well i don't have any problem with abruptly ending snare tails, but yeah... making your Xfade value .5 of your trigger pad value is a good rule of thumb with BD. no "bad splice points" should ever happen. doesn't for me.
The 'I love ProTools' thread
- Thread starter Ermz
- Start date
I found that it tended to misalign the trigger point from the transient some of the time. At worst it would trigger on something that seemed entirely arbitrary. There was something 'hands on' about manual splicing that I always dug. Sure, that's a self-serving excuse, and you can probably see right through it, but it's all I have to hold onto, man 
In all seriousness though until I get the work that warrants a HD system, I'm happy taking extra time to edit because the work itself doesn't demand the workflow of a HD system. You get what you pay for, and if the clients pay enough to buy me a HD system, they'll get the boons associated with it. I'm happy to work whichever way is available.
In all seriousness though until I get the work that warrants a HD system, I'm happy taking extra time to edit because the work itself doesn't demand the workflow of a HD system. You get what you pay for, and if the clients pay enough to buy me a HD system, they'll get the boons associated with it. I'm happy to work whichever way is available.
Lasse Lammert
HCAF Blitzkrieg
AudioGeekZine
arsehole know-it-all
I've seen some people do the whole song at once with Beat Detective and with very few problems. I haven't had much luck with that.
I used to do it 4-8 bars at a time because I thought it would be more accurate, but I was making more mistakes, putting things in the wrong spot a 16th early or late etc.
Now I do about 24 bars or a song section at a time, collection mode. You've got to make sure BD is seeing your selection right, show the trigger times (with PT 8 the trigger colors blend in with the region colors too much IMO, I've got to have the trigger times shown) then look through, most of the time things will be on x|xx|000 or x|xx|480. Look for anything strange.
I used to do it 4-8 bars at a time because I thought it would be more accurate, but I was making more mistakes, putting things in the wrong spot a 16th early or late etc.
Now I do about 24 bars or a song section at a time, collection mode. You've got to make sure BD is seeing your selection right, show the trigger times (with PT 8 the trigger colors blend in with the region colors too much IMO, I've got to have the trigger times shown) then look through, most of the time things will be on x|xx|000 or x|xx|480. Look for anything strange.
it sounds like you're really just missing out on a critical stage of BD... auditing the automatic results, using varying zooms.I found that it tended to misalign the trigger point from the transient some of the time. At worst it would trigger on something that seemed entirely arbitrary. There was something 'hands on' about manual splicing that I always dug. Sure, that's a self-serving excuse, and you can probably see right through it, but it's all I have to hold onto, man
on the first audit pass you just look for false and missing beat triggers... zoom in to where you can view a few to several bars at a time, depending on the density of the playing, and delete and add them quickly as you run through that pass... this goes very fast normally if you have chosen the best detection analysis algorithm and sensitivity to start with... just scroll through using your scroll wheel and add and delete as necessary... very quick. make certain you have beat triggers on every transient that should have one, and no "falsies" remaining.
Then zoom a bit further in for the second pass.. make sure that "Show Trigger Time" is checked in the Region Separation pane, and that the radial button for Sub-Beats is indicated.... now scroll-wheel through and make sure the trigger times... the beat/sub-beat that BD will move each transient to, or closer to, depending on the Strength setting you have chosen in the Region Conform pane... is correct for each beat trigger. This goes much quicker than you'd think, especially once you're used to what the subdivisions will read as for each type of rhythm. Correct each one as necessary by simply double-clicking on the beat trigger and entering the correct value. Your success here will depend highly upon your remembering to Capture Selection and choose the appropriate subdivision in the "Contains:" drop-down in the Selection Pane for each section you work on.
to get tweaky you can add a third step to the audit process: zoom very far in and move through each trigger via the Scroll Next button in BD... and manually moving beat triggers to fine-tune their location on the transient as desired.
then just make sure you've set a Trigger Pad of around 10-20 ms in the Region Separation pane and hit Separate. move to the Region Conform pane, set your desired Strength, and hit conform... if you've taken Audit step #2 everything should move to the correct location, or closer to it as per your Strength selection. then move to the Edit Smoothing pane, make sure the Fill and Crossfade radial button is indicated, and set a Crossfade time for one half the trigger pad value you set earlier in the Separation pane. Hit Smooth.
Boom! done... no "bad splices".
Awesome guide James. One thing that always puzzled me was that sometimes BD didn't allow me to add my own beat markers too close to pre-existing ones. So for instance say I have two beats, a 16th apart, I can add a beat marker directly on one, but the second wont allow me to move any closer than halfway through the waveform. Why does BD sometimes impose a limitation like this? I kept thinking I was missing some vital feature or setting, but I never found it.
AudioGeekZine
arsehole know-it-all
+1 to all that.
In addition, the nudge functions are your friend. Especially nudge within region boundaries. 200 samples works for me most of the time.
I usually just conform and fill gaps. I'll crossfade using batch fades for the whole song. That takes a long time but the fades slow down the editing otherwise.
In addition, the nudge functions are your friend. Especially nudge within region boundaries. 200 samples works for me most of the time.
I usually just conform and fill gaps. I'll crossfade using batch fades for the whole song. That takes a long time but the fades slow down the editing otherwise.
