Q: Mastering and Recording Enhancement

[Marik]

OK, here is the deal.

I spent about 80 hours of editing classical piano commercial recording.
All is great except one thing, due to the nature of that particular piano register from D to A second octave sounds horroble--very flat and uninteresting, like balalayka. It is out of question to release the recording in this shape.

I am about to lose a few grands, (not talking about other expenses like hall rent, tuner services, union fees, my transportation, etc., which comes to a few other grands, the performer put from his pocket).

If I EQ the top that register gets better, but then the whole recording becomes quite shrill. It is impossible to find a good balance even with 7 band parametric.

The reverb helps, but then all the other registers start "swimming".

It is obvious, that register should be extracted from the rest, and then processed separately.

HOW???

ANY OTHER SUGGESTIONS?

 

[rodabod]

Marik, don't get too worried!

Did you just have a close pair on the piano itself, or did you have any distant mics which you can play with?

I'm inclined to think that if the piano sounded bad in the upper registers, then it's always going to sound bad, but it's hard to say without being able to hear it. Do you have any short samples which highlight the "problem". Maybe you can decribe it better - is the higher register dull?

There are some powerful digital filters which you could probably use along with other processing, but I personally dislike the idea of heavily processing Classical material.

 

[beatpoet]

Let's hear it!

 

[Svart]

try frequency based compression?

 

[NewYorkDave]

Yeah, I was gonna suggest multiband compression, too. It's an effective tool when used wisely and sparingly.

 

[pucho812]

I would suggest some dynamic eq :thumb:

 

[Marik]

Thank you Gentlemen for suggestions!

Unfortunately, I don't own and even don't know how to use a compressor . Many years ago I learnt that it is a no no for classical, so never bothered.

The recording was made with MS and there are a few things I've discovered (I am working with PT)
1) magnifying soundwaves almost to the max. I payed attention that the "S" has a delay of about 3ms. I shifted the "M" forward (by 3ms) and lotsa phase distortions seem to disappear, also resulting in not as "straight into your face sound".
2) On "M" channel the "flattness of sound" is much worse, as the hall sound masks it on "S", so I EQed it so that between 600Hz--1.5KHz the "S" has a peak, and "M" has a dip.
3) To simulate a frequency dependant reverb, I duplicated stereo matrixed track then cut everything below 600Hz, and put additional reverb on that part.
Shifting this track one or couple microseconds gives interesting results, but I still have to play with it.
Yesterday I played with it like 8 hours and it seems there is a chance that the recording can be saved.

Please keep suggestions coming.

I will make and post sound clips later.

Thanks, M

 

[TomWaterman]

Hmm any idea where the 3ms delay came from in the M/S array?
I bet you didn't have your mics three feet apart so is that a converter thing?
I had a random occurance of this once on a PTHD rig but never duplicatd the problem again...killed my drum overheads!!

I take it no amount of EQ has solved this flatness you hear on the M?
Is it dull, lifeless...or boxy, unclear? I know describing sound is hard but I'm not quite following the exact problem tonally...

P.S you can always make a multiband processing tool in PT with EQ3 7Band.
Mult the track three times or however many bands you need and then split each one up into frequency sections using the HPF/LPFs.
Combining all three (or however many you do) sections will return close to the original frequency spectrum, albeit with some nasty phase skewing around the crossover points. Linear phase EQ or nice filters like a Bessel are good.
You can place a PT compressor or any other type of insert device after the filter, even subtle exciters or harmonic distortion to change a frequency band without altering another.

Play with the filter positions and slopes to achieve the most transparent split first.

I tend to do three bands (or however many are needed) and keep an original on another track so I can A/B the summed version with original without any other processing just to be sure the filters are not having more of a negative effect than the original problem.

Once sorted you can even automate the level of one band with PT fader to achieve a manual dynamic EQ..or multiband comp effect.

But I have honestly never needed or wanted to do that kind of processing on a classical recording...it breaks all of the purist rules and then some...so tell no-one you tried it! LOL
-T

 

[SSLtech]

If it's for a professional release, I would suggest a professional solution...

Ordinarily I'd suggest that you try to fix it yourself, but I'm a little concerned about your lack of familiarity with things like frequency-deopendant compression, or dynamic EQ.

