Ear shave for sure. K+N filters are recommended. As they are an oiled high flow filter they are as good or better then stock. K+N's are proven and tested to filter better the OEM filters and flow better.
I'm going to use them myself. The P/O for my bike did the "Coastering". So I'm going to complete the process and remove the pods and surge box
Go with the ear shave. Look in the Carbs and Fuel threads for VoId Docs' ear shave procedure. It a greatly detailed how to.
Unfortunately that is not the case. K&N filters have certainly been tested plenty of times, and every single time they have failed miserably compared to stock filters. Do your own research online. I could post enough information on it to crash the server. I have posted at least 20 different tests already, here is some more information, from a BMW car forum.
Rob Levinson from UUC Motorwerks says that "it is really doing every E34 owner a disservice for three reasons:
• First, conclusive tests from independent labs show that K&N filters, no matter how well-oiled and cleaned, let through huge amounts of large particulate that damages engines. Oil analysis shows 5x as much silicate contamination. Additionally, the oil from these filters tends to muck up the MAF and cause an engine-damaging lean-run condition.
• Second, an open element filter like that sucks in hot underhood air and subjects the filter to "fan wash", the swirl that creates a vacuum effect. On the E36 M3, we measure a 15hp loss from open cone filters.
• Third, the E34 airbox is designed with an integral velocity-stack style venturi at the opening. Running one of these engines on a dyno, you can cause a 5hp drop just from putting your finger on the stack lip. Removing that airbox undoes a lot of good BMW engineering. Yes, these filters make a "really cool" noise... but noise does not equal power. Stick with a stock paper element in the factory airbox, you're not getting any more power with a K&N garbage-filter, just damaging your motor."
Jim Conforti (AKA the Land Shark) did some testing:
This was a scientific test, not one done by filter manufacturer X to show that their filters are better than manufacturer Y. The test results are pretty irrefutable as the test lab tests and designs filters where "screw ups" are absolutely NOT allowable (I can't say any more for security. Think "Glow in the Dark").
A scientific test was done on TEST filters where air was loaded with ACCTD (some standardized "test dust" called AC Coarse Test Dust) and sucked through the TEST filter then through an analysis membrane. From the Quantity of dust injected and the amount that gets through the TEST filter and is then captured on the analysis membrane we can calculate the efficiency of the TEST filter in Question.
BMW Stock Filter, Eff. Area of Media: 8.4 sq ft.
K&N Replacement, Eff. Area of Media: 1.6 sq ft.
The filters are the SAME size. They both fit in the STOCK BMW M3 airbox. The difference is that the STOCK filter has 65 pleats 1.5" deep and the K&N only 29 pleats each 0.75" deep.
Now, remember this ratio: " 5.25:1". It's the ratio of the AREA of STOCK to K&N. It's very important and will come into play later.
The STOCK filter efficiency started at 93.4% at 0 loading and increased to 99.2% efficiency as the loading increased to a max tested of 38.8 gm/sq ft of dust.
The K&N filter efficiency started at 85.2% at 0 loading and increased to 98.1% at the max tested loading of 41.38 gm/sq ft.
Now, I hear you. "Jim, that's only a FEW PERCENT". But is it?
Let's look. If we had 100 grams of dust on a new BMW filter we would let through a total of 6.6 grams of dust in. If we used the new K&N filter we get 14.8 grams of dust. That's 224% (TWO HUNDRED TWENTY FOUR PERCENT!!) more dust ingested initially, stock vs. "free flow" and this ratio is pretty much held. Somewhere between 200-300% more dirt gets "ingested" anywhere across loading equivalence. The more INTERESTING thing is when you look at what happens to the DP or Differential Pressure at a constant airflow as you dirty both filters equally with time.
The test used a rate of 75gr of dust per 20 min. Here's where the AREA difference comes MAJORLY into play. See, even though the BMW filter flows a bit less at the SAME loading, it also LOADS UP 5.25 times SLOWER due to it's LARGER effective area. So what happens is that the K&N initially flows better, but as the dirt continues coming in, the K&N eventually flows WORSE while still letting MORE dirt in.
Now, does any of this additional dirt cause problems? I dunno. I suppose we could have a few people do some independent oil analyses on different motors using both K&Ns and Stock filters. Get enough of them, and you'd have a good statistical basis. For me though, it's simple: More DIRT = BAD.
The additional short-term airflow might make sense on a track car. IMHO, it doesn't for the street.
-- Jim Conforti