AudioGeekZine
arsehole know-it-all
How is that possible? I select what I need edited. It's not like I can get that wrong. BD can interpret it wrong, but I certainly know what I'm selecting when I do it
first, i corrected a gang of typos already now.. so re-read it for added clarity.
no, that sometimes just happens... i deal with this by reselecting a smaller section... or just manually fixing these small discrepencies after the BD process.
screw all that.. the normal BD workflow, for me, is to consolidate after you've corrected various sections of the song... not each and every time you Fill and Crossfade, but before the amount of cuts and fades can bog down your system... this works best in my experience... AGZ's way leaves with a zillion cuts in a long complex metal tune... would bog down any system to one degree or another after a certain point.
^ The second point is why I set up a RAID0 partition running the two quickest 7,200RPM drives I could find. To be able to increase the amount of leway before needing to consolidate is nice. That is one thing that really rocks about Elastic Audio. The crux of the pressure is on the CPU, which can handle it, rather than our grossly outdated storage technology. Doesn't help that EA commonly sounds like arse on drums. Honestly how do you get it to NOT screw with phase!?
Lasse Lammert
HCAF Blitzkrieg
How is that possible? I select what I need edited. It's not like I can get that wrong. BD can interpret it wrong, but I certainly know what I'm selecting when I do it
make sure to select whole bars and hit "capture selection"
the "whole bars" though have to be "as intended" not by the grid prior to BD
AudioGeekZine
arsehole know-it-all
That's what I'm finding. I've even put half the tracks on another drive, but as soon as you put fades in (which goes on the same drive as the project file) it slows down again.
Unfortunately I can't really consolidate everything because the clients not in the room and I don't want to commit the editing until they've given the OK. I don't know the songs well so things like, was that a sloppy 16th or triplets? needs to be undoable.
There are often 17,000 fade files in a song that's fully chopped and fixed.
Making a cut or adjusting a fade can take up to 15 seconds to respond.
I'm considering a solid state drive as a solution.
Ermz, no RAID for PT. Not supported.
AudioGeekZine
arsehole know-it-all
Sometimes having a gap before and after seems to help if it won't capture the selection right. Sometimes not having a gap.
So if you selection in the edit window is from 100|1 to 108|1 sometimes BD will see it as 99|4 to 107|4
If you get a popup asking if you want to move the regions to a new bar|beat location that's a good indication that your selection was fucked to begin with.
sometimes it's still fine and you can just use the region quantize function instead.
Hardware RAID works fine. It's done at a controller level which is for all intents and purposes invisible to the OS. I've edited drums on it just yesterday. Worked fine. Digi don't 'officially' support it, but there's no need to.
Also, I use playlists to get around the system bog down when editing. I will duplicate my chopped section (usually half a song), then go back to the 'comped' playlist and just consolidate there.
Also, you can always keep an unedited parallel copy of the regions on another playlist so you can always revert back to the raw tracks to re-edit from. There's really no reason not to consolidate as you go (unless your RAID set-up negates the need to a large degree)
Also, I use playlists to get around the system bog down when editing. I will duplicate my chopped section (usually half a song), then go back to the 'comped' playlist and just consolidate there.
Also, you can always keep an unedited parallel copy of the regions on another playlist so you can always revert back to the raw tracks to re-edit from. There's really no reason not to consolidate as you go (unless your RAID set-up negates the need to a large degree
)using a RAID0 array right now.... PT sees it as one 500gB drive.. it's transparent to PT.
i think you're thinking too hard about this... lol.
just choose the bars you intend to edit, in slip mode... going from the exact transient of the first bar, be it snare or kick or whatever, to the transient of whichever one it is on the next bar after the section you want to edit... then separate that selection use Apple +E or B in Keyboard Focus.... then open BD... then capture, set sub-beat resolution, analyze, audit & correct, separate, confrom, and finally smooth... move on.
and Ermz... don't bother with EA on drums... leave it alone. stick to BD.
AudioGeekZine
arsehole know-it-all
just trying to explain to Ermz the importance of having the selection captured correctly by BD. if your edit selection and captured selection don't match, when you conform the regions will go all over the place if you don't manually change every trigger time. With the right selection, you don't have to change any of them.
You've been doing this for years longer than us, How long would it take you do do a whole song? Average metal drumming complexity?
You've been doing this for years longer than us, How long would it take you do do a whole song? Average metal drumming complexity?
I understand, but I already do all this. I've BD'd enough drum tracks to understand what it likes and doesn't like to see. As James mentioned above it's just a quirk with BD with no way to resolve from within BD itself unless one compromises by making different selections. Conform presents similar issues at times with the way it snaps. I usually have more success using 'Quantize', as you do.
huh? you can't control how PT initially interprets your selection if you accurately choose your region to edit.... there's no way to do without changing your selection, and that's just extra, unnecessary work. just audit.. if the first marker is off, just correct it... the others after it should.. and do.. auto-correct themselves once you do that. i see where you're getting with this.. but you have to create gaps in the audio to do what you are saying... that's sucks on it's own, nevermind the extra steps you are creating.
i'm kinda tweaky, so on average i'd say 3 hours-ish... but that depends quite subjectively on what you mean by "whole song" in terms of length, and how you interpret "average complexity"