Half the skill of using these beasts is a familiarity with what the 'downside' or associated artifacts are. Gaining that familiarity usually comes from making a mistake... it's the best way to learn where the "limit" (no pun intended) is, that way you can stay THIS SIDE of the limit in future.

My personal reccommendation would be Bob Katz at Digital Domain. He's a former professional classical recording engineer who designs, builds and sells equipment, so he's technically VERY familiar, but -most importantly- he has a STUNNING monitoring setup and an excellent set of (VERY experienced) ears.

I'd insist that you at the very least speak to him before making any other step: I recently had him master an album I recorded with the Bach Festival Society, and he did excellent, sympathetic work. He also spotted something that I'd missed slightly: the overall level of eachtune relative to it's absolute dynamic marking... I'd set a couple of quieter tunes a little too hot, and he revised their level within the album downwards slightly so that they sat more comfortably when played next to fortissimo tunes... I had gotten so wrapped up in thinking of it as a collection of single, individual pieces that I'd 'isolated' them in my mind, and that made a big difference, too!

Seriously, -talk to Bob. You'll find his contact info at http://www.digido.com and do mention that Keith Andrews is still eternally grateful for his assistance. He's also very good at soothing nervous clients if they are concerned about any issue.

Keith

 

[rodabod]

I'm not sure it's unreasonable to have a delay between your M and S mics depending on the setup. If it sounds better "corrected" though, then possibly there was a problem there and you have rectified this.

The suggestion of compressors does make me a little anxious - I also don't really think they are a good idea with Classical material, and I also don't see how they really apply to your problem here. I suppose multi-band compression can act like an eq. by reducing levels within a given band, but I don't think it's the sort of thing I would try with Classical, but this is just a personal opinion.

It sounds like you've had some success with eq'ing the separate tracks, so I'd maybe persevere with a little more experimentation and maybe try some different eq's if you have access.

The reverb techniques sound quite clever - perhaps you could apply this technique to just the high end of the M channel if this is where you believe the problem lies. You'd still need to consider summing again to stereo though.

I think Keef's suggestion sounds good (and relieving!) especially coming from an experienced engineer, so give that a shot and see how you get on.

 

[Viitalahde]

Yep, I too recommend a mastering engineer who's intimately familiar with the genre (not a rock guy). Never heard any Bob Katz's work, but I've heard good things about him (although I don't always agree with his book! :razz: ).

A set of new ears is your best bet.

 

[Gold]

Bob Katz would be good.
Michael MacDonald at algorhythms # verizon.net replace # with @
Mark Donahue at http://www.soundmirror.com

All would be good.

 

[TomWaterman]

Oh yeah thats totally the right option - I sort of bypassed the fact that this was already an expensive project you hoped to release.

The multiband idea was FYI only....if you felt like trying something dangerous!!
If you can afford a 2nd opinion by someone very qualified then that seems ideal and appropriate.

Good heads up Keef, Paul etc...
-T

 

[Marik]

[quote author="TomWaterman"]
Is it dull, lifeless...or boxy, unclear? I know describing sound is hard but I'm not quite following the exact problem tonally...

[/quote]

Those who know Steinways well will immediately recognize the problem, which is a dull, lifeless, boxy sound in a middle range. Everything around that range sounds beautufull, but in second octave there is a sense that the sound board just does not support the sound, resulting in a COMPRESSED, lifeless, and ugly sound, with no color or dynamics, whatsoever in a small range between D2 to A2.

The best comparison I could come up with is a violin played with no vibrato in that range, on forte, with no gradations.

[quote author="Viitalahde"]

A set of new ears is your best bet.[/quote]

Well, among other things I have a Doctorate degree in piano performance, and the schools I went through so far include Moscow Conservatory, among others, if it means anything.

As such I believe I have a pretty good idea as for how classical piano should sound like, esp. considering the fact I am a concert pianist myself (well, among others).

Besides, I have some experience in mastering, and among others, successfully remastered for commercial release recordings which were screwed up on Deutsche Grammophone, and discarded as "hopeless", so I beleive I'd be up to task, myself.

At this point I beleive not in the set of new ears, but the set of new unorthodox ideas. That would be the most helpful.

