I'm no expert in this, so I'll just say what I've heard and what has made sense.
Everything that requires individual thought, imagination, etc. Will start with your take on it.
If you are to go for some form of degree don't do the typical thing, its a complete and utter waist if you just knuckle under and work hard.
If you can use your time, experience and skill properly you can aim for a degree while achieving something great.

Get told to make a lighting engine? Strive to make a flexible material based lighting engine with indirect illumination and ambient occlusion post-processing.
Get told to make a water shader? Make a full featured water shader allowing for wave variation and control with fantastic reflections.
Get told to render an outline? Make a shader supporting gradient outlines with Fresnel effects and line variation.

Those are just rough exampled but the point is, if your going to try for a degree, don't just try to pass, this is one of the cases where it would pay off to show off.

how do you prove that you can do the job?

Well my idea would be to make a blog/library/portfolio of code samples that implement various techniques/algorithms etc. Maybe a small game.
I know that is definitely too much work for one persons spare time to make a complete graphics engine from the scratch.
You think that will help?

For the programmer vs. software engineer issue: from my observation of many developers, it appears that there is very little correlation between having a PhD and being a good software engineer (vs. say a BS or MS degree) in the average case. I have seen many PhDs (Computer Science) who are poor software engineers, and many non-PhDs who are top notch, with no clear bias one way or the other. What I will say though, is that if I had to rank the top ten uber-developers that I have worked with, more then half have PhDs, and the others have Masters or Bachelors degrees. I believe the explanation (apart from confirmation bias :-) ) is that a PhD is really what you make of it (I know it sounds cliche, but stay with me) -- it is a huge opportunity to develop advanced quantitative and analytical skills, and dive deep into a few areas of cutting edge research and participate there. These opportinities generally don't exist in industry (which tends to be a bottom line culture), and even many PhDs steer clear from such work since quite frankly, it's hard. Those that stay on a steep learning curve for all 9 years of their academic lives, will tend to bring the same curve into industry -- they will ramp up quickly, and have a remarkable toolset from which to draw on.

In terms of advice, as others have said, there are plenty of jobs for PhDs in industry - whether or not you are likely to have research freedom is somewhat dubious though (you have to be a damn good researcher to achieve this - lots of high profile publications). If you want to do a PhD, and want the full benefit, make sure that it is as challenging as possible without being over ambituous - easier said than done, I know: When you do your grauduate coursework (during your MS, assuming a US degree), you will be able to do work that builds on top of your BS, similarly when you do your PhD (thesis part), make sure that it is something that actually requires the knowledge and skill you developed in your MS. Many of the less succesful PhDs I have seen did thesis work that they likely could have done with the tools gleaned from 2 years of undergraduate study - this is a warning sign. Also, for your undergraduate, I strongly suggest having a second major in some other hard science (math, applied math, statistics, physics, chemistry, etc.), and incorporating this field into your PhD (i.e., becoming more familiar with the "hard sciences") - the cool thing about graphics, is that there is room to incorporate all the usual suspects into your research.

My background: I have a PhD in a mix between computer graphics and computational geometry, and have spent over 10 years in the graphics software (non-game) industry, and have worked as both a software engineer and manager.

I've recently realized that I really love academic work when it comes to computer graphics programming. I am close to finishing my undergraduate degree where I spent a year working on a research project, and am thinking about whether I want to pursue a PhD. The thing is, I don't really want to teach, and I don't want to have to scramble together some grants to work on my own cutting-edge research project. I've heard that a PhD will get you either a teaching or research career. Research sounds fun to me, but I would much rather be doing work at a company like Intel or nVidia working on a real-life product than dreaming up some new algorithm for cloth simulation. My ultimate goal in getting a PhD would be to master the subject, but I don't want to whittle away my job opportunities just because I'm overqualified.

