Audio Snap - Useful For What?

I've been trying to use Audio Snap for various things over the last couple years with pretty much no success. At one time I thought I had it working for setting the project tempo based on the original recording and even posted a how-to here about it. The how-to was a long laborious list of things you had to do, most of which required a lot of manual intervention and workarounds, fiddling with markers and manipulating the beat map, etc.
 
After Platinum was released with the audio snap improvements, I tried it again and it seemed to be even worse at getting the transients correct. I posted here about the issues I was having and was informed that audio snap was not really designed for program material and that Live would be a better solution for warping audio. The responses I got were correct, Live worked perfectly and with no intervention. Fair enough.
 
Last week I had a mono rhythm guitar track where the sound was perfect but it had a couple minor timing glitches, a little rushing and dragging here and there. I thought ok, now this is something that audio snap should be able to handle and was probably designed for. I was surprised to find that the transient detection was even worse, with around 50% of the transient markers missing the actual transients, sometimes by up to a 16th note. Not wanting to spend hours on the track, I just dragged it into Live, exported the audio, and imported it back into Sonar. Again, Live got it right with no intervention.
 
The utility of a warping tool is all about being able to bring live audio into a DAW environment and grid the live performance to the project to allow further parts to be added etc. It's pretty rare to want to go the other way and match the project to timing inconsistencies of a live performance.
 
Audio snap might work perfectly for percussion and drum parts but I've never needed to use it for that since I program all my drum parts. However, I recently tried it with the Amen Break and it had similar problems with that as well.
 
What are the use cases that Audio Snap works on? The ones that it just gets right?
So far I'm batting 1000 picking the wrong ones.
 
 
 

BTW there is also another method that can be used in Sonar. The Smart Tool can be used to crop/stretch audio clips with a simple drag move. (I think you hover the cursor over the start/end of a clip and hold down the Alt key to switch to the Timing Tool). What you would do for the task you described is create splits in the audio clip where you want to apply the stretching, engage the Timing Tool (or whatever the Crop/Stretch thingie is called) and stretch the start/ends of the clips to where they are needed. From there you can release the time stretching tool, drag the edges of the clips over each and create a crossfade for a smooth transition and bounce the clips together to remove the splits.
 
The Loop Constructor also has some neat time stretching abilities but probably not quite as well suited for corrections such as this.

aaaand a super useful function of audiosnap is the fact you can merge/lock transients across tracks. This is good for multi mic'd drum recordings. You can do beat correction across all the drum tracks at once to avoid phase issues. I don't know how well other programs deal with this type of thing.
 
AS, in theory, is ultra powerful but definitely requires a lot of patience and knowledge. Simple warping of already mixed material is probably better done in other programs with simpler schemes or within the Loop Constructor or using the Timing Tool.
 
Cheers. I'll shut up now.

Hey guys, thanks for the responses.
It sounds like everyone has had similar experience with it and have given up as well.
Does anyone actually use it in production?
If so, for what?
 

I often use AUDIO SNAP for timing corrections - mostly on the odd note hear and there, eg matching a BASS and KICK drum to get a solid groove
 
Just a note for others - don't forget that the "online" rendering is just a "preview", bounce the corrected section to get a MUCH better quality correction

TomHelvey
Hey guys, thanks for the responses.
It sounds like everyone has had similar experience with it and have given up as well.
Does anyone actually use it in production?
If so, for what?
 

I don't use it that much, but that is because my workflow doesn't require it that often.  Even for things like chopping breakbeats, i have discovered that i would rather chop and map to pads to be triggered than correct the time and lose the swing that gives them the feel that makes them special.
 
I think what frustrates people about audiosnap is similar to what frustrates people with tools like Melodyne.  The sample videos and marketing materials give some the impression that it is a quick fix tool.  The most basic of tools can be time savers, but honestly, the real power in either takes a lot of time and effort to get desired results.  Some ask why bother with this tool if so much work still needs to be done?  i suppose that is a question everyone has to reconcile on their own.  For me , i like having it on hand and occasionally use it. no tool that is in essence a repair tool should be a regularly used feature.  Many times you are better off just doing another take. Fix it in post is a poor strategy.  I think chuck uses audiosnap a lot and has made videos that show practical applications

Maybe itr wasn't chuck.  I know it was on youtube. here is one from karl tha tmay help: https://www.youtube.com/w...o&feature=youtu.be

Someone on the forum posted a really detailed use of audiosnap with a multi-track drum session that was well done but i can't recall who.

I use AudioSnap for only a few things: lining up tight vocal harmonies, correcting bass guitar sloppiness, lining up hand percussion, and for generating a tempo map from a live drum track. The latter works well if you have a separate kick or snare mike to key off of. It's the only scenario in which AudioSnap is a simple "click here for magic" operation.
 
