Issues with Auto-Align Post 2.3.1 (ARA) in Nuendo 14: Clip Gain Duplication, Direct Offline Processing Loss, Lane Conflicts?

Auto-Align Post
1

Currently using Auto-Align Post 2.3.1 (ARA version) with Nuendo 14, and encountering the following issues. I’d appreciate any guidance regarding proper workflow or known limitations.

■Observed Issues

  1. Direct Offline Processing (DOP) effects such as noise reduction (e.g., iZotope RX) sometimes get lost or deactivated, especially after copying, trimming, or moving events between lanes.
  2. Clip Gain seems to be applied twice in some cases.
    For example, when Clip Gain is adjusted before using Auto-Align Post 2, the gain is sometimes “stacked” or applied again after AAP2 processing, resulting in unexpected loudness.
  3. Interactions with lane modes (show lanes, comping, etc.) may cause DOP or AAP2 processing to disappear or behave inconsistently.

■Questions

  • Are these known limitations or expected behaviors?
  • Is there a recommended workflow (e.g., AAP2 first → then DOP → then gain adjustments)?
  • What precautions can I take to avoid losing processing after edits?
2

Hi @Bensan ,

That’s defintley seems as unexpected behavior.

In order to properly triage it, it’ll be great if you can send a minimal project to reproduce such issues. We can then review if it’s an issue on our end that we need to fix or a general issue that needs to be reviewed by Steinberg.

I guess a stable workflow would be, first applying AAP first and make extension permenant and only then DOP and gain.

Hope this makes sense.
Tal.

3

Hi Tal,

Thanks for your reply.

I tried to reproduce the “clip gain duplication” issue using a small test project with minimal audio data, but I couldn’t get it to happen reliably — it seems to occur randomly. In some sessions, clip gain seems to be applied twice after AAP2 processing, but in others it works fine. So far, I haven’t found a consistent way to trigger it.

Also, sometimes in the AAP2 (ARA) panel, I see the message “Reference track is invalid.” Despite that, the alignment appears to work correctly and the phase seems to be aligned.

Here’s a link to a test project (zipped) that you can check:

There’s another issue I’d like to mention:

If I follow the workflow:

  1. Cut and edit clips
  2. Apply RX noise reduction via Direct Offline Processing
  3. Then use AAP2 to align the phase

…in some cases, especially with ambient or background noise (as opposed to strong dialogue), the noise reduction becomes partially ineffective at the clip boundaries. It seems like RX processing is being undone or ignored just at the edges. My guess is that when there’s not enough audio energy for AAP2 to detect phase accurately, it may affect how the underlying DOP behaves at those points.

While this isn’t a big problem in short-form projects, most of my work involves long-form content like films and TV shows — often over an hour in length. In those projects, every time I reopen the session, AAP2 needs to recheck all clips, which can be extremely time-consuming and greatly reduces the usability benefit of using AAP2 in the first place.

I’d really love to hear how others manage their workflow with AAP2 in long-form productions. Any tips or best practices would be appreciated.

Thanks again,
Bensan

4

Thanks @Bensan ,
Our support will overview the provided files.

“Reference track is invalid” - This message suggests that some edit such as the clip or reference track content was moved, trimmed or simply removed from the ARA (eg you remove the AAP instance on the reference track) resulting discrepencies from the time the alignment took place.

It doesn’t change the aligned content. but… since AAP no longer able to assure the relations between the reference track and the clip. It shows this warning.

There’s another issue I’d like to mention:

If I follow the workflow:

  1. Cut and edit clips
  2. Apply RX noise reduction via Direct Offline Processing
  3. Then use AAP2 to align the phase

If I may suggest, applying AAP before other processes would be best.

Most noise reduction tools use filters which can alter phase and gating.
So, it’s always better to first do AAP on the source audio rather than the processed one.

I’m not sure what do you mean by “It seems like RX processing is being undone or ignored just at the edges” but AAP can’t really undo noise reduction.

Tal.

6

Hi Tal,

Thanks for your reply.

Do you mean that AAP needs to be applied first?

However, since Direct Offline Processing cannot be used after applying AAP, I think it is necessary to cancel AAP once in order to apply DOP.

Or does everyone render all clips at the point when AAP is applied?

However, after processing with AAP, there may be cases where you need to edit the crossfade position of the clip or adjust the clip gain level.

In that case, I think you have to remove the AAP process in ARA once, adjust the fade position, change the clip gain, or take a different approach to noise processing, and then apply the AAP process again, but if you’ve already rendered, that’s impossible.

If that’s the case, I think you need to create a new lane with AAP applied while keeping the original clip in lane mode. What do you all think about this?

I noticed that switching lanes at this point seems to cause many errors.

As far as I can tell, if you don’t render while the AAP processing is active, and instead adjust the crossfade position of the clip, change the clip gain level, or select a clip in another lane in lane mode, the AAP processing is removed, which is indicated by the message “Reference track is invalid.”

Then, at some point, the AAP processing is applied again, and the clip gain that has been changed is further changed, resulting in double processing.

If I reselect the clip with AAP applied and reapply AAP via ARA, there should be no issues, correct? Or do I need to first select “No Extension” from the Extension section to remove the ARA application, then reapply AAP?

The most reliable method would be to apply AAP processing to all tracks where you want to align the phase, then render them at that point to minimize errors.

However, this method prevents you from adjusting crossfade positions or redoing partial noise processing.

Therefore, I think it’s best to keep unprocessed lanes that haven’t had AAP applied separate.

How do others keep unprocessed clips? Are they not using lane mode?

“It seems like RX processing is being undone or ignored just at the edges”

means that after applying noise processing to the entire clip with RX and then aligning the phase with AAP, the first and last parts of the clip, which have not been processed for noise, are moved and appear.

This is a common error that occurs when connecting clips with a crossfade.

If there’s a better method, please let me know.

Thanks,

Bensan

7

Hi Bensan,

Or does everyone render all clips at the point when AAP is applied?

Ideally yes. meaning you’d like to do noise reduction / cleaning on the material.

In that case, I think you have to remove the AAP process in ARA once, adjust the fade position, change the clip gain, or take a different approach to noise processing, and then apply the AAP process again, but if you’ve already rendered, that’s impossible.

Indeed, unless Nuendo enabled different approach. Lately (with 2.3.1) we’ve added internal handles to allow having the extra range without re-aligning, but for making it persistent, the host must render the extra area.

If that’s the case, I think you need to create a new lane with AAP applied while keeping the original clip in lane mode. What do you all think about this?

That sounds the most reasonable solution as it is for now, new lane with AAP’ed clips while having originals in another lane.

If I reselect the clip with AAP applied and reapply AAP via ARA, there should be no issues, correct? Or do I need to first select “No Extension” from the Extension section to remove the ARA application, then reapply AAP?

Currently, only single ARA instance is supported. just keep in mind that “re-applying” AAP on material will actually re-align it so it might sound different.

If I may suggest, maybe you’ll find AAP-Nuendo users at Steinberg’s forum

But once we have some way of reproducing the issue we can then evaluate it better, fix it on our end if possible or contact Steinberg to better tackle this.

Tal