Will elastic audio ever get fixed?

[SomeOtherGuy]

Originally Posted by midnightrambler

(And boy have I complained about the sound of it on here over the years )

Let's keep on them!

[Top Jimmy]

Originally Posted by SomeOtherGuy

Jimmy - would you mind just explaining in a little more detail about these steps? I'm interested in trying this on some other tracks. If I understand this, you create the bar and beat markers in Beat Detective, but after that I'm not sure how to recreate what you're doing. (I always work in ticks mode, so that wouldn't be something I could turn on afterwards, since it's already on.) Thanks!

Once you've generated bar|beat markers, all you then need to do is enable elastic audio on the audio track you want to quantize, put the audio track in ticks mode rather than sample mode, and then turn the conductor off. The EA audio tracks will then quantize to the manual tempo grid. Commit by turning EA off and putting the track back in sample mode. When I use this method, I always reduce the EA event sensitivity to 0.

[TNM]

Originally Posted by midnightrambler

Thanks for this!

Yes the Logic one really does sound quite a lot better.

Both elastic versions (yours and TNM's) sound pretty bad, they both have that horrible flutter/wobble on the very first G which pretty much gives them away right from the start.

TopJimmy's version is pretty good actually but obviously that required a fair amount more work on his part, and even then it still has a couple of low end artefacts, particularly on the 2nd phrase (the Gm 1st inversion).

All of which just gets me even more depressed about the current state of Elastic audio really. (And boy have I complained about the sound of it on here over the years )

i didn't do anything other than import his file and quantize to 1/8th at 100% transient sensitivity.

I only hear one warble in there that i can manually fix.

Someotherguy forgets to mention that logic fails the automatic test terribly some time too by putting way too many transient markers.

I mean it worked on one piano.
let's do a comparison of 20 different sources including drums/guitar/synth/vocal

Regardless, a bit of work i could get it perfect..

I am testing cubase's now which uses elastique pro and i think it's great.

Nevertheless nothing has the workflow of EA.. it's a joy to actually work with.

I have tried to explain there is absolutely nothing wrong with radius as a stretch algirithm, it's the placement of the markers.. that's all that matters.. even one sample wrong might cause issues. That's why people hand slice.
(or in this case, manually put in the warp markers).

How about Someotherguy gets a simple audio file, a full mix say, 4 bars, and stretch it.. no quantising or anything else, just a pure stretch from say 130 bpm to 100.. then i do exactly the same with x form.. at least this might prove once and for all that X FORM is not the issue and that all of logic's *stretch* modes actually produce more artefacts.

That said, it brings to light that Avid needs to work on their transient auto detection and auto audio quantize. Fair enough. But the issue is NOT X FORM! This is all I have been trying to say!

I am happy to do the file manually, even if it take me half an hour, to get a perfect result to prove it can be done, IF someotherguy wants to hear it.. Or anyone else.. otherwise i won't bother, cause for my uses i have never had an issue.

But i am happy to try if you guys want it.

PS there is also the part where we hear what we want to, placebo ear. I am admitting i suffer from that as much as anyone else. One time I had an EQ plugin that i was bypassing off and on, this was just a few months back, and i could clearly hear the difference every time i clicked the mouse and bypassed or enabled it.. Well guess what, it was disabled the entire time with it's own *internal* bypass! ROFL!

I believe we can move our head 1 inch, and then click a mouse, and that can change everything.. even just the noise of the mouse click can refocus our ears and confuse them.

Obviously i hate Logic with a passion after being tortured by it for 20 years, so i DO want EA to be as good... and we all hear different things.. for example some can't hear the artefacts in the logic version at the end and I swear I can, and i can only hear one artefact in the middle of the avid and he can hear plenty LOL..

Now that others have backed him up, i am more inclined to believe that the error is mine here, hence the offer of doing it manually.

Also, i WILL admit, the first half of the logic one is supremely pristine.. zero artefacts.
Anyway, YMMV and all that, and peace to all!

I am sorry to someotherguy for the tension and conflict we had, i have made a very strict decision as of the other day since that post, to stop all conflict with anyone here from now on. It is just not worth it and so petty when life throws 100x things worse at us .. we don't need to bicker at a forum.

PEACE!

[TNM]

edit, after hearing some results from further tweaking i have new opinions. Did not like the non X form method from Jimmy.. (sorry).

