Select For Production Amp Distribution: Fill & Download for Free

What is a funny joke that only an engineer would understand?Salesman: I have a customer that is asking how reliable our widget is.Engineer: 5% per year.Salesman: What does that mean?Engineer: That means that after a year in service, 5% of them fail.Salesman: That’s horrible. I can’t tell them that.Engineer: (After thinking for ten seconds), how about we tell them that the mean time between failures is 20 years?Salesman: Yeah. That’s great. That’s what I’ll tell them.Sadly, that wasn’t a joke...Engineer: We need to tighten up the final test procedure, we’re shipping crap.Manager: Our procedures are fine. Prove that we’re shipping crap.Engineer: I just pulled 100 units out of shipping and 10 were bad.Manager: That just proves my point. Those 10 didn’t get shipped, did they?Sadly, that wasn’t a joke...Engineer: We need to put a hold on shipping thingamajigs. Due to a defective part, the thingamajigs burst into flames after 72 hours of use. We can get replacement parts in 3 days.Manager: It’s the last day of the quarter. If we don’t ship I’ll miss my forecast.Engineer: This is potentially life threatening. We can’t ship.Manager: Let’s ship them and issue an immediate recall.Sadly, that wasn’t a joke...From before the invention of the internet …Engineer: The vice president of human resources is opening all our mail to be sure that no one is getting job offers from headhunters. Its adding up to 5 working days to getting technical data. It is really slowing this rush, bet the company project. It will probably cause thirty day slip in the delivery date.Manager: Mail handling is not a proper subject for engineers to discuss.Sadly, that wasn’t a joke...Manager: We need some IP (intellectual property) on the whatchamacallit.Engineer: We don’t have any IP. We buy the unit ready built from China and put our name on it.Salesman: I’ve got an idea. It comes with computer software. Let’s patent the idea of “File Open”.Engineer: You can’t file for a patent on the concept of file open.Salesman: Why not? Nobody else has patented it. Patents are awarded based on first to file. We’ll be the first to file. We’ll get rich, I’ll be a VP.Engineer: We would have to have actually invented it.Salesman; Who’s to say we didn’t invent it.Engineer: When you apply for a patent, you have to declare an oath that you are the original inventor. You would issue an oath saying that they were the original inventor of “File Open”?Salesman: Me. I invented.Engineer: No you didn’t.Salesman: Yes I did.Engineer: “File Open” has been around since before you were born. You didn’t invent it. Claiming so would be fraud. Up to five years imprisonment.Salesman: Yiiiiieeee! You’re not being a team player. We really wanted that patent and you stopped us.Engineer: (rolls eyes and starts looking at job openings).Sadly, that wasn’t a joke, either..When Six-Sigma became the quality program de jour, the quality organization had (usually one per division) black-belts and green belts. The black belt was highly trained and probably had at least a master’s degree and probably a Ph.D. The green belts had bachelor’s degrees. If we were in the Air Force, that would correspond roughly to Majors and Second Lieutenants. These guys were supposed to get our quality level up to six-sigma. Their mission was to seek out and solve quality programs. Here is a typical interaction between a green-belt and a design engineer.Green Belt: “You have to minimize the RMS deviation. It says so in the book.”Engineer: “My designs meet spec 100% of the time if the parts are in spec.”Green Belt: “So, lets optimize it for bad parts.”Engineer: “It is economically better to buy good parts.”Green Belt: “Well, let me see if I can do better.”Engineer: “OK.”Green Belt: “What is the sigma of the resistance value?”Engineer: “I don’t know, but when I buy 1% resisters, I always get 1% resisters.”Green Belt: “OK, well take 1% as the six-sigma deviation. So, it looks like this (draws a narrow bell curve).”Engineer: ”No. At the factory that makes the resisters, sigma is about 3%. They select out the 1% resisters and sell them to me. They sell the others as 5% resisters.”Green Belt: “OK, then a uniform distribution from .99 to 1.01.”Engineer: “No, they sort out the 0.5% resisters and sell them to someone else. It looks like this (two-hump distribution). We call it a dog-bone distribution.”Green Belt: “That’s not in Minitab.”Engineer: “But that is what we get.”Green Belt: “You need to buy 5% resisters so we can optimize it.”Engineer: “Our cost for parts and labor is about $1000. There are about 300 resisters; they have an average cost of 1.2 cents. Which ones do you want to optimize?”Green Belt goes away and comes back with a book. “See, this is how it’s done.”Green Belt shows engineer an example in his book of an NPN common emitter amplifier that was optimized by design of experiments. There are many equations and derivatives and graphs.Engineer: “In what sense is that circuit optimal?”Green Belt: “Minimum output variation.”Engineer: “I see. The output