Any news on Adobe Audition 4?

[gnu23]

IIRC Craig Anderton had an article in SoS in the last year about using AA's spectrum feature to "paint out" all kinds of artifacts in audio. It was a good read and informative. If SF can do something similar, I'd check for Anderton's article as it may be helpful. Sorry I don't have the link handy.

That said, I use AA only because I used CoolEdit from back in the day. I use AA because in essence I have always used AA, even when it had a different name. With respect to the SF vs AA ddiscussion, I don't have a dog in that hunt. If SF works for you, great. If AA works for you, equally great.

[drjohnny79]

As far as I know Wavelab has an advanced spectrum editor too.

And regarding Samplitude vs AA, Elastique ts/ps is for realtime processing, Radius in AA3 is offline and supposed to be better quality.

/Johnny

[birrbits]

The spectrum editing .........

It's for more than removing breaths. I would even say it's overkill for that. But you could even use it to remove a kick drum from a drum beat, and much more.

Like removing the stupid dog barking outside right in the middle of an outstanding drum session...

[Optomadic]

What caught my attention when Adobe 1st put Domain editing into 1.5 was how it could be also used creativly to do things you could never do before like move a dominant frequencies from one peice of audio to another or take a kick snare or other hit out of a recording. In AA2 it got the lasso and that opened even more possibilities! This definatly changed the way I looked at music and editing forever.

I think that WL got Algorithmix Renovator. I know that the Samp desk got it. This is also an amazing piece of software that suspiciously came out around the same time AA got FDE.

Now if only Adobe would aquire WIDIsoft then they could put the truetone editor into AA4 !



Those little rectangles are MIDI notes

Imagine being able to paint/detect MIDI notes over audio in the Spectral view! I mean its halfway there already as AA does show the note and cents at the top of the frequency analysis box.

[BONES]

You know, I think it is all possible, and most of it reasonably simple, already in any audio editor. It seems to me that this particular implementation has just opened your eyes to different possibilities. A lot of what you describe, like changing the amplitude of individual notes, is really just EQ, which is really just a specialised use of basic filters. It is all achievable by other means and, unless the temporal aspect is nailed, it is probably safer to do it in the normal waveform view.

There's the healing brush in spectral view which works much like the healing brush in Photoshop. Paint over a click or other sound, and it's gone, often without affecting other audio.

How does that work temporally? Does it affect the audio over its full length or do you define a time region? Photoshop works with single images and doing the same thing in After Effects, which works with moving images, can be a disaster. hence my curiosity about how Audition handles it over time.

Then there's the pan and phase views. For instance, in pan view, open a Norah Jones track (say, or anything with a fairly dry, forward vocal). Draw a marquee round the vocal - it's that bright bit in along the middle of the display. Apply effects to reduce the level and add loads of reverb. Voila, a totally rebalanced version of the original with Norah singing from the back of a church.

Again, isolating parts like that is easily done with filters, you are just saving the effort of having to layer it to add reverb. It does sound cool but not worth putting up with Adobe for. I imagine there are also things that SF does better than Audition, like supporting Acid krap [useless to me]. It would largely depend on your priorities. Sometimes I still miss Cool Edit but if Audition has moved on, I probably wouldn't find it a better replacement than SF, and I've had nothing but great experiences with their customer service.

I have to agree with koolkeys. You can't compare AA to a host with a built in sequencer, full MIDI and VSTi support (yet!). I'm not sure it'l ever be such an application but if Adobe did somehow integrate more powerful sequencer and even sampler like features it would make it a killer app for sure.

Really? Most of what an audio editor does requires an entirely different head-space to most of what I use my host for. On the rare occasions that I need one in the other, I have a great live link from ORION which basically gives me SF built-in to ORION's Playlist. I'd hate to see SF move more towards the full host kind of thing as it would be useless duplication of effort. OTOH, putting multi-track features in would make it a much better audio editor. They exist in Vegas so I don't see why it couldn't easily be done.