Zplane elastique is the worst of the lot by far.. Apple have overhauled logic's complex algorithm in 10.4. I just did a 50% stretch and it was mostly clear.

So, it might just be that it IS better than x form now, or at least avid's default x form settings.. for some reason x form is eating transients completely when slowing something down,

but if i use the plugin version and tweak it it sounds great.

So.. something is wrong with avi'd auto settings and this is an arbitrary fix! They should do that!

[TNM]

Hi, I just thought I would post Cubase's automatic audio quantise on the piano.

I have no idea whether this is good or bad as i have been unable to discern to a degree that others have..

To me yes after careful listening Logic was the best but pt was perfectly fine, and others completely disagreed with that opinion..

So to me this sounds ok too but i am sure most will hate it.

Nevertheless I thought it was worth posting just to have the cubase comparison in the mix.. One thing i noticed about Cubase's result is the timing itself is super tight.. that I am sure of.. just the quality i am not so sure about. Again it seems ok to me.

All I did was import the file into a blank 48k project set at 100 bpm

I enabled musical mode for the file and set it to 100BPM
enabled hitpoints and quantized to 1/8th with the audiowarp button enabled.

Took all of 5 seconds.. I did not adjust any hitpoints or play around with moving ANYTHING manually.

So this is cubase's auto quantise at 1/8 result.

https://www.dropbox.com/s/w4hcs9raef...ntize.wav?dl=0

[SomeOtherGuy]

Originally Posted by TNM

Hi, I just thought I would post Cubase's automatic audio quantise on the piano.

I have no idea whether this is good or bad as i have been unable to discern to a degree that others have..

To me yes after careful listening Logic was the best but pt was perfectly fine, and others completely disagreed with that opinion..

So to me this sounds ok too but i am sure most will hate it.

Nevertheless I thought it was worth posting just to have the cubase comparison in the mix.. One thing i noticed about Cubase's result is the timing itself is super tight.. that I am sure of.. just the quality i am not so sure about. Again it seems ok to me.

All I did was import the file into a blank 48k project set at 100 bpm

I enabled musical mode for the file and set it to 100BPM
enabled hitpoints and quantized to 1/8th with the audiowarp button enabled.

Took all of 5 seconds.. I did not adjust any hitpoints or play around with moving ANYTHING manually.

So this is cubase's auto quantise at 1/8 result.

https://www.dropbox.com/s/w4hcs9raef...ntize.wav?dl=0

That's pretty good, thanks for doing that TNM. It does have slight transient artifacts at the start of each phrase, but overall it's certainly much better than EA.

[TNM]

you're welcome.. yeah i thought it was one of the better ones also.

I noticed cubase's default transient detection was incredibly accurate vs PT.

That must have had something to do with it.

I can hear the PT EA issue now, there's like a distortion there when the transients hit.

Something is very wrong with Avid's default internal settings for X form and it seems, upon doing some research, that people have been complaining about it for many years.. Surely that would be an arbitrary fix.. to literally just change a couple under the hood parameters..

Anyway, cubase is decent and since that's the DAW i am going to i am more than happy with that. Using it more and more each day.. Cubase is the future.

It has none of Logic's weird bugs, lacks nothing of Logic's features other than Alchemy and inferior FX, but simply adds a lot more on top, and with better performance too (i discovered that logic can only use 50% of whatever logical cores your computer has, when tracks are in live monitor mode.. Cubase can spread across all of them, as does Reaper, and PT as well, as long as you don't use under 128 buffer).

midi routing is 100x easier, it has a full drum midi editor, instant audio engine response time, way better ADC than Logic, WAY better audio editor. VST Expression, inline midi like pro tools, and much much more. Not to mention the chord track which you can literally move all your pre composed chords to (you don't have to use the chord track itself just to compose, you can drag your already done composition), where you can slave samples to follow.. I mean no other DAW does that. I literally import a sample vocal and no more hunting to get it in tune.. Just drag my midi to cubase chord track and tell the vocal to follow that.. some tweaking may be required but really, it's such a powerful program.