is zero with no variation. Also, it invariably fails to amplify.”Green Belt: “You are putting too much emphasis on amplifying. Minimum output variation is the correct thing to optimize. It says so right here.”Engineer: “You have the collector attached to a negative supply. Not only will it not work like that, we don’t have a negative supply.”Green Belt: “My analysis shows that we need a negative supply.”Engineer: “Also, your bias circuit draws 12,000 amps (it used sub-milliohm resistors) and 144 kilowatts of power. This is a handheld, battery powered, instrument and you have added a power supply the size of a washing machine that will need its own dedicated branch circuit for powering and will need water cooling.”Green Belt: “You’re just looking for reasons not to like it.”Engineer: “Well, then, maybe you should contact the resister manufactures and see if they will share their process parameters with you so will have the sigma of the resistance values.”Green Belt: “You don’t have it?”Engineer: “No.”Green Belt: “No wonder you don’t understand this.”Engineer: “You know what, there is a real crying need right now to optimize the width of the packing tape we use on the box.”Green Belt: “It’s not amenable to this analysis.”Engineer: “Why not?”Green Belt: “There are too many variables.”Engineer: “Maybe the division Six Sigma Black-Belt can help you there. That’s his purpose, isn’t it? Why don’t you talk to him about it?”He went away and the next time the engineer saw the green-belt, he asked the green-belt what he was going to do about that packing tape.Green Belt: “The Black-Belt said we couldn’t change anything that was already in production because of ISO-9000.”Engineer: “Who’d of thought that?”That also, was not a joke.

Lots of great stuff. I'm mostly familiar with systems research (Amp Lab and Par Lab), but some of that is just ridiculous. Not to say that there isn't any cool non-systems research; this is just what I'm the most familiar with.Communication-Avoiding AlgorithmsThe basic idea is simple: these days, doing some computation on the CPU is much faster than accessing memory. So we need to update our fundamental theory of algorithms to account for communication. "Communication" covers accessing caches, memory, other cores or even information over the network.The group (BeBOP) has developed a theory for modelling the amount of communication an algorithm does. This allows them to do things like prove lower bounds for the number of memory accesses needed to solve a given problem. This is a much better theoretical model of an algorithm's runtime than normal big-O because it accounts for the increased time cost of accessing memory, as well as the "shape" of the memory (multi-level caches, network latency...).This theory allows the group to develop algorithms close to the communication bound. This has resulted in speedups for, well, pretty much all linear algebra algorithms. Yes, really, all of them. Including dense matrix multiply and Strassen's. Another example speedup was in breadth-first search which let its author place 18th on the Graph 500 rankings with a single node. Boy, did that get noticed: now the majority of the top entrants use his algorithm.Performance improvement from "communication-avoiding parallel Strassens" (CAPS), from Grey Ballard's dissertation talk. That's an improvement for dense matrix multiplication!Many of these algorithms are improvements because they perform asymptotically less communication. This group is producing asymptotic improvements to some of the most fundamental algorithms in CS.(Yeah, the logo is pretty weak.)Apache Spark™ is a framework for running large distributed computations developed at the AMPLab. It's like Hadoop but better in literally every way. And I'm not be facetious either: it's not a trade-off, it's strictly better. It's:fast: up to 100x faster in memory and 10x faster on diskflexible: the core abstraction (Resilient Distributed Datasets or RDDs) is more expressive than map reduce, allowing more complicated processing like streaming computationseasy: Spark has a sane interface in Scala, Java and Python including support for interactive development. I've used Hadoop and its interface is... hideous.On top of the framework itself, the Amp Lab has developed some interesting frameworks for more specific tasks:Shark: distributed SQL engineGraph X: efficient large-scale graph processingBlink DB: a database the can approximate SQL queries with bounded errors, making it much faster over large amounts of dataSpark is especially impressive in having significant industry buy-in almost immediately after being published. Companies are already using Spark commercially, with more considering it. The project has also spawned a startup: Databricks.Here's a nice (if slightly sensationalist) article about it: Welcome to Berkeley: Where Hadoop isn't nearly fast enough.SEJITSSelective Embedded Just-In-Time Specialization (SEJITS) is a set of techniques for enabling scientists and domain experts (who are not necessarily expert programmers) to easily write high-performance