Is there anyone here who has a PhD in Computer Graphics (or even something else)? Do you feel like it was really worth it versus just working in industry with a Bachelors degree? The main issue I have is that I don't want a job where I'm just pushing out product releases for eternity. I want to work on something innovative. I guess it just seems like a lot of stuff with computer graphics is happening at the graduate level--unless you happen to get a job at Valve or DICE where you get to build your own game engine. At the same time, I want to be treated as someone who has authority and is respected. For instance, it seems to me that a majority of game development studios work their people like slaves and pay them peanuts. That just doesn't sound sustainable to me (and it's not, just look at the turnover rates).

I hope some of that made sense. I guess I'd just like to hear your input on the subject if you have any.
Thanks!

Every master in the industry I've ever read about didn't ask this question. They found something they were passionate about, the innovation you speak of, and pursue it rain hail or shine, because its what they want to work on. Examples are Chris Sawyer (phd in computer sciences, started porting games to make an income, wrote his own game engine from that experience that was innovative at the time, combined it with something that was fun and made millions selling RollerCoaster Tycoon), John Carmack (Read "Masters of Doom" - very interesting story), Steve Jobs (watch his Stanford Uni speech on Youtube), Or Gabe Newell who ran from Microsoft to follow a dream of making a better variant of Doom. I think you got to find what interests you and follow that direction; and new things may pop up on the way that change your path. As Steve Jobs said in that speech, looking back on his calligraphy classes, he used those skills to create the font system that he so kindly put "Microsoft copied from Mac", had he never had those classes, the typography we use on computers today may never have come to realisation.

Research sounds fun to me, but I would much rather be doing work at a company like Intel or nVidia working on a real-life product than dreaming up some new algorithm for cloth simulation. My ultimate goal in getting a PhD would be to master the subject, but I don't want to whittle away my job opportunities just because I'm overqualified.

I'm not sure where you are, but in the UK some universities have an industrial doctorate centre, which offer practice based PhDs (The national centre for computer animation [NCCA] offer Phd's of this sort). Typically these involve working for a company, where the research conducted there counts towards your Phd. In general, this isn't a bad plan imho. It gives you access to the companies core frameworks, therefore cutting down your work load a bit. The only small caveat is that I'd imagine the quality of the company would affect how much academic freedom you actually have whilst working for the company. (i.e. A company experimenting with 'freemium' content may not be the most inspiring of places to work..... )

Is there anyone here who has a PhD in Computer Graphics (or even something else)? Do you feel like it was really worth it versus just working in industry with a Bachelors degree?

I had an option to do one a few years ago, but in the end turned it down, and am actually glad I did. I think at the time I wouldn't have been fully prepared, so I'd have spent my time 'learning' rather than 'researching a novel idea'. The best place to learn the trade is within the industry imho. If a PhD appeals, then consider spending a year in industry before starting on the Phd (all though that carries certain risks, because you may get too used to the wage!). The only reason I say this is that industry gives you a very good education in how to deliver features on time - something that will help you achieve much better results on the PhD.

The main issue I have is that I don't want a job where I'm just pushing out product releases for eternity.

You'll struggle to find a programming job that doesn't have a product release cycle. If you do find one, realise that the company may be about to fold ;)

I want to work on something innovative. I guess it just seems like a lot of stuff with computer graphics is happening at the graduate level--unless you happen to get a job at Valve or DICE where you get to build your own game engine.

There are many companies out there in the field of middleware who build lots of interesting tech (I work for one of them). Generally it's slightly less pressure than working on a game, involves less overtime, and encourages new and novel approaches to existing problems. It's definitely more innovative & productive than an academic environment (where team work is generally unheard of), however that comes at a price - about 75% of the time you will be writing docs / writing unit tests / doing design work, you know, just standard software dev work....

Go for it! I did a PhD directly after my masters in computer science engineering (with the final master thesis inside a game company by the way). Doing a PhD was the coolest experience ever. Finally all that knowledge that at the time at least was cutting edge (physics algos, rendering algos etc...) that I researched in the game community was useful :-) It is hard to find a job as a researcher in a high tech company but the search is worth it and the reward is a life long ( I hope ;-)) adventure where you refine your skills to the utmost perfection.