[HERCsMusic]

The Spectral view features a temporal axis and a frequency axis, so it is more than simply a broad EQ filter. You can choose one particular spot in the audio to 'heal'. Yes, this makes it different and more flexible.
SF, AA (CEP), WL..they were only ever specialised audio editors and certainly when I started to make music 10 years ago, there was a distinct difference between a host like Cubase and an editor like Sound Forge or Cool Edit 96.
The feature bleed over into host territory is only a recent development and probably has confused a lot of people.
They are not full blow hosts. They are specialised audio editors that have added in some basic host functionality. Works in progress really.

[BONES]

The Spectral view features a temporal axis and a frequency axis, so it is more than simply a broad EQ filter. You can choose one particular spot in the audio to 'heal'. Yes, this makes it different and more flexible.

Not really. I mean, Cool Edit 2000 had a Spectral view and I could select a region and apply a filter to it, which is pretty much the same thing, just without the fancy visuals to help you. Given that this is an audio process, I don't see that the visual is necessarily gong to be all that helpful as you will still need to connect what you are seeing to what you hear. I'm not dismissing it, just saying that it is really only offering an incremental improvement and is hardly any kind of killer feature [like the Healing brush in Photoshop which is far less useful in practice than it seems when you first encounter it].

[HERCsMusic]

I'm not fully aware of the 'spectral view' in CEP2000, nor am I saying that AA is the best thing since sliced bread. I'm just clarifying that the Spectral view (much like any spectral domain editing) does feature a temporal axis, so it is more than a simple EQ applied in a host across the whole piece of audio. This makes it a somewhat different beast and useful for sound design at that level of detail if you want it.
How Adobe market it is up to them.
I personally like having a seperate specialised editor for audio (much like a specialised image editor) because that's the way I've always worked and I find the features and tools in such an editor far more comprehensive than simply applying effects on the master bus in Orion, Live etc. To me that's not really detailed enough and does not give me enough control.
But, horses for courses. If you like doing it all in orion then hey, you can save yourself a load of cash.

[Optomadic]

Being able to access the frequency domain directly in spectral view, people that can not grasp the concept of filtering, qing the bandwidth of that filter or how much to apply or subtract can get a clear understanding through the graphic interface.

2ndly (and more importantly), you can do things like create a que that isolates bands cylindrically with varying size, something that is impossible with a normal parametric alone. You gain the ability to curve a que's frequency. You could attempt to do this multiple passes of a conventional filter, but you will never be able to curve a linear que.

[Optomadic]

I'm not dismissing it, just saying that it is really only offering an incremental improvement and is hardly any kind of killer feature [like the Healing brush in Photoshop which is far less useful in practice than it seems when you first encounter it].

Well the difference is you can make a large customized selection over a user determined span of audio. Try doing that with automation a filter and a que (good luck)

I can see where this could be a threat to pro eq heads and engineers (umm the same right ).

Just wait until the next Melodyne comes and then we'l talk about real advances in frequency domain editing

[HERCsMusic]

I'm not dismissing it, just saying that it is really only offering an incremental improvement and is hardly any kind of killer feature [like the Healing brush in Photoshop which is far less useful in practice than it seems when you first encounter it].

Well the difference is you can make a large customized selection over a user determined span of audio. Try doing that with automation a filter and a que (good luck)

I can see where this could be a threat to pro eq heads and engineers (umm the same right ).

Well, umm...AA is a Pro app as far as I know and is used in studios across the world increasingly, so those EQ heads are probably already using the feature

[BONES]

Maybe they are, maybe they aren't. I was very dismissive of Spectral view in CE2k because it didn't speak to me. Maybe it would be worth the effort to learn what it says now but I still think most experienced engineers will still want to use their ears to do that stuff, not their eyes.

I'm not fully aware of the 'spectral view' in CEP2000, nor am I saying that AA is the best thing since sliced bread. I'm just clarifying that the Spectral view (much like any spectral domain editing) does feature a temporal axis, so it is more than a simple EQ applied in a host across the whole piece of audio.
This makes it a somewhat different beast and useful for sound design at that level of detail if you want it.

Sure, but with any audio editor, you can make a selection before applying it anyway. So that becomes your temporal selection and the filter settings, frequency and width, are your spectral selections. If your EQ or filter has a graphical read-out, like the Ozone stuff bundled with SF, you will be getting the same information, just in a more musically relevant way.

But, horses for courses. If you like doing it all in orion then hey, you can save yourself a load of cash.