parallel code. It is based on designing domain-specific languages (DSLs) that embody a set of common parallel programming patterns the group identified.The idea is to combine "productive" languages like Python with highly optimized parallel kernels for core algorithms. The SEJITS framework specializes high-level code at runtime to be efficiently executed. The system supports a specific subset of the host language and produces code that is specific to the general algorithmic task performed and the specific parallel architecture the code is running on.A particularly interesting feature of the system is that the specializer is implemented in the host high-level language. This makes it easy to extend the SEJITS system with new specializers for different tasks or architectures after the fact.Here's an implementation of SEJITS: ucb-sejits/ctree.Program SynthesisProgram synthesis involves automatically generating code according to some specification. It's a special kind of automatic programming. Essentially, this involves searching through the space of possible programs. Of course, these spaces are mind-bogglingly large (I wrote an answer about this), so we have to be pretty clever about how we search through them.There are two fundamental ideas developed here for program synthesis: CEGIS and synthesis by sketching. This research is mostly led by Ras Bodik.CEGISCounter-example guided inductive synthesis (CEGIS) is a general algorithm for efficiently synthesizing a general-purpose program. The goal is to synthesize a program that matches a logical specification: for example, to synthesize a program which has identical behavior as a slower program or a program in another language.In practice, it resolves around two insights. The first is using some way of synthesizing a program given a set of input-output pairs: this can be efficiently done with modern Satisfiability Modulo Theories (SMT) solvers. Given an SMT solver, we can synthesize a program that conforms to a particular set of inputs and outputs. But how do we go from there to fully implementing the logical specification? The next part of the trick is using the SMT solver for verification; if our program does not match our specification on some input, the SMT solver returns this input. This gives us another input/output pair which we can plug back into the synthesis system.The CEGIS algorithm, from Armando Solar-Lezama's PhD thesis.SketchingThe second big idea is synthesis by sketching. For certain kinds of code, the synthesizer is much better than a human. However, many other things like boilerplate and the general structure of an algorithm are very difficult for the solver but basically trivial for a programmer. We want some easy way for the programmer to provide insight for the solver.Enter sketching. The idea is actually really simple: we let the programmer sketch out the skeleton of the code, leaving holes for the synthesizer to fill in. For example, if you're writing a transpose function, you might know that you need to use the shufps SIMD instruction and two arrays, but not know any of the specifics. You could write a sketch like this:int[16] transpose_simd(int[16] M) implements transpose {  int[16] S = 0, T = 0;   repeat (??) S[??::4] = shufps(M[??::4], M[??::4], ??);  repeat (??) T[??::4] = shufps(S[??::4], S[??::4], ??);  return T; } The synthesizer would then automatically fill in the ?? holes for you, giving you a result like this:int[16] transpose_simd(int[16] M) implements transpose {  int[16] = S = 0, T = 0;   S[4::4] = shufps(M[6::4], M[2::4], 11001000b);  S[0::4] = shufps(M[11::4], M[6::4], 10010110b);  S[12::4] = shufps(M[0::4], M[2::4], 10001101b);  S[8::4] = shufps(M[8::4], M[12::4], 11010111b);  T[4::4] = shufps(M[11::4], M[1::4], 10111100b);  T[12::4] = shufps(M[3::4], M[8::4], 11000011b);  T[8::4] = shufps(M[4::4], M[9::4], 11100010b);  T[0::4] = shufps(M[12::4], M[0::4], 10110100b);  return T; } That would have been a pain to do by hand! I wrote more about program synthesis in general here.My section on program synthesis is the longest because that's also what I've been working on. So I'll take the chance to plug that project as well: Chlorophyll, a synthesis-aided compiler for low-level spatial architectures. The idea is to show how we can use program synthesis to easily implement an optimizing compiler for a really weird architecture (GreenArrays). The compiler also uses synthesis techniques to automatically partition and distribute code over multiple cores. I think it's pretty neat.Parallel Web BrowsersFinally, something of an honorary note for a project from the recent past: parallel web browsers. In particular, there was a focus on efficiently implementing a parallel layout engine using attribute grammars and program synthesis.I don't think this project is active any more, but it's interesting because it heavily influence Mozilla's Servo project. This included several of the students from the Berkeley team interning at Mozilla. I think it's neat to see research like this being put into more active development for a real project like this.

