Since this is a situation that comes up fairly often, I'm going to go beyond the law itself (which is pretty straightforward, playing the recording is a clear copyright infringement), to get into how you would analyze the pros and cons over deciding whether to use the copyrighted recording or not.
The playing a recording of music subject to a valid U.S. copyright to a crowd at a concert or other public performance (without regard to whether one had to pay for admission to the concert) would be a clear act of copyright infringement in the U.S. without a license from the owner of the recording, if it were performed in the United States. This is clearly not "fair use".
If the recording is played in Brazil, to slightly oversimplify, U.S. law would govern whether a valid copyright exists with regard to that music recording (it is safe to assume that a valid copyright is in place), and Brazilian copyright law would govern what the legal consequences of playing the copyrights recording at a concern without a license from the copyright holder would be.
Realistically, however, it is still extremely likely that playing the copyrighted recording at a concern in Brazil would legally infringe the copyright and subject you to the risk of a civil lawsuit for money damages by the copyright owner.
On the other hand, as a practical matter, the likelihood that the U.S. record label will learn about the fact that you played the recording subject to their copyright in Brazil, and will furthermore take the effort to sue you over it, given the modest amount of money they would get if they won the lawsuit, relative to the costs and risks of bringing the lawsuit and monitoring all concerts in Brazil to learn of these infringements, is pretty low.
I would guess that the odds of being sued by the record company for a performance in Brazil are probably less than 5% unless you are on national or international television at a high profile event, or something like that, and probably less than 25% even then. The lower profile your performance is, and the less likely it is to get on national or international TV or the Internet, the less likely it is that you will be sued. If this is just a small gathering at a low profile bar in a small suburb in Brazil, your odds of being sued are probably well under 1%, but your budget to pay to get a license for the performance from the record company is also probably very small (you may only be getting paid $100-$200 (U.S.) for the performance, or less).
Also, while it is extremely likely that playing this recording would violate Brazilian copyright law, it is also likely that the amount of money that could be awarded to the record company in a lawsuit brought in Brazil under Brazilian copyright law would be significantly less than the amount of money that could be awarded to the record company in a lawsuit brought in the U.S., in the same situation. This is mostly because:
The U.S. has, and Brazil probably has, a specific a minimum amount of statutory damages that can be awarded per intentional infringing performance of a recording of a copyrighted work in lieu of actual lost profits to the record company. But, at current exchange rates (1 Brazilian Real currently equals about 0.27 United States Dollars), the statutory damage amount in the U.S. is probably much larger than the statutory damage amount (if any) in the Brazilian case.
The going royalty rate to play recorded music pursuant to a license in Brazil is probably lower relative to the revenues from the event than the going royalty rate for a similar license in the U.S.
The record company probably gets its attorneys' fees and court costs if it wins the lawsuit in both the U.S. and Brazil. But, the attorneys' fees in the U.S. are likely to be much larger than the attorneys' fees in Brazil, both because lawyers charge more per hour after adjusting for exchange rates in the U.S. than in Brazil, and because the number of hours of work involved in bringing a copyright lawsuit in the U.S. to conclusion is much larger than the number of hours of work involved in bringing a copyright lawsuit in Brazil to conclusion.
In the U.S., the record company would be likely to win $30,000-$50,000 (U.S.) for damages, attorneys' fees and litigation costs combined (or more) if it won a copyright infringement lawsuit in a case like this one. In Brazil, the award to the record company in a basically identical case if it won would probably be much lower (my extremely rough back of napkin estimate would be $5,000 to $15,000 U.S. equivalent).
Usually, however, it would almost always be possible to settle the case for much less than you would have to pay if the record company took the case to trial and won in a negotiated settlement, because the record company would be able to avoid most of the attorneys' fees and litigation costs involved. And, it wouldn't make sense to fight a lawsuit from a record company like this very hard at all once it was brought, because you would almost surely lose. A settlement could easily be 10% to 50% of what you would have had to pay if you took the case to trial and lost and a judgment was imposed upon you.
Of course, even adjusting for exchange rates, the smaller amount of money owed in Brazil would still be a big amount relative to the amount of money made by music bands performing concerts (who also tend to be paid much less per concert in Brazil than they are in the U.S.). And, of course, being sued is also a big hassle and inconvenience in your life and may result in you suffering from bad publicity.