So far the best worked the idea of splitting the top range and treating it separately and then mixing back to the main signal.

Since it is MS I matrixed it "overwidth" stereo, so after duplicating the track and crossing at around 600Hz I can reduce the width by means of "pan", and then put part above 600Hz spread to full, with additional reverb, which adds needed spaciousness and color.

All would be good, except the fact (of course) there is a discrepancy, where the keyboard sounds like splitted into two halves.

Another option is I don't have much modulation, but chorus worked little better, as it would give some "vibrato", making sound more lively and colorful on top.
It works the right direction (i.e. the kind of direction what vibrato on violin would would do), but sounds way too artificial, as it is.

Anything that would give that effect in a natural way (well, as much as it is possible?)

 

[SSLtech]

Well, I can't tell you how that will sound, though I only ever used any chorusing on acoustic pianos in popular music, never on classical/symphonic type stuff...

If you have the tools to do that, by all means try it... though I've a seriously limited expectation of what that approach could yield...

It's not so much for the new 'ears' per se that I recommend a decent mastering engineer, or even the particular 'newness' of the ideas, so much as it is for the experience with available tools and trickery.

Bob in particular is known for his work in "ctepeo" enhancement, and in fact invented and developed hardware and software which does stuff unachieveable by other conventional methods... it works very well on some material, less so on other stuff... you have to try.

Certainly, the impulse responses of M/S must be simultaneous/instantaneously identical, so assuming the M/S delay discrepancy is corrected, anything should be possible... 3ms equates to a difference distance of around 1 meter... -YIKES!!! -How did that happen? :shock:

Anyhow, good luck with the process... try a few things by all means, but I would re-state that a GOOD mastering engineer may very well be able to assist and enhance the music. -A bad one will sell you stuff you don't want, like or need.

The last studio I worked had three pianos: a Yamaha, a Steinway and a Bachmann upright... A studio which I presently also maintain over in California has ELEVEN pianos; at least one in every room, and a couple of rooms with two pianos in... including one concert Steinway, the frame of which has been signed by several VERY well-known players. -I LOVE pianos and their different characters, though I certainly can't claim to have the ear of Marik...

Keith

 

[Viitalahde]

[quote author="Marik"]Well, among other things I have a Doctorate degree in piano performance, and the schools I went through so far include Moscow Conservatory, among others, if it means anything.[/quote]

Uh, I didn't mean it that way.

Just in the way that an experienced ME will too be able to understand the problem and know how to address it without breaking it up.. Or when not to.

 

[SSLtech]

[quote author="Viitalahde"]Or when not to.[/quote]
...A very important part, that bit.

As I tried to suggest earlier, a very valuable part of learning to use ANY tool or technique is to learn its limits.

Keith

 

[Gold]

I know I think about EQ and compression very differently than a mixer does. I wouldn't discount the value of someone who has spent years dealing with problems like yours. As experienced as you are the mindset of a mastering engineer will be different. I know when clients walk in with the list of things they are most concerned about I smile and nod. Then I do the mental translation and distill the problem down to it's essence. Often I work in areas and ways that have nothing to do with what the mixer imagines needs to be done. Yet when I ask if it's sounding more like what they're after they say yes. Most will do a short sample to see if it's what you are after.

 

[Viitalahde]

[quote author="Gold"]I know when clients walk in with the list of things they are most concerned about I smile and nod. Then I do the mental translation and distill the problem down to it's essence.[/quote]

Hehe, I have some learning to do about this but I'm getting better. :wink:

 

[Marik]

OK,

This project was put away for awhile. Now I got back to it and found a few things.
As you might remember, it was recorded as MS, so...

The M sounds fine, the M+S sounds good, but the M-S is what gives that kind of very thin sound with some weird string buzzes, uneveness, etc. I guess there are some kind of phase problems.

So, when I put "M" on one channel and "S" on another--all sounds good--round and beautiful. When I matrix MS into stereo and put "M" in the middle--sounds $hit. If I pan the "M" about half way towards the M+S channel--all is good... except (of course) there is that kinda weird sense the piano is not in the middle of the stage, but somewhere on the side.

What would be suggestion? Is there any way to matrix the damn thing to get a nice stereo without getting into those problems?

Best, M