In fact, the only thing I feel Logic has any real advantage with these days is the complete package you get for the price.. very appealing to those wanting an all in one. Cubase is 3x the cost.
Logic's new studio instruments, much MUCH better fx plugins, and Alchemy.. You actually need steinberg absolute collection to match Logic's instruments so cubase becomes like $1100 vs $199. That said, since Apple will never fix fundamental issues with Logic like the ridiculous automation being out of time problem, i think it's better to just spend the money (i'd still use PT also btw, over Logic any day.. regardless of EA).

Only really good thing out of the box that Cubase has FX wise is the saturator/tube in the mixer, the new EQ, the multiband transient designer, and halion sonic SE.. But it just whips Logic in almost every other way.. so many advantages. Off topic but had to say it cause i am really happy and think i *finally* might have found my forever DAW. Finally!

Cubase IS the future! Rockin'!

[RichrdC]

Originally Posted by BScout

Have no problem warping grouped tracks together. I pick my master track and put my warp markers there and adjust and all the grouped tracks follow. Internal of the group, they all stay together.

Don't understand the different transients issue.

I haven't ever grouped tracks in EA.. So to do so,is it just a matter of highlighting the tracks you want to include, or do you need to go to the Mix window and actually create the group there? Thanks, richrdc

[Stephen Bond]

Interesting notes from the 2018.12 PT release regarding Elastic Audio.

http://avid.force.com/pkb/articles/r...-Release-Notes

Elastic Audio
Maintaining phase coherency with Elastic Audio pitch processing (PTSW-20602)
When there is Elastic Audio Pitch processing enabled on a track, switching from Polyphonic, Rhythmic, or X-Form to Monophonic or Varispeed can disrupt phase coherency. To preserve phase coherency in this case, be sure to clear all Elastic Audio Pitch processing from the track before switching to the Monophonic or Varispeed algorithm.

Drift in an audio file may occur when using Elastic Audio and the Monophonic or X-Form algorithms (PTSW-33768)
Drift in an audio clip may occur when using elastic audio depending on how much the clips is expanded or compressed when using the Monophonic or X-Form algorithms. If you are using elastic audio with material that contains transient information in it that you would like to keep from drifting, you should use the Polyphonic or Rhythmic algorithms.

Default Input Gain preference is only for Elastic Audio imported at session tempo (PTSW-34096)
The Default Input Gain preference in the Elastic Audio section of the Processing Preferences page is for Elastic Audio Imported at Ses- sion Tempo only. This preference option will not function unless Preview in Context is enabled in Workspace, and/or “Drag and Drop from Desktop Conforms To Session Tempo” is enabled in the Processing Preferences page.

Displayed Timebase is incorrect after importing Elastic Audio tracks (PTSW-46897)
The incorrect timebase is displayed after importing Elastic Audio tracks to existing tracks with opposite timebase. You will need to man- ually select the correct timebase for the affected tracks.

Elastic Audio clips on Playlist lanes are not rendered (PTSW-47311)
In Playlists View, any Elastic Audio clips on Playlist lanes are not rendered (waveform is “off-line”) when changing the track’s Elastic Audio processing from Real-Time to Rendered. Using the object grabber to select any off-line clips, promoting them to the main playlist or copying and pasting the clips in place will render using the selected Elastic Audio algorithm.

Clip Groups containing Elastic Audio are not recognized as Elastic if the Clip Group itself has not had any Elastic functions applied (PTSW-34335)
If a clip group contains Elastic Audio inside it, but no Elastic Audio operations have been performed on the outermost level of the clip group itself, then the clip group will not register as Elastic Audio. When used to create a new track, the track will not automatically be Elastic Audio–enabled, and the elastic audio inside the clip group will be rendered using the default Elastic Audio plug-in for the ses- sion. As a workaround, add a warp marker to the clip group to force it to register as an elastic clip. Or, create an Elastic Audio–enabled track first, then add the clip group to it.

Cannot import Elastic Audio tracks with the Consolidate From Source Media option (PTSW-33894)
Use Copy from Source Media to import Elastic Audio tracks.

Timing inaccuracy in audio files converted from CAF and MP3 with Preview In Context enabled (PTSW-46707)
Audio files converted from CAF and MP3 with Preview In Context enabled can cause timing inaccuracies (gaps in the audio and inner- clip asynchronous transients issues). This is inherent with compressed file formats.

Stephen

[AlexLakis]

Interesting! Where are you getting this information (the link doesn't work)?


Would love to know more!