I'm not sure where +16 comes from. I see a peak hp with manifolds of 328 at 5200, and 349 at 5000 with headers (see dyno sheets at 15:09 and 30.40). That's a 21 peak hp gain. But the gains are throughout the rpm range, not just at peak.
Here's what I see headers being worth over the manifolds, according to Nick's dyno:
3000 rpm +6
3500 rpm +13
4000 rpm +21
4500 rpm +16
5000 rpm +29
5200 rpm +20
As for my own experience, I replaced the stock manifolds on my old otherwise stock 73 Duster 340 with cheap BlackJack headers, and replaced the 68-70 340 manifolds on the mildly modified 340 in my 65 Barracuda and the mildly modified 360 in my 67 Barracuda coupe with TTI. In all cases, the power difference was amazing, at all speeds, low to high. Makes me wonder about the small gain at 3K on Nick's dyno. But ???
I don't have a problem with anyone who keeps his manifolds - I did for a long time, because in the old days, headers leaked, rattled, had to be clearanced with a BFH, and the A body ones hung so low that two tubes got flattened (my stepson knocked a hole in his in less than two weeks; I never beat mine that badly in two years of daily driving, but they did get flattened). The ones I put on my Dad's 440 had so little plug wire clearance that we routinely shorted out two plug wires - a problem we almost fixed with Hemi spark plug wire ceramic insulators.
Manifolds were much quieter, fit better, were more durable, and still made decent power for a street cruiser.
But there are good headers today that fit, don't leak and don't hang low. There's really no downside to running them except for the cost (and I sold the 68-70 manifolds in my two Barracudas for enough to pay for TTI headers).
