I don't think you're going to EQ this out.
Any compression will only make it worse.
You could possibly apply a noise gate (fast attack/slow rel.) to remove the parts between vocal expressions.

Personally I'd go with my "SM58 of plugins" IQ4Gui and attempt to knock down cymbals etc. whenever they get loud.

How much bleed and is it just between vocal phrases or constant (constant => get your face in that mic, low shelving EQ were made for a reason)

I recently had a very similar situation. I've been watching this topic to see if anybody had the magic bullet for this. This way my experience:

The drummer was behind plexiglass and the vocal mic nearest the drums (AKG 414 xls, set to cardioid pattern 10ft. away), did a great job rejecting almost all of the drums. Much to my surprise, the other vocal mic (AKG C3000, cardioid 40ft. away) was full of drums. I think this mic was in a spot that happened to be at a gathering point for the reflected drum sounds. I listened intently to the mic nearest the drums and was satisfied with the separation, but should have listened just as intently to the far mic - I should have known better.

No broadband noise reduction could strip out drums without chewing up the vocal. The BNR was tolerable with the kick, but the female vocals were degraded too much when I tried to work out the snare and cymbals.

Gating the drums out of the vocal channels was never an option, it would only draw attention to the drums coming and going with every phrase.

Any EQ aggressive enough to affect the drums was destroying the tone of the vocal. Framing the vocal within a high-pass and low-pass filter were the only half-acceptable results. But it still left an unnatural boxy ambience to the drums - as you would expect from a mic so far away + limiting the bandwidth so much - not good.

In the end, I called the vocalists back in to the same venue to overdub the vocals. They were a very young band, so they were very willing to indulge me, since it wasn't going to cost them anything.

That may not be an option for you, but I couldn't find any way to get a good mix with all the drums present in the vocals. And I couldn't find a way to strip the drums out of the vocal tracks in PT. The rest of the tracks were good, so it seemed a shame to sacrifice them because of my oversight and the location not lending itself to better isolation. Handing the client a nasty sounding mix wasn't an option, so I bit the bullet and re-recorded what needed re-recorded. Lesson learned.

If someone out there has a solution, please enlighten us.

You nailed the solution. Re-record it.

Hours and hours and hours of editing will leave you with a nasty product. The human voice and the ambiance of the room all fall across the drums voice and it's ambiance. Cutting either out of the other will leave you with crud.

Use the original as a scratch/reference vocal, and re-record it.

The 1st full band recording I did was 24 songs in two nights. Slam-bam. Bootcamp for me. They insisted on doing it live.
While we did end up ODing a line or two for performance reasons, it was a nightmare to mix. Here's a few high(low)lights:

1. HOURS and HOURS of drawing automation on the vocals. Pull it back a little when not singing to keep all the weird drum noise down, but not so much so that there's big swoops in levels as the vocals come in. And you have to make the envelope just right so you don't notice when the extra drums DO come in.
2. EQ all over the place. Duplicating tracks, compressing one, 'verbing the other (still a technique I use).
3. When we did OD vocals, in order to make up for the lack of drum noise, I had to cut a section of the drum room mic that matched the same measures as the vocal line. And then copy it to a new track. And then line it up, and play w/ levels until it was really hard to determine that the line had been overdubbed.
4. In all, I spent 10+ hours a week mixing this, over 4 months.

Granted, I record in a church, and it's nearly impossible to put up another usable mic when there's drums. Your situation is likely not so dramatic.
And don't get me wrong - for my first full band, I'm thrilled w/ the results.
The recording sounds really good, and I learned an enormous amount.

Still.
Save yourself the headache. Re-record it.
(Lesson learned here)

soapfloats:
I know. I know. *pats reassuringly*

Hello everyone,
thanks for your responses. I know it is impossible now to isolate the vocal from the bleed.

This project is my good friend that recently past away so unfortunately re-recording his voice is not an option. I will just have to leave the track the way it is. His sister is coming from Germany to take care of the legalities. I was hoping to make him sound just a touch more in tune on a couple spots.

I am happy with what I have, however I was hoping to use pitch correction on a few parts that needed a little help.

The problem comes when the pitch correction changes occurs it also changes the pitch of the acoustic guitar bleed. This in turn comes through on the mix and sounds terrible.

Soloution: Keep it simple stupid.

I think I will post the song later, I am sure many will enjoy it . Bret

Don't give up entirely, Jammster There are a few things you can do to fix it up a touch.

While it's not ideal, if you got good tracks otherwise of the vox and guitar, try some frequency slotting. Find the most crucial ranges for the vocal, and cut everything else. Do the opposite for the guitar - cut the ranges where the vocal sits, and leave the rest intact.
That might make the vox track clean enough to allow the pitch correction to give less unpleasant results.

If it's a word/line that's somewhere else in the song, try copying that and pasting it over the bad note.

While I haven't done any pitch correction (something in my musical nature recoils at the idea of "auto-tune"), I did spend a lot of time playing with "less than ideal" tracks - I'm sure there's a few other things you can do if you have the time to really dig in and experiment.

Hopefully someone else has some ideas, too.

Good luck!

Forgot - none of the above will be likely to give you a real fix (true iso).
You may be able to come up w/ something that you can live with, however.

Try duping the track and doing different things to each. I played around w/ dupes and compression, EQ, reverb, and delay. It's really a useful tool.

No magic bullet, but on re-reading dvdhawk's post, wanted to clarify a bit.

For some inspiration, below is a link to one of those songs, complete w/ this forum's critiques. Yes, it did take 4 months. But I'm interested to see if you think it turned out as well as I did.

{old-link-removed}

In between the vocal phrases, the vox track was reading @ -15dBu regularly. Just from the drums in the room!
And yes, I did try to iso the singer as much as possible - big heavy glass doors draped in blankets, wood "ceiling" - basically a glorified childhood "fort". I'm wiser now.

soapfloats,

I would like to thank you for your inspirational words. You know your right! Something in me just likes to hear the notes better, but why? I agree with you about using pitch correction, I'm still just trying this for fun mainly.

This is not my song, and often when I engineer and produce for friends I want to "fix" them to sound better than they are. My good friend would not agree about the use of pitch correction. Its just that he was a smoker and did not have great pitch some of the time and often the use of the pitch correction plugin will make his recordings much better on long sustaining notes.

I just posted the song for critique here. Its listed as Blue Forever:

(Dead Link Removed)

I can post the track in question as well, but I a pretty sure I am going to just leave the mix as is for now.

I plan to post more songs from this project for fun in the future.
Thanks for all the input,
Bret

By the way, forgot to mention that the track in question is the Root or Lead Vocal(Lower pitch voice) that was initially recorded as a scratch track. At the time we recorded it we felt no need to overdub a new root vocal.

soapfloats,

one more thing. I really liked the song in your link. Did you ever post the finished product? I could not see the result that your asking about.

Thanks again,
Bret

Hi Bret,

We're all sorry to hear you lost a good friend. I'm sure you want this to sound as good as possible, but I'm sure his sister will cherish any recording you might have - imperfections and all.

Best of luck.

You can try inverting the guitar track and mixing it with the vocal.

Copy the guitar track, and invert it.
Solo the inverted guitar and the vocal.
Loop on a part that is just guitar bleed, and gradually bringing in the inverted track. With an RMS volume meter on the master bus you should see the volume go down, and then back up. You want to find that sweet spot that is the lowest volume.

This trick has worked for me to some degree, but it's not perfect. If you have time, you could mess with EQ on the inverted track to match the tone of the guitar bleed to the guitar track.