Bass works best if you only use initial transients to quantize a note start to the beat. I'll typically delete all but the first detected transient, so I'm essentially just moving whole notes around with little or no stretching. Sometimes it's easier to just split-and-nudge instead of bringing out AudioSnap, if there are only a couple notes that need tweaking.
 
The vocal alignment procedure is labor-intensive because you have fewer reliable transients to lock onto. I typically delete 50-80% of the detected transients and often move the remaining ones visually or insert my own. This works OK because for vocals you're usually only trying to line up consonants, starts of phrases and durations of held notes. It's rarely worth the effort for a lead vocal alone, but it pays off for multi-part harmonies.
 
At least in these limited applications, AudioSnap is a wonderful convenience that really works. I just wouldn't trust it on anything that didn't have well-defined transients (e.g. strummed guitar).
 

mixmkr
I've tried it several times trying to create a tempo map to a freely played instrument part...so I could add MIDI drums, etc.  It has always been a failure and trying to figure it out each time was laborious to say the least.

This has been my primary purpose and frustration as well. I have not sought to stretch or make something fit to tempo, but rather SONAR's tempo map match the audio. For this I have sort of given up with AS and fallen back on using Shift-M, which is significantly more forgiving to me.
 
One thing that bothers me lately is I have popped open a few low-end VSTs recently and the bpm in them is spot on loading a file, even for a commercial mix... it is almost like the beat detection algorithm has EQ baked into it to isolate frequency bands. I am not sure how AS works, but using "non-percussive" often falls apart on me, when other programs/VSTs seem to be fine with an entire mix (things not even doing any AS functions even).
 
Then again, it may be the lack of patience with AS that really gets me.

Good question, actually. The way I see it is this. The most important element in getting AudioSnap to work perfectly is to have a perfect transient map. Whether AS can do this or not depends on the track itself. At one extreme, if it's a four-on-the-floor kick drum, AS will create a perfect transient map. At the other extreme, if it's program material the transient map will be pretty much useless.
 
Now, I freely admit I also have a philosophical prejudice, which is not to correct something unless it SOUNDS wrong. Therefore I listen to a track. If something sounds wrong, I fix it manually by either grabbing an existing transient and moving it manually, or inserting a transient and moving it manually. No beat/transient corrector can decide what sounds wrong, only what it thinks looks wrong. 
 
However, I have used AS in "auto quantize" mode. For example, I was doing a remix of a dance tune that had a recorded electronic kick drum. I wanted to layer a drum loop played by a human, but when the two kicks hit, there was often flamming. I copied the human drum part, filtered out the highs, used the transient shaper to give real strong transients, and applied AS to create a pool. The transient map was perfect, and it was painless to then snap the electronic kick to the pool and sync up with the human drums.
 
There was also a rhythm guitar part where I wanted to have metronomic precision because it was being played against electronic drums. So I "auto-corrected" the part with AS. When I played it back, there were two places where AS really screwed up but the rest was fine. So I hit "undo," split the clips with the problems, applied AS to the rest, then fixed the two problem areas manually.
 
So the bottom line is this. AS can be a "magic" solution, but most of the time you'll need to at least edit the transient map. How much you'll need to edit it depends on the track. Rather than take the time to create a perfect transient map and auto-correct, I'll just use the Transients edit filter, fix what needs to be fixed, and be done with it. This can take anywhere from a few seconds to several minutes per track. 
 
Finally, it took me a while to develop AS chops. I didn't use it for a long time because it was just too much hassle and I never got really good results. However, in the course of using it, I've learned what it does well, how to compensate for what it doesn't do well, and developed my AS skills to where I use it frequently and (I think) effectively.
 
Does that help?
 
 
 
 

I use it. I think it's very good. BUT I think it doesn't do much well automatically. You have to go in and do a lot of manual work to get good results. For example, forget about using the quantise function unless you're willing to zoom in and fine tune the transient positions. That's got better in the latest release, but it's still not perfect.
 
Audio Snap's real weakness is in workflow design. It's full of obscure gotchas and opaque procedures.
 
 

My main use for audiosnap is extracting tempos from freely played material. Most of my songs start life on acoustic guitar or sometimes piano and then I use that as the framework for the song to build on and then disregard the scratch track once I've got the timing and feel out of it.
 
I've done much the same when recording bands but usually take the timing from the kick and/or snare and go from there.
 
I'd also +1 on the fact that it's not a simple one click tool but a bit of time understanding how it works and the tools available was time well spent for me.

You need a degree in AS to master it .. Personally I haven't used it in so long , it's just like ACT is a waste of time and workflow killer . Add this transient.delete that one , use it like this or that , seriously !