NO, artists don't get paid if they are with a record label, The record label lends them money that is to be paid back if/when the artist makes it Suppose that a music label gives a band a $250,000 advance to record an album. The label agrees to do so in return for 90% of the sales. In addition, the label will specify certain standards for production of the album, for example, which studios the band will engage. The label may even hold the advance and make all disbursements on the band's behalf, ensuring the funds are used exactly as agreed. In other words, the $250,000 advance is not simply pocketed by the band -- it is to be spent on album production and the band's reasonable expenses during production. The album is recorded, and sells 200,000* copies at $10 each, yielding $2m. The record company takes 90% of this as agreed, leaving the band with $200,000 of their own. This is the situation without recoupment.With recoupment, the label advances the band $250,000 as before. As before, the band (or the label, on the band's behalf) spends substantially all of the advance to produce an album meeting the label's specifications. The album again sells 200,000 copies at $10 each, yielding $2m. The record company takes 100% of the first $250,000 in sales (recouping the $250,000 advance to the band), and 90% of all further sales, as agreed. The band's net is effectively reduced to 0% of the first $250,000 in sales and 10% of the further $1.75m, leaving them with a total of $175,000.*Recoupment deals primarily involve new artists, or those without a proven sales history, so selling 200,000 copies would be fairly unusual. More typically, 10,000 or fewer copies are sold by a new artist. In this more typical case, the label recoups only $100,000 of the $250,000 advance, along with other non-recoupable expenses incurred in getting the album out into the world.---------------------------------------------------------------------------Here is article is from the early ’90s, and originally appeared in Maximum Rock ‘n’ Roll magazine. While some of the information and figures listed here are dated, it is still a useful and informative article. And no, we don’t know how to reach Steve Albini.The Problem With Music by Steve AlbiniWhenever I talk to a band who are about to sign with a major label, I always end up thinking of them in a particular context. I imagine a trench, about four feet wide and five feet deep, maybe sixty yards long, filled with runny, decaying shit. I imagine these people, some of them good friends, some of them barely acquaintances, at one end of this trench. I also imagine a faceless industry lackey at the other end holding a fountain pen and a contract waiting to be signed. Nobody can see what’s printed on the contract. It’s too far away, and besides, the shit stench is making everybody’s eyes water. The lackey shouts to everybody that the first one to swim the trench gets to sign the contract. Everybody dives in the trench and they struggle furiously to get to the other end. Two people arrive simultaneously and begin wrestling furiously, clawing each other and dunking each other under the shit. Eventually, one of them capitulates, and there’s only one contestant left. He reaches for the pen, but the Lackey says “Actually, I think you need a little more development. Swim again, please. Backstroke”. And he does of course.Every major label involved in the hunt for new bands now has on staff a high-profile point man, an “A & R” rep who can present a comfortable face to any prospective band. The initials stand for “Artist and Repertoire.” because historically, the A & R staff would select artists to record music that they had also selected, out of an available pool of each. This is still the case, though not openly. These guys are universally young [about the same age as the bands being wooed], and nowadays they always have some obvious underground rock credibility flag they can wave.Lyle Preslar, former guitarist for Minor Threat, is one of them. Terry Tolkin, former NY independent booking agent and assistant manager at Touch and Go is one of them. Al Smith, former soundman at CBGB is one of them. Mike Gitter, former editor of XXX fanzine and contributor to Rip, Kerrang and other lowbrow rags is one of them. Many of the annoying turds who used to staff college radio stations are in their ranks as well. There are several reasons A & R scouts are always young. The explanation usually copped-to is that the scout will be “hip to the current musical “scene.” A more important reason is that the bands will intuitively trust someone they think is a peer, and who speaks fondly of the same formative rock and roll experiences. The A & R person is the first person to make contact with the band, and as such is the first person to promise them the moon. Who better to promise them the moon than an idealistic young turk who expects to be calling the shots in a few years, and who has had no previous experience with a big record company. Hell, he’s as naive as the band he’s duping. When he tells them no one will interfere in their