 is the best sounding (what do I call it...? Limiter? Dynamics package since I use the maximizer and multiband dynamics together?) on the market. It easily replaced 3 plugins on my master chain in one shot and I haven't even scratched the surface of what I can do with it.

[Rossd25]

Another thing that I like about Ozone (which you can also do in Final Plug 5) is that you can audition the different dither types and shapes in 8 bits. This has helped me tremendously in understanding dither and what it does and which dither to pick.

[MAzevedo]

The Ozone 5 vs. Advanced thing is really aggravating. The 'big' bonus of Advanced is the meter plug, which I have zero interest in. I'd love to have separate plugs for each module, and the Transient Recovery and extra bits in the Dynamics section are really appealing, but I'm having a hard time justifying an extra $300 for those features. I wish iZotope had done Ozone 5 at $250, Ozone 5 Advanced at $500, and the metering as a separate $500 plug than forcing it into a bundle with the actual mastering processors.

[Michael Pane]

Unless I've missed it, I've read through the whole thread and no one has mentioned the big S word ( No wise cracks please ). I'm in 10.6.3 on mac. Just did a fresh OS install and will be in 10.6.8 I think it is soon. Anyways, am I not seeing any posts on this because its solid I assume. Logic 9 was not as solid as it is now etc after an update, any comments on V5 stability vs V4 stability? Is it a tight release? I want to taste this this new limiter everyone is raving about but not at any hits in the S word....

[MattGray]

Originally Posted by Michael Pane

Unless I've missed it, I've read through the whole thread and no one has mentioned the big S word ( No wise cracks please ). I'm in 10.6.3 on mac. Just did a fresh OS install and will be in 10.6.8 I think it is soon. Anyways, am I not seeing any posts on this because its solid I assume. Logic 9 was not as solid as it is now etc after an update, any comments on V5 stability vs V4 stability? Is it a tight release? I want to taste this this new limiter everyone is raving about but not at any hits in the S word....

Extremely stable, even in the beta versions I don't recall the previous 5-10 builds crashing any host. There is a 5.0.1 patch about to be released with further small bug fixes. But stability has never been a major concern with any iZotope products that I can think of which is a true testament to the programming skills & the internal & external beta testers. Stability is great.

[beau_mckee]

Originally Posted by marchhare

Just out of curiosity, you say you're only using 2db of limiting, how loud are your masters? The reason I ask is because I always wonder how people get reasonably competitive levels using so little limiting. Are you clipping before the limiter and feeding it a really hot level? Thanks.

I am still learning and refining techniques myself, but it really helps to focus on compression and balancing before the limiter, so your just skimming off the peak transients to get it that bit louder/ consistent on the final limit. Im sure that's generally the consensus, but im sure someone may be able to offer a different perspective for their wanted outcome

[Michael Pane]

Originally Posted by MattGray

Extremely stable, even in the beta versions I don't recall the previous 5-10 builds crashing any host. There is a 5.0.1 patch about to be released with further small bug fixes. But stability has never been a major concern with any iZotope products that I can think of which is a true testament to the programming skills & the internal & external beta testers. Stability is great.

Thanks for the insight Matt. Much appreciated.

[thermos]

Just wanted to say again the limiter is awesome. Really another level for me.
I hope the improvements make it into the standard version, I hear its release action with as little as 1.5 db of reduction. Worth it though as the sound is great.

[Thomas W. Bethe]

There is a new version available. 5.01 is out and downloadable.

[mljung]

Originally Posted by Thomas W. Bethe

There is a new version available. 5.1 is out and downloadable.

I think it's still 5.01 which was released Nov. 22, at least that's what it seems like according to iZotopes web-site.

::
Mads

[Thomas W. Bethe]

Originally Posted by mljung

I think it's still 5.01 which was released Nov. 22, at least that's what it seems like according to iZotopes web-site.

::
Mads

You are sooooooo right and I was soooooooo wrong. It is version 5.01 and not 5.1 Sorry about the confusion....

[ChubbsMaloney]

Originally Posted by AMIEL

Is that true?

what do you think Ozone 4 users?

Originally Posted by janjaal

do you guys really think the limiter sounds better than fabfilter pro-L, i doubt so, i'm just fooling around with both limiters, and i'm loosing lots of lowends when i push ozone's limiter.. pro-l keeps all the dynamic, when but i'm just fooling around..
what i really like about this ozone is its multiband compressor, much easier to work with compare to uad multiband,
the eq sounds very nice too,
but everyone is raving about the limiter, i think its the weakest feature....

