8E8s

great. Can't wait to get home and try it.

Yup. Heard great things about Serato's algo too.

Also, Izotope's "Radius" code is an option. Check it out. It's just an algorithm that you install and can select within time and pitch machine. It's significantly better than Logic's stock algorithms. It's marketed as a product for tempo manipulation of audio. But essentially decoupling tempo and pitch is what this algorithm is all about so it does pitch manipulation wonderfully. I've no idea how it compares in price to third party full-app solutions like antares and celemony.

Probably as David suggests your comp threshold is the key here. Is your kicks level remaining above the threshold for some time? because if you can get the kick to just tickle the threshold with its initial attack that might solve the problem. What if you used the compressor's side chain filter to key in on just the initial attack of the kick, just that click which has a lot around 1KHz?

Maybe try making the ducking more subtle by just ducking selective low frequencies from the bass. Then is might not be so noticeable if the ducking lasts a but too long. At the same time, the gate plug you'll use with it's attack release, hold and hysterisis might offer a little more control. Try this: bus the bass to an aux. On that aux insert a gain plug and invert the waveform. Next, insert a noise gate plug with sidechain set to the kick's track. Leave it set for full reduction. Set threshold etc. When the kick hits and opens the gate, the inverted bass waveform will partially (or totally - depending on output level) cancel the signal from the bass track. To make the effect more subtle, insert a linear phase eq and use the filters to carve out the frequencies you don't want removed from the bass. Just remember that it's all opposite because of the phase flip. Like if you want the mids and highs in the bass track to remain then roll them off using a steep filter in the eq. More about this technique here: http://logicprohelp.com/viewtopic.php?t=18921

That's interesting. I frequently have a comp on one or more tracks in most projects. The folder isn't open, though.

Sounds very good.

Lately this has been happening almost anytime I leave Logic running and leave my macbook pro for 8 or so hours. The computer is set to never sleep, is left plugged in, but does have its screen set to turn off after one hour of inactivity. No other programs are running when this happens, except of course the plethora of processes revealed in activity monitor. No new software has been installed and no system upgrades have taken place. Yet in the last month or so this problem has become more and more frequent. Any similar experiences and/or ideas to fix?

Well this is the coolest trick in the mixing book! Works beautifully to keep piano from masking male vocals. thanks. One variant I just tried is to use the mutipressor on the aux track that has the inverted waveform instead of using the linear phase eq. Just solo one or two of the frequency bands in the mutipressor, set the crossovers accordingly and adjust levels for each band. Make sure no compression or expansion is happening.

Ok so if you route your 8 tracks to a bus then the easiest thing to do might be to simply put Logic's automation mode into Touch for the associated Aux track that the Bus feeds. Play the song and as you approach your bad section, use the mouse to change the eq filter in real time, gradually increasing the amount of cut and then returning it to default later. Normally I would just punch one of those band bypass buttons in and out on the channel eq plug but you indicated that you wanted a smooth change so as not to suddenly vary the tonality of the chimes. When you return the automation mode on the Aux track to READ Logic will play back your automation steps. The other option that was mentioned before was not sending those 8 tracks to a common bus but instead automating one of them as mentioned above, then switching the pointer tool to the automation select tool (hit the escape key to see the available tools). Hit "A" to open automation lanes on the track and use the automation select tool to select the section where you see the automation to the filter. Copy it. Then paste that into the automation lanes of the eight other tracks in the same place on the timeline. You can probably just option-drag the selected automation section using the pointer tool down to the other tracks. That method's using 8 eq plugs though even if only for a few seconds. I've had good luck in the past removing low frequency spikes and "plosives" etc. by just drawing them out even if there is superimposed audio in the waveform. For example in the sample editor if you can see your high freq. chime waveform and then suddenly in the middle of it there is a longer wavelength wave (the cough) superimposed with it, you can sometimes literally redraw the high frequency (small wavelength) tone without it. See the attached .png file for a picture of what I mean. Did you also consider finding some other good chimes audio from another section of the song and copying it to the cough region, crossfading etc. That would be my first option.

Instead of automating an eq plug, did you try drawing out the cough in the sample editor?

Ok, this problem is solved. It's actually quite easy. I used audio to midi groove template to force the midi tracks to quantize to the audio overhead tracks. Simple.

Yeah, that might be the easiest option. Lose some flexibility in mixing but skip all the editing. I can keep the original tracks and re bounce later with the snare up or down and recompress if the drum mix isn't sitting well.

Yes, thanks. What's different in these cases is that all the drum tracks are sample aligned and stay that way throughout the work. But when I tempo compress the audio by 8 bpm in Time and Pitch Machine I lose this. The way the transients move out of alignment on the different tracks is rather extreme and unpredictable - which I guess I have to expect from even a good time compression algorithm run separately on different tracks. So any audio-to-score work I do really has to be all done from one track and that should be the track I'm keeping as audio - in my case, the drum overheads, which I don't need or want to quantize. What's evident here is the way to divide one track by the transients of the other. Which means I can get relatively close to isolating just the snares from the overheads track (using the snare track whose transients are "close enough" after time compression) and then generate a midi score from the snares on the overheads track. . . or just the same for the kicks etc. COo.l Thanks, 8's

Actually, it's only one file. I just uploaded a zipped version because I didn't know if the .pst would go.