creative process, he probably even believes it. When he sits down with the band for the first time, over a plate of angel hair pasta, he can tell them with all sincerity that when they sign with company X, they’re really signing with him and he’s on their side. Remember that great gig I saw you at in ’85? Didn’t we have a blast. By now all rock bands are wise enough to be suspicious of music industry scum. There is a pervasive caricature in popular culture of a portly, middle aged ex-hipster talking a mile-a-minute, using outdated jargon and calling everybody “baby.” After meeting “their” A & R guy, the band will say to themselves and everyone else, “He’s not like a record company guy at all! He’s like one of us.” And they will be right. That’s one of the reasons he was hired.These A & R guys are not allowed to write contracts. What they do is present the band with a letter of intent, or “deal memo,” which loosely states some terms, and affirms that the band will sign with the label once a contract has been agreed on. The spookiest thing about this harmless sounding little memo, is that it is, for all legal purposes, a binding document. That is, once the band signs it, they are under obligation to conclude a deal with the label. If the label presents them with a contract that the band don’t want to sign, all the label has to do is wait. There are a hundred other bands willing to sign the exact same contract, so the label is in a position of strength. These letters never have any terms of expiration, so the band remain bound by the deal memo until a contract is signed, no matter how long that takes. The band cannot sign to another laborer or even put out its own material unless they are released from their agreement, which never happens. Make no mistake about it: once a band has signed a letter of intent, they will either eventually sign a contract that suits the label or they will be destroyed.One of my favorite bands was held hostage for the better part of two years by a slick young “He’s not like a label guy at all,” A & R rep, on the basis of such a deal memo. He had failed to come through on any of his promises [something he did with similar effect to another well-known band], and so the band wanted out. Another label expressed interest, but when the A & R man was asked to release the band, he said he would need money or points, or possibly both, before he would consider it. The new label was afraid the price would be too dear, and they said no thanks. On the cusp of making their signature album, an excellent band, humiliated, broke up from the stress and the many months of inactivity. There’s this band. They’re pretty ordinary, but they’re also pretty good, so they’ve attracted some attention. They’re signed to a moderate-sized “independent” label owned by a distribution company, and they have another two albums owed to the label. They’re a little ambitious. They’d like to get signed by a major label so they can have some security you know, get some good equipment, tour in a proper tour bus — nothing fancy, just a little reward for all the hard work. To that end, they got a manager. He knows some of the label guys, and he can shop their next project to all the right people. He takes his cut, sure, but it’s only 15%, and if he can get them signed then it’s money well spent. Anyways, it doesn’t cost them anything if it doesn’t work. 15% of nothing isn’t much! One day an A & R scout calls them, says he’s ‘been following them for a while now, and when their manager mentioned them to him, it just “clicked.” Would they like to meet with him about the possibility of working out a deal with his label? Wow. Big Break time. They meet the guy, and y’know what — he’s not what they expected from a label guy. He’s young and dresses pretty much like the band does. He knows all their favorite bands. He’s like one of them. He tells them he wants to go to bat for them, to try to get them everything they want. He says anything is possible with the right attitude.They conclude the evening by taking home a copy of a deal memo they wrote out and signed on the spot. The A & R guy was full of great ideas, even talked about using a name producer. Butch Vig is out of the question-he wants 100 g’s and three points, but they can get Don Fleming for $30,000 plus three points. Even that’s a little steep, so maybe they’ll go with that guy who used to be in David Letterman’s band. He only wants three points. Or they can have just anybody record it (like Warton Tiers, maybe– cost you 5 or 7 grand] and have Andy Wallace remix it for 4 grand a track plus 2 points. It was a lot to think about. Well, they like this guy and they trust him. Besides, they already signed the deal memo. He must have been serious about wanting them to sign. They break the news to their current label, and the label manager says he wants them to succeed, so they have his blessing. He will need to be compensated, of course, for the remaining albums left on their contract, but he’ll work it out with the label himself.Sub Pop made millions from selling off Nirvana, and Twin