Timing Discrepancy - Record Auto features

In the Record Window are features such as “Auto Start On Sound Detection” and “Auto Stop after Duration”. I like these wonderful WL features and have come to use them more and more. However, I have always noticed that the resulting file’s duration seldom matches the intended times I had programmed. In the past I just shrugged this off. But I find more and more that computer savvy clients (and internet train spotters) get shaken when exact timings are not seen. I ran some tests this morning to provided factual data:

TEST 1: Auto Stop after Duration = 0:00:20
Pass1 = 0:00:20.225
Pass2 = 0:00:20.317
Pass3 = 0:00:20.201

TEST 2: As above plus Auto Start on Sound Detection = -71dB/200ms (-40dBFS trigger signal)
Pass1 = 0:00:20.540 (measured preroll = 183ms)
Pass2 = 0:00:20.563 (measured preroll = 181ms)
Pass3 = 0:00:20.517 (measured preroll = 198ms)

As you can see, each pass is a different duration from each other and NONE of them match the intended, programmed duration of 0:00:20.000. Very strange! Especially in Test 1, where there is no tolerance involved with signal detection. Exact and repeatable results should occur. It does seem the trigger for Auto Start is reasonably reliable.

So what’s the deal?

This feature does not “cut” the recording length to the programmed duration.
But WaveLab regularly checks the time, and when the duration has been exceeded, recording is ordered to stop. This means the actual recorded length is slightly longer. In what can this be a problem?

In specific cases, this can be an immense problem. The kind of problem where big clients get mega-upset and will pull entire revenue streams from you.

Here is an example: Lets say you have a song with 23 mix/remix versions. Lets say one version is 5:00 long, so I set AutoDuration to 5:00. Wavelab ends up writing a 5:00.501 long file. No biggy, you think, its still basically 5.00, right? The track listing is documented as 5:00 in duration for that version.

Well, sometime later (possibly even years) the Label is looking for that particular mix version for commercial purposes but the timing of the audio on file does not match their label copy. Not matching because their system rounds up the time to 5:01. They try to cross reference against titles but it is misread or there are typos, or some other comedy of errors. In the end they mistakenly send the wrong version and manufacture 100,000 copies only to have a fan blog about the mistake online. Costly and embarrassing! Guess where the finger will point. This is not a hypothetical scenario. I have seen this happen in the real world a number of times over the decades.

I can think of other reasons, but I think the point is made. Mostly, WL’s inaccuracy will pose no problem. But when there is, it is a giant problem. Even if the problem is only a perceived problem (subjective rather than objective) we will feel the pain.

Besides, is there some difficulty in having the recording end exactly on the programmed value?

I think the Auto Record function is meant to automate recording, not to deliver finished files. I hardly use it, but when I do, the result of the automated recording is the start of my job, not the end. BTW, cutting to an exact length should be no problem in the batch processor.

Sure, I take your point. This has been the status quo for production masters: a direct-burned CD, rendered DDP or master BWF.

But as CD’s become less significant and single oriented file masters increase, its awesome to make direct captures, with relevant metadata, and nail it in one step. Elegant and efficient!

The capabilities are already present in the Record Window to accomplish this. They just need to be tweaked for accuracy.

I don’t see the relevance with CD or DDP, but I guess don’t understand your process here. In WL, you are automatically recording (I guess with sound detection) an already finished product from somewhere else of which you already know the exact length?

It is sufficient that I do use and value this feature; further justification is not necessary to continue the topic. What is more to the point is we have another peculiar behavior where WL does not produce an expected result. One which requires the user to triple-check WL’s work and mop up any slop.

case in point:
“Audio-In-Pauses” toggle does not update CD timings thus presenting misleading and incorrect values:

…another case in point:
“Render according to Markers” with plugs present in clips can produce misaligned audio:

“Stop After Duration” produces arbitrary and unrepeatable file lengths:

It’s not as if I am asking for some new outrageous feature. I am simply advocating for pre-existing tools to be tightened up. I am advocating for ACCURACY. I am advocating for WYSIWYG. Why should anyone be against that?

I wasn’t asking for justification, I was simply interested in your workflow. Not anymore. Actually, I couldn’t care less about your workflow. Good luck getting WL behaving to your expectations. Free tip: the easy way is vice versa.