Also, there is almost no chance that you would be prosecuted criminally for infringing on this copyright, in either the U.S. or Brazil. Usually criminal prosecutions are reserved for people who mass produce exact recording copies and sell them to the general public in large numbers (often even deceiving the purchasers regarding the legality of the recording that they are buying).
On average, the risk you face if you do this in Brazil without a license is on the order of $100-$200 or less (U.S.) per performance, and maybe more like $20-$30 per performance for an infringement involving a small gig in a small town.
A license from the record company would probably cost at least $50-$100 (U.S.) per performance plus something like at least $150 to $500 of legal/agent fees to identify the right person to contact at the record company and enter into an agreement to give you the license (although there are economies of scale with the legal/agent fees that wouldn't be much different if you got permission for 1 performance or 20 performances). So, all total, it would probably cost $200 to $600 (U.S.) to use the recording legally at just one performance.
(The large transaction cost involved in hiring someone to negotiate the license is the big problem that the copyright law system has, in general, with trying to use copyrighted material with a license so as not to infringe the copyright. The transaction costs are often much larger than the actual license fee in a case involving a small isolated performance or lower dollar amount, small run of copies.)
Whether you play the copyrighted recording at a concert at a light show, knowing that if you are sued for doing so you will almost certainly lose, rather than not performing it at all, or spending the time and money necessary to get a license to do so from the record company, is a business decision that you will have to make. You make this decision by evaluating the likely damages if you lose, the likelihood that you will be sued, and the cost of getting a license so that you can perform the work legally.
On average, you probably do better infringing the copyright by playing the recording than you do paying the record company for permission, mostly because the likelihood that the record company will sue you in Brazil is so small. But, in the worst case scenario where the record company sues you and refuses to settle the case with you, you could easily be out $15,000 when you could have avoided that judgment by paying $600.
On the other hand, if you had played a recording of your own cover of the music you want to play, instead of the copyrighted recording, you could pretty much eliminate the risk entirely, simply by paying the automatic mandatory licensing fee for the cover that you or the venue is paying anyway.
So, the question is whether the recording is so much better than your own cover of the recorded material that it would be worth facing an average risk of about $200 per performance (or less in the case of a small show off the beaten path) that could be more like $15,000 in a worse case scenario.
The answer might be yes, or it might be no. That is something that you would have to evaluate from a business perspective and as a musician.
If this show could likely be the difference between you being discovered and becoming an A list star group in Brazil, and your group getting shitty concert gigs at ill attended bars for chump change for the next few years, because a talent scout who never records performances, will be coming, you might very well decide that it is worth the risk to provide the best possible performance that you can.
But, if this is a small show before an audience that may not care that much about quality but is likely to have someone who might snitch to a record company in the audience (e.g. a high school student from L.A. who posts a lot of recorded performances to their YouTube channel watched by lots of people in the recording industry, but who isn't very influential in helping bands break out to the big time), it probably isn't worth the risk.
And, as these examples illustrate, it is a tricky decision, because most of the time, the performances most likely to be a huge help to your band in the long run, or that are likely to make a lot of money in that one gig, are also the performances which the record company is most likely to sue you over.
CAVEAT: The dollar amounts and percentages that I quote in this answer are subjective estimates based upon my lifetime experience, what I have read about the recording industry, and what I have learned as a lawyer. But, I haven't done or referred to any formal statistical study to come up with the dollar amounts or the percentages. If anything, my guesses on the dollar amounts involved may tend to be a little low, and my guesses on the percentage risk may tend to be a little high. But, I am comfortable that these numbers are on the right order of magnitude to be realistic estimates and that these numbers are pretty close to being correct relative to each other. In other words, the errors made in any particular estimate are likely to be balanced out be similar errors in the same direction in other estimates, so these estimates are likely to be a better guide to decision making than the accuracy of the individual figures suggests.
Probably the greatest uncertainty is in the amount of transaction costs that would be incurred to obtain a license to perform the work legally. The right person with the right experience might be able to be much more efficient in this regard, but it could also be much more expensive than I estimate to do right.
Also, in real life, no lawsuit is 100% likely to succeed in court. Even in the clearest cases, judges in lawsuits that go to trial, in my experience, get it wrong at least 5%-10% of the time even when the case is very clear cut (this is also not based upon an academic study that is exactly on point, but I do have quite a bit more first hand experience with this and I have read a fair share of all of the academic literature out there on how likely it is for judge's to make incorrect decisions in lawsuits and criminal prosecutions that go to trial).