Tone hasn’t done bad either: 50 grand for the Babes and 60 grand for the Poster Children– without having to sell a single additional record. It’ll be something modest. The new label doesn’t mind, so long as it’s recoupable out of royalties. Well, they get the final contract, and it’s not quite what they expected. They figure it’s better to be safe than sorry and they turn it over to a lawyer–one who says he’s experienced in entertainment law and he hammers out a few bugs. They’re still not sure about it, but the lawyer says he’s seen a lot of contracts, and theirs is pretty good. They’ll be great royalty: 13% [less a 1O% packaging deduction]. Wasn’t it Buffalo Tom that were only getting 12% less 10? Whatever. The old label only wants 50 grand, an no points. Hell, Sub Pop got 3 points when they let Nirvana go. They’re signed for four years, with options on each year, for a total of over a million dollars! That’s a lot of money in any man’s English. The first year’s advance alone is $250,000. Just think about it, a quarter million, just for being in a rock band! Their manager thinks it’s a great deal, especially the large advance. Besides, he knows a publishing company that will take the band on if they get signed, and even give them an advance of 20 grand, so they’ll be making that money too. The manager says publishing is pretty mysterious, and nobody really knows where all the money comes from, but the lawyer can look that contract over too. Hell, it’s free money. Their booking agent is excited about the band signing to a major. He says they can maybe average $1,000 or $2,000 a night from now on. That’s enough to justify a five week tour, and with tour support, they can use a proper crew, buy some good equipment and even get a tour bus! Buses are pretty expensive, but if you figure in the price of a hotel room for everybody In the band and crew, they’re actually about the same cost. Some bands like Therapy? and Sloan and Stereolab use buses on their tours even when they’re getting paid only a couple hundred bucks a night, and this tour should earn at least a grand or two every night. It’ll be worth it. The band will be more comfortable and will play better.The agent says a band on a major label can get a merchandising company to pay them an advance on T-shirt sales! ridiculous! There’s a gold mine here! The lawyer Should look over the merchandising contract, just to be safe. They get drunk at the signing party. Polaroids are taken and everybody looks thrilled. The label picked them up in a limo. They decided to go with the producer who used to be in Letterman’s band. He had these technicians come in and tune the drums for them and tweak their amps and guitars. He had a guy bring in a slew of expensive old “vintage” microphones. Boy, were they “warm.” He even had a guy come in and check the phase of all the equipment in the control room! Boy, was he professional. He used a bunch of equipment on them and by the end of it, they all agreed that it sounded very “punchy,” yet “warm.” All that hard work paid off. With the help of a video, the album went like hotcakes! They sold a quarter million copies! Here is the math that will explain just how fucked they are: These figures are representative of amounts that appear in record contracts daily. There’s no need to skew the figures to make the scenario look bad, since real-life examples more than abound. income is bold and underlined, expenses are not.Advance: ................................................... $ 250,000Manager’s cut:........................................... $ 37,500Legal fees: ................................................. $ 10,000Recording Budget: ..................................... $ 150,000Producer’s advance.................................... $ 50,000Studio fee.................................................. $ 52,500Drum Amp, Mic “Doctors”:.......................... $ 3,000Recording tape........................................... $ 8,000Equipment rental:...................................... $ 5,000Transportation:.......................................... $ 5,000Lodging while recording:............................. $ 10,000Catering:.................................................... $ 3,000Mastering:................................................. $ 10,000Tape copies, reference CDs, misc. expenses: $ 2,000Video budget: ............................................ $ 30,000Cameras: ................................................... 8,000Crew:......................................................... $ 5,000Processing and transfers:............................ $ 3,000Off-line:..................................................... $ 2,000On-line editing:.......................................... $ 3,000Catering:.................................................... $ 1,000Stage and construction:.............................. $ 3,000Copies, couriers, transportation:................. $ 2,000Director’s fee:............................................ $ 3,000Album Artwork:.......................................... $ 5,000Promotional photo shoot and duplication:... $ 2,000Band fund:................................................. $ 15,000New fancy professional drum kit:................ $ 5,000New fancy professional guitars [2]:.............. $ 3,000New fancy professional guitar amp rigs [2]:.. $ 4,000New fancy potato-shaped bass guitar:......... $ 1,000New fancy rack of lights bass amp:.............. $ 1,000Rehearsal space rental:............................... $ 500Big blowout party for their friends:............. $ 500Tour expense [5 weeks]:............................. $ 50,875Bus:........................................................... $ 25,000Crew [3]:.................................................... $ 7,500Food and per diems:................................... $ 7,875Fuel:.......................................................... $ 3,000Consumable supplies:................................. $ 3,500Wardrobe:................................................. $ 1,000Promotion:................................................ $ 3,000Tour gross income: .................................... $ 50,000Agent’s cut:............................................... $ 7,500Manager’s cut:.......................................... $ 7,500Merchandising advance: ............................ $ 20,000Manager’s cut:........................................... $ 3,000Lawyer’s fee:............................................. $ 1,000Publishing advance: ................................... $ 20,000Manager’s cut:........................................... 3,000Lawyer’s fee:............................................. $ 1,000Record sales:.............................................. 250,000 @ $12 =$3,000,000Gross retail revenue Royalty:...................... [13% of 90% of retail]:.................................................................. $ 351,000Less advance:............................................. $ 250,000Producer’s points:...................................... [3% less $50,000 advance]:.................................................................. $ 40,000Promotional budget:.................................. $ 25,000Recoupable buyout from previous label:...... $ 50,000Net royalty: ............................................... $ -14,000Record company income:Record wholesale price:.............................. $6.50 x 250,000 =.................................................................. $1,625,000 gross incomeArtist Royalties:.......................................... $ 351,000Deficit from royalties:................................. $ 14,000Manufacturing, packaging and distribution: @ $2.20 per record: $ 550,000Gross profit:............................................... $ 710,000The Balance Sheet: This is how much each playergot paid at the end of the game.Record company:....................................... $ 710,000Producer:................................................... $ 90,000Manager:................................................... $ 51,000Studio:....................................................... $ 52,500Previous label:............................................ $ 50,000Agent:....................................................... $ 7,500Lawyer:...................................................... $ 12,000Band member net income each:.................. $ 4,031.25The band is now 1/4 of the way through its contract, has made the music industry more than 3 million dollars richer, but is in the hole $14,000 on royalties. The band members have each earned about 1/3 as much as they would working at a 7-11, but they got to ride in a tour bus for a month.The next album will be about the same, except that the record company will insist they spend more time and money on it. Since the previous one never “recouped,” the band will have no leverage, and will oblige. The next tour will be about the same, except the merchandising advance will have already been paid, and the band, strangely enough, won’t have earned any royalties from their T-shirts yet. Maybe the T-shirt guys have figured out how to count money like record company guys. Some of your friends are probably already this fucked.Steve Albini is an independent and corporate rock record producer most widely known for having produced Nirvana’s “In Utero”.OK what about making money from Spotify, pandora, youtube etcEach 10,000 albums sold on iTunes (or 100,000 song downloads) generates $70,000 in revenue for the solo artist or band. To achieve the same revenue per 10,000 fans in streams, the band has to generate 30 million streaming plays (as detailed above) if they are distributing their music across the most common streaming services including Spotify and YouTube.In 2013 the top 1% of new releases (which happen to be those 620 titles selling 20k units or more) totaled over 77% of the new release market share leaving the remaining 99% of new releases to divide up the remaining 23% of sales.This appears to confirm our suspicion that the internet has not created a new middle class of empowered, independent and DIY artists but sadly has sentenced them to be hobbyists and non-professionals.