No, what I was saying is that I can right-click on an audio clip in ORION and edit it in SF [or WaveLab or Audition or whatever], which effectively gives me all that power built-into my host, as it updates after every save.

Being able to access the frequency domain directly in spectral view, people that can not grasp the concept of filtering, qing the bandwidth of that filter or how much to apply or subtract can get a clear understanding through the graphic interface.

Nonsense! The main reason I never really worked with the Spectral view in CE2k was because I had no idea how the graphic really related to what I was hearing. e.g. I can see where my breaths are in the normal waveform view without any effort at all but in Spectral view everything looks too homogeneous. SO it would become a big effort to learn something that is ultimately not very musical and could not be applied anywhere else, as opposed to taking the time to understand filters, which is knowledge you can use everywhere.

2ndly (and more importantly), you can do things like create a que that isolates bands cylindrically with varying size, something that is impossible with a normal parametric alone.

Unless you automate or modulate settings, a-la sidechain or Peak Follower, etc. Again, these are far more musical solutions

Well the difference is you can make a large customized selection over a user determined span of audio. Try doing that with automation a filter and a que (good luck)

Can you feather the selection? If so, what feather radius would amount to an 18dB fall-off? See, that is the essential difference - using this method doesn't teach you anything about audio, nor does it allow experienced engineers and producers to leverage their accumulated knowledge to solve a problem. It seems you just have trouble understanding how things work.

I can see where this could be a threat to pro eq heads and engineers (umm the same right ).

Its not a threat at all, its just a different way of doing something that is already well understood, like an EQ with a graphical display. The issue I see is that it is knowledge that requires a lot of effort to learn, yet has little or no application outside of this tool.that use knowledge most people already have. it has its place, and for newbies it is probably a great way to get decent results quite quickly, but it seems to me to ultimately be a bit of a dead-end, or an extra thing you'll need to learn on top of all the stuff you need to learn anyway that can do the same job.

[koolkeys]

I don't think it has anything to do with understanding something technical or taking the easy way out instead. It's just another way to do something, only in an easier way.

Playing back that great vocal take and didn't realize the neighbor's dog barked in the middle of the chorus, maybe even during a word? This is a situation where it would be quite easy to load up the spectral editing view to find that bark and take it out. Again, just a simple example. And to be honest, it's really hard to explain all the benefits. But suffice it to say that it IS useful. I won't go as far to say that it's impossible to accomplish everything the spectral editor can do without using the spectral editor. But it sure does simplify some things, and there's nothing wrong with that.

Although I do hold to the belief that you can do things in this view that are near impossible to do without it. But that's just an opinion not based on knowing what everyone uses it for.

Brent

[Ozpeter]

The Cool Edit 2000 spectral display allowed you to select vertical slices only, not horizontal. It has virtually nothing to do with the spectral editing in Audition, which allows you to make a number of irregular selections at once and apply whatever effect you want to them in one pass. Whether it would even be possible another way, it would be a whole lot harder.

You cannot isolate a vocal in the centre of a mix with eq. If you could, why have so many DSP engineers worked for so long to achieve it by other means?

And what if you want to isolate the vocal at the centre from the another vocal duetting half-left? Some kind of filter? I don't think so.

[HERCsMusic]

Bones, I believe that you previously stated something about the spectral view not nailing the timeline? I was merely pointing out that it does feature a temporal axis.
It actually is fairly straightforward to learn. It's not something arcane and under certain circumstances it is useful and does make some processes much easier than stacking filters, EQs and visualisers in a regular host.
It sounds complex and no doubt the coding is pretty hefty, but the visual representation of sound can make things like removing noise at particular regions in the sample very easy and intuitive.
No one is suggesting that using ears is redundant in such a view. It's just that you can use eyes and ears together. The whole point of such an editor is more detailed control and this is what it provides. I doubt anyone would disagree with that. If they do, they don't understand what a good audio editor is capable of or under what circumstances you'd use it.
I'm not questioning your workflow. You do whatever is good mate. But for some of us, the spectral view feature, with all bells and whistles, is a boon in certain cases.
It's also a way of doing things that are very difficult to do in any other way. Maybe not a revolution, but more an evolution of already understood processes for some.