Originally Posted by greggybud

Let me get this straight....so with the exception of Jerry Tubb...plus Dan Collins & John Scrip who probably haven't had time to chime in yet...I should assume the new Ozone is the golden ticket to mastering?

Then Brian Gardner, Bob Ludwig, Doug Sax, Emily Lazar, Brad Blackwood, or most anyone at Sterling...considering their years of experience, should dump all their hardware, keep their rooms, and just convert to Ozone?

I'm simply trying to emphasize that this tool with marketing hype like "Your mix isn't finished until Ozone is on it" is being marketed toward DIY. That's do-it-yourself musicians, producers, engineers, and anyone else who wants to risk their sonic outcome by saving money. Of course risk is greatly reduced by user experience.

Too many GS contributors...judging from often innocent type questions asked...are lead to believe the sonic results will be at the same level of those names mentioned above. These same contributors continually are misled that the majority of loudness is achieved by whatever limiter is placed at the end of the chain, hence the never ending "whats the best limiter" questions. I find it amazing users can distinguish the color, tone, and distortion types when any limiter is being used for just an additional -1db to -2dbRMS gain. I must be going deaf! Actually perhaps that issue could be largely addressed before it even hits the limiter?

Please don't misunderstand me, I sat in on a session with a well known ME who used the L-2 when he could have chosen much more expensive hardware. These are all just tools. But the marketing hype behind Ozone is misleading at best. The proof IMO is the type of questions being asked here on a regular basis often by Ozone users, usually about loudness and multiband compression.

exactly!...heard this all before with Ozone 4...almost point for point...transparent blah...wonderffulblah...better than blah....mutlband blah


It's a DIY plug...Ozone makes good plugs...not great ones.

[mljung]

I wonder how to save options-settings!?

If I recall correctly older versions of Ozone remembered changes made in the options dialog. Anyhow I need to know!

Please also check how the presets work, the compare system seems only to work sometimes, or maybe I just haven't figured out how it works exactly..!?

::
Mads

[thermos]

Originally Posted by mljung

I wonder how to save options-settings!?

If I recall correctly older versions of Ozone remembered changes made in the options dialog. Anyhow I need to know!

Please also check how the presets work, the compare system seems only to work sometimes, or maybe I just haven't figured out how it works exactly..!?

::
Mads

I can't figure this out either....

[nott]

any news about a standalone or single plug ozone meter?
the integrated advanced metering is really frustrating!!!

[Mike H]

Originally Posted by MattGray

Looks like Alexey is onto this with a further revision, but honestly at around 1.7 it sounds pretty natural to me on a fairly wide range of material if it's not pushed too far & you keep ISP protection off. You can also see the actual peak reduction in the 'Gain Reduction Trace' display. I can't see that if you're actually limiting/clipping fast percussive transients and the release is set fast how you could perceive much in the way of nervous pumping because it's acting more like clipping. But I know what you're saying/hearing & that was more obvious to me with longer release settings not shorter ones. Try 1.7 for me with up to 2dB of GR & tell me it's not acceptable musically?

Lets hope Alexey can improve the release algo even further so it becomes more flexible/natural at longer release settings.

Matt, have you used version 5.02 yet?
I am just wondering what IRC3 Release setting you find is equivalent to your 1.7 max with version 5.01.

[Alexey Lukin]

Some technical details on what has been changed in IRC 3 algorithm in version 5.02.

[nott]

Originally Posted by Alexey Lukin

Some technical details on what has been changed in IRC 3 algorithm in version 5.02.

hi alexey! i think we need a single meter plug. yet the meter integration isn't useful!!!

[cowudders]

Originally Posted by MattGray

The exciter is also extremely useable & versatile. Running in single broadband (parallel) mode I'm preferring some settings to the HEDD's excellent processing.

Hello, maybe I'm too dumb, but how do you set up the 'parallel' routing on OZ5? With Alloy this is no problem, ease to do in 'Graphs'.

Or do you mean mixing it in via the 'Mix' slider?

thanks!

[Alexey Lukin]

Yes, parallel compression is controlled by the mix slider.

[cowudders]

Originally Posted by Alexey Lukin

Yes, parallel compression is controlled by the mix slider.

Alexey, thanks!

Ozone 5 is quite a monster of a plug-in, congratulations!