Eric Cardenas Posted August 7, 2013 Share Posted August 7, 2013 • When set to Inter Sample Peak Detection the Adaptive Limiter fails to report the delay it is inducing to be able to look ahead. • If ISPD is not engaged the Out Ceiling parameter ceases to function as a ceiling parameter. Quote Link to comment Share on other sites More sharing options...
Eriksimon Posted August 7, 2013 Share Posted August 7, 2013 • If ISPD is not engaged the Out Ceiling parameter ceases to function as a ceiling parameter. I get different results with this 4 measures of test bounce, with the only setting changed is the Out ceiling, Audioleak shows a ± 2 dB difference in level, which to me means it works: (just now I noticed that the optimal lookahead "advice" has changed from 15 ms (which you cannot set) to 0 ms (which you also cannot set... ), where I did not do anything but bounce, set the ceiling to a different value, and bounce again...) Quote Link to comment Share on other sites More sharing options...
Eric Cardenas Posted August 8, 2013 Author Share Posted August 8, 2013 Thanks for checking this Erik! Now, on the instance with a - 2dB, try changing the Gain level to 5 dB and make your measurements. Do you still get a -2 reading? Quote Link to comment Share on other sites More sharing options...
Eriksimon Posted August 8, 2013 Share Posted August 8, 2013 ...on the instance with a - 2dB, try changing the Gain level to 5 dB and make your measurements.Do you still get a -2 reading? Yep, also with 12 dB gain. Although from the pics it looks like some peak values are staying put, and I'd say the difference is slighly less than 2 dB. But there is a clear difference. 5 dB 12 dB Quote Link to comment Share on other sites More sharing options...
Eric Cardenas Posted August 8, 2013 Author Share Posted August 8, 2013 Thanks for checking again but that is not consistent with my findings. I took a loop and routed it to 2 Aux channels. Both have an Adaptive limiter inserted. The second Aux channel has a polarity invert. These are the settings: ------------ Input Scale -------- Gain ------ Out Ceiling ----------- Aux 1 ------- 0.0 dB ---------- 5.0 dB ------- -2.0 dB ---------------- Aux 2 ------- 0.0 dB ---------- 3.0 dB -------- 0.0 dB ---------------- I have inserted a Multi Meter on the output and the two signals are nulling perfectly. I have also provided the project for others to test. AD_Limiter_Broken.zip Quote Link to comment Share on other sites More sharing options...
Eriksimon Posted August 8, 2013 Share Posted August 8, 2013 Haven't looked at your project yet - but wasn't there some very serious bug involving auxes? Quote Link to comment Share on other sites More sharing options...
Eric Cardenas Posted August 8, 2013 Author Share Posted August 8, 2013 wasn't there some very serious bug involving auxes? Yes! But it will not come into play in this case. Quote Link to comment Share on other sites More sharing options...
Eriksimon Posted August 8, 2013 Share Posted August 8, 2013 (edited) Yep, silent. No argueing that - the mystery deepens... As far as I can tell I did not make any methodic mistake in my setup - so I should have gotten identical AudioLeak pics from these bounces... First I'll disable the Gain plugin and bounce each track soloed, see what happens... OK, I solo-bounced both "routes", dropped them on Audioleak and this is what I got, different looking waveforms - actually, after taking a more careful look, the waveforms are pretty much identical, it's only the longterm equivalent (leq) level lines (yellow and white) that differ between the pics... But, I then imported the two bounces, routed them to an aux, flipped the phase on one of 'm and hear: they null . I look at both waveforms in Logics' Sample editor, they do look identical. So the bug is confirmed and I am left with the question why on earth Audioleak is telling me that these files are different...? :? Edited August 8, 2013 by Eriksimon Quote Link to comment Share on other sites More sharing options...
