How to achieve sample accurate drum replacement
- googacky
- pushin' record
- Posts: 295
- Joined: Wed May 07, 2003 2:39 am
- Location: St. Louis, MO
- Contact:
How to achieve sample accurate drum replacement
I'm not one to use drum triggering on a regular basis, but I am on my current project. I'm wanting to blend my sample with the acoustic sound and that's where my problem lies. I'm using Sound Replacer and I'm noticing that it's not sample accurate. I hear the replacement and original phasing to varying degrees. I nudge the sample to be accurate with the original for one hit, but hits down the line won't be quite accurate enough. I figure I could tab to transient and drop in the sample rather than using Sound Replacer, but I like that the program can track dynamics. Am I missing something here or is tab to transient the only way to tighten things up.
-
- pushin' record
- Posts: 257
- Joined: Thu Jun 26, 2008 5:59 pm
- Location: Oak Ridge
- Contact:
- googacky
- pushin' record
- Posts: 295
- Joined: Wed May 07, 2003 2:39 am
- Location: St. Louis, MO
- Contact:
I do have peak align on. It's not 180 degrees out. I'm talking about very small differences here. The samples will come back with a pretty small tolerance--probably twenty samples or less--but it's enough for me to hear. I took the time to tighten up one track by hand and it certainly sounded better. It also took about two and a half hours for one song. I counted the first 100 kick hits and they happened in the first thirty four seconds of the song. I'm not looking forward to doing all of this by hand, but I'm thinking it's inevitable.
-
- gettin' sounds
- Posts: 147
- Joined: Fri Jun 01, 2007 3:13 am
- Location: MA
I trigger without Replacer, like this (using protools):
1) select/highlight the original track
2) consolidate it and save it as a new playlist
3) run beat detective on it. Set the emphasis to find the transient peaks and choose 'separate'.
4) create a new audio track (make sure it's placed above the original)
5) group the 2 tracks
6) select/enable the "tab to transient" box. http://www.audiomidi.com/classroom/prot ... s/fig2.jpg
7) import/drag your drum sample onto the new track and copy it (Ctrl-C)
now tab through the transients and paste the sample at each transient point.
9) mix the real and sample tracks to taste.
It's a lot of tab-paste-tab-paste-tab-paste, but I can usually get through a song in 1-2 minutes.
1) select/highlight the original track
2) consolidate it and save it as a new playlist
3) run beat detective on it. Set the emphasis to find the transient peaks and choose 'separate'.
4) create a new audio track (make sure it's placed above the original)
5) group the 2 tracks
6) select/enable the "tab to transient" box. http://www.audiomidi.com/classroom/prot ... s/fig2.jpg
7) import/drag your drum sample onto the new track and copy it (Ctrl-C)
now tab through the transients and paste the sample at each transient point.
9) mix the real and sample tracks to taste.
It's a lot of tab-paste-tab-paste-tab-paste, but I can usually get through a song in 1-2 minutes.
-
- pluggin' in mics
- Posts: 35
- Joined: Tue Apr 17, 2007 9:17 am
- Location: Central NY
-
- gettin' sounds
- Posts: 147
- Joined: Fri Jun 01, 2007 3:13 am
- Location: MA
-
- pluggin' in mics
- Posts: 35
- Joined: Tue Apr 17, 2007 9:17 am
- Location: Central NY
-
- gettin' sounds
- Posts: 147
- Joined: Fri Jun 01, 2007 3:13 am
- Location: MA
- farview
- tinnitus
- Posts: 1204
- Joined: Tue Aug 31, 2004 1:42 pm
- Location: St. Charles (chicago) IL
- Contact:
Have you tried Drumagog?
Other than that, my only other suggestion is to de-emphesize the attack of the original drumtrack. I'm assuming that you aren't trying to stack sounds to double up the attack, you are either trying to use the sample for a more consistent attack or to get more 'boom' than the origianl track has. EQ out the transient of the sound that you aren't using the transient from.
Or, don't mix them so close to 50/50.
Other than that, my only other suggestion is to de-emphesize the attack of the original drumtrack. I'm assuming that you aren't trying to stack sounds to double up the attack, you are either trying to use the sample for a more consistent attack or to get more 'boom' than the origianl track has. EQ out the transient of the sound that you aren't using the transient from.
Or, don't mix them so close to 50/50.
- googacky
- pushin' record
- Posts: 295
- Joined: Wed May 07, 2003 2:39 am
- Location: St. Louis, MO
- Contact:
It's strange that disabling peak align would help me align the peaks, but I'll play around with that.
My only qualms about the tab to transient version of accomplishing this task is that the sample replaced version won't be following the dynamics of the original take. I worked out a system to get the best of both worlds, but damned if it doesn't take a bunch of time. What I've done is copied my original kick track to another track, tabbed to transients and separated each hit. I then tabbed to transient on my Sound Replacer track (so the dynamics are accounted for) and split up all the hits there. I then nudged the whole sample track down a bit (to account for early triggers), tabbed to each hit on my original kick track and held control while clicking its corresponding replacement, snapping it in line with the original kick. Once this is done, I figure out the sample delay with time adjuster to lock the original and sample in the best phase coherency I can hear. To add an extra level of anal retentiveness, I then went through the track tabbing to each hit and visually checking alignment. Some hits were still five or so samples off, so I'd nudge them back on. I suppose tab to transient has a small amount of variance.
The project I'm working on is pretty crazy and there are a ton of kick hits per track. The first song took me over two hours to finish up and it's about half as long as a couple of the tracks. I compared the original Sound Replacer version with my manually aligned version and the manually aligned version is much more phase coherent. I think there's a lot of keyboard action in my future.
My only qualms about the tab to transient version of accomplishing this task is that the sample replaced version won't be following the dynamics of the original take. I worked out a system to get the best of both worlds, but damned if it doesn't take a bunch of time. What I've done is copied my original kick track to another track, tabbed to transients and separated each hit. I then tabbed to transient on my Sound Replacer track (so the dynamics are accounted for) and split up all the hits there. I then nudged the whole sample track down a bit (to account for early triggers), tabbed to each hit on my original kick track and held control while clicking its corresponding replacement, snapping it in line with the original kick. Once this is done, I figure out the sample delay with time adjuster to lock the original and sample in the best phase coherency I can hear. To add an extra level of anal retentiveness, I then went through the track tabbing to each hit and visually checking alignment. Some hits were still five or so samples off, so I'd nudge them back on. I suppose tab to transient has a small amount of variance.
The project I'm working on is pretty crazy and there are a ton of kick hits per track. The first song took me over two hours to finish up and it's about half as long as a couple of the tracks. I compared the original Sound Replacer version with my manually aligned version and the manually aligned version is much more phase coherent. I think there's a lot of keyboard action in my future.
-
- re-cappin' neve
- Posts: 778
- Joined: Fri Dec 30, 2005 12:07 pm
- Location: Sunnyside Queens, NY
Well, that's Sound Replacer for you. I think the Peak Align feature does some type of stretching to help it align. Although it's always worked better for me with it off. I only heard it during rapid fills though. I dunno.firesine wrote:That's ProTools for ya!googacky wrote:It's strange that disabling peak align would help me align the peaks, but I'll play around with that.
"If there's one ironclad rule of pop history, it's this: The monkey types Hamlet only once."
Who is online
Users browsing this forum: No registered users and 58 guests