Eric Cardenas Posted August 8, 2013 Author Share Posted August 8, 2013 Yep, silent. No argueing that - the mystery deepens... As far as I can tell I did not make any methodic mistake in my setup - so I should have gotten identical AudioLeak pics from these bounces... First I'll disable the Gain and bounce each track soloed, see what happens... I'll see if I get different resuls if I set this up from an empty template... Thanks for the confirmation! Your last two screen shots with the 12 dB increase seem kind of telling now that I look closer at them. Try comparing the + 12 dB Gain, -2 dB Output Ceiling file with a new one set to + 10 dB, 0 dB Output Ceiling. Quote Link to comment Share on other sites More sharing options...
Eriksimon Posted August 8, 2013 Share Posted August 8, 2013 12 -2, 10 0, tried that, yep, also nulling. Now you try and raise the Input Scale on both AdLs (stop Logic playback, set both AdLimiters to same input scale, start playback). The silence starts to give way from around ± 1.5 dB and up! Is the Adaptive part of the Limiting finally kicking in? Quote Link to comment Share on other sites More sharing options...
Eric Cardenas Posted August 8, 2013 Author Share Posted August 8, 2013 Now you try No. what's happening here is that you are clipping the input and then raising the clipped signal differently. That's why it start's to fall apart. Also Remove DC kick's in differently depending on the level. Quote Link to comment Share on other sites More sharing options...
Eriksimon Posted August 8, 2013 Share Posted August 8, 2013 I'm not sure why a clipped signal would not nul against an equally clipped signal? One thing is clear though: the AdLimiter is not behaving at all orderly and as a good plugin should ! It basically ignores the ceiling, and seems to treat it as yet another gain parameter, not as the absolute maximum it is supposed to be... I did some other extreme settings, and even heavily clipping the input also nulls against not clipping it: I'm running out of emoticons. Quote Link to comment Share on other sites More sharing options...
Eric Cardenas Posted August 8, 2013 Author Share Posted August 8, 2013 I'm not sure why a clipped signal would not nul against an equally clipped signal? They are not equally clipped due to the Output Ceiling parameter processing in-between the Input and the Gain parameter Quote Link to comment Share on other sites More sharing options...
Eric Cardenas Posted August 8, 2013 Author Share Posted August 8, 2013 I did some other extreme settings, and even heavily clipping the input also nulls against not clipping it: I can not confirm this. Quote Link to comment Share on other sites More sharing options...
Eriksimon Posted August 8, 2013 Share Posted August 8, 2013 (edited) I did some other extreme settings, and even heavily clipping the input also nulls against not clipping it: I can not confirm this. Are you sure? Can you try this project (yours, resaved with the pictured settings - it still nulls here.) Later edit: I made an error here. I lowered the audio tracks' fader and forgot that I did. 00 AD Limiter Broken II.zip Edited August 8, 2013 by Eriksimon Quote Link to comment Share on other sites More sharing options...
Eric Cardenas Posted August 8, 2013 Author Share Posted August 8, 2013 You are not clipping the Input. Quote Link to comment Share on other sites More sharing options...
Eriksimon Posted August 8, 2013 Share Posted August 8, 2013 Ouch! I lowered the loops' track fader and forgot about it , my bad, clipping it makes it no longer nulls. Did I say "oops" already? Bedtime... Quote Link to comment Share on other sites More sharing options...
Eric Cardenas Posted August 8, 2013 Author Share Posted August 8, 2013 Thanks for confirming the bugs! Bedtime... Sleep tight. http://i231.photobucket.com/albums/ee28/pinkdolphin11/smilies/7_11_116.gif Quote Link to comment Share on other sites More sharing options...
Eric Cardenas Posted August 20, 2013 Author Share Posted August 20, 2013 The Out Ceiling bug is fixed in Logic Pro 10.0.2. The non-reported plug-in latency is still a problem with ISPD. Quote Link to comment Share on other sites More sharing options...
Eric Cardenas Posted January 20, 2016 Author Share Posted January 20, 2016 There is still non-compensated latency when using the True Peak detection (ISPD equivalent). Logic Pro X 10.2.1. Quote Link to comment Share on other sites More sharing options...
Eric Cardenas Posted March 8, 2016 Author Share Posted March 8, 2016 And... continues in Logic Pro X 10.2.2. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.