I want to replace the interior bulbs in my car to LEDs because I want brighter cleaner white light from LEDs rather than dim yellow white light from halogen bulbs.  They are 31mm festoon bulbs but I'm somewhat confused about SMDs. There are festoon bulbs which have a single SMD, 2, 4, 8 up to 16 SMDs. Does more SMDs necessarily mean it will be brighter?

Not exactly.
The problem is, there are many SMD LED types (3528, 5050,...), each of them of different power and so output per single LED.

So the only information you have for the light output is the light output specified by the seller. We all know, how the sellers play with these, but you have nothing else.
The second specification you should be careful about is the power input: The cheepeese makers tend to cram huge power into very small formats, resulting only into short life and frequent failures. So for a 5W festoon stay away from input power above about 1W.

For an interior lighting I would rather not use the direct incandescent replacements, but rather the small panels (well, unless there is some focusing lens in front of the incandescent, then you can not replace it). Normally the typical car interior light design is, about 70% of the light output get wasted (it is still way less fuel consumption than a better optic would cause by it's extra weight). So with a flat panel, you may easily suffice with about 30% of the lumen output, so in other words may easily get a lot of extra light.

But generally I would be careful to boost any interior light, mainly those occasionally used while driving around the driver seat - the Titanic disaster was to big extend caused by the high onboard illumination levels blinding the crew and they then failed to spot the iceberg in time...
With the interior lights and night driving the problem is exactly the same and that could be the main reason,m why some lights are intentionally designed with so low output.

Thanks for the detailed explanation. I should point out I'm replacing the front map lights, rear dome lights, boot light with LED. They aren't on whilst driving. Also the front sidelights in the headlights which are. Anyway, I'm looking to go with the Philips X-treme Vision range. There is the W5W available in 4000k, 6000k and 8000k which fits the front map lights, boot light and side light.

The festoon for the rear dome light is also available in the same colour temperatures. The reason for my question is that its a single SMD. I was just wondering if the 16 SMD from a generic manufacturer would be brighter. I obviously wanted to go with the Philips.

Philips:

http://www.p4c.philips.com/cgi-bin/cpindex.pl?ctn=129416000KX1&hlt=Link_Overview&scy=RU&slg=AEN
http://www.philips.co.uk/c-p/129416000KX1/x-tremevision-led-led-interior-lighting

16 SMD:

http://www.amazon.co.uk/2Pcs-31mm-White-Festoon-Light/dp/B00NQA3VDY/

I see the 16 SMD one uses SMD Type: 1210,3528

Assuming each of the LED's in the "16LED" one is driven by 15mA, the overall power input of the LED's would be about 0.7W.
The Philips states the overall power input as 0.5W, with the ballast losses I would expect about 0.4W ending in the LED itself.
The raw efficacy could be similar, so the "16LED" may give off about twice the brutto light output.
But the Philips gives off all the light at the place where is normally the filament, so where the lens is focused to give off the desired pattern (illuminate the lap of the passenger sitting there, so the map he is holding, etc), while the "16LED" will give only about 1/4 of it's light there and the rest will go who knows where (glare of some other people in the car,...).
So if really changing for the LED, I would rather go for the Philips one...

I've ordered the Philips one but I also ordered the 16 SMD to compare just out of interest as its only £2.

The Philips one is rated at 45 lumens:

http://www.usa.philips.com/c-p/129416000KX1/x-tremevision-led-led-interior-lighting/specifications

The manufacturer has replied to my question on Amazon and is claiming the 16 SMD one is rated at 96 lumens which is probably grossly over-exaggerated:

http://www.amazon.co.uk/White-16-SMD-DE3175-DE3022-Lights/dp/B00BG9KVS0

It could be, when new. The 100 lm/W is quite realistic for a 6..7000K LED's. But the question is, how that evolve over time, mainly with all the accumulated heat. The cheepeese will most likely use 2S8P configuration with just serial resistors ballast (plus a diode bridge to correct the polarity), so dissipate about 2W on that board, feeding the LED's by about 1W or so (so the 100 lm output; 1W on 16 chips would mean each driven by about 20mA; quite likely scenario).
I would expect when the car voltage goes above 12V (engine running), the current will go even higher, so way higher temperature.

The Philips has 0.5W power input and most likely uses high efficiency DCDC to feed the LED by nearly the same power. With 100lm/W LED that yields the nearly 50 lm output.
But the quarter power means way colder operation, so only that aspect would mean longer life expectation.
Plus the active regulator will make the power independent on the input voltage, so whatever the car voltage would do, the power will stay constant.
And the better optical coupling in the fixture will make at the end way more of the useful light...

I installed both bulbs in turn and took some photos for comparison. The Philips is on the left and the generic 16 SMD sold by Sourcingmap on Amazon is on the right.




I actually think the 16 SMD is brighter but decided to leave the Philips one in as I think its a better choice long term.

Lookint into the glare of the lamp does not show how bright it is. What is the light level and distribution inside the car ?

Both of them are fine. But as Medved said with the Philips the lens is focused to give off the desired pattern. I'm leaving with the Philips one in as it will no doubt prove to be more reliable as well.

Well, I was expecting really a kind of lens, from the pictures here this one does look just as a prismatic difuser (same pattern across the complete surface, regardless if it is below the lamp or in the corner; really aimed neam shaping optics will have it's features centered around the lamp position above the lens), this one does not look like doing much more than just diffusing the sharp light source formed by the filament. There the raw lumen output will likely bring more light on the desired area, but you have to compare that really in the car, when reading a map or so.
But with the lower power of the Philips I would expect less reliability problems and better output stability over time. However I doubt such lantern would ever make so many running hours for these effects to become noticeable. And even if so, it would be ON for just short moments (maximum a minute or so) with long resting time, so the thermal inertia will be sufficient to keep the temperature at reasonable levels.

Our car dome light has two bulbs (one for each front seat): both turn on when you open the doors, then the driver and the passenger can control their own bulb.

I replaced the incandescent bulbs by led. I simply bought them on eBay and chose the one who stated the highest lumen flux (I knew the lumen statement is rarely accurate).

I also tried to have a white led bulb for the driver (for general illumination when opening the doors) and a deep red incandescent bulb for the other light, so that the passenger could use this light without any glare for the driver. But one led bulb directed to the driver's seat was to dim, and the red light wasn't very practical, so I went back to the dual led configuration. I use a Zebralight (with a firefly mode) when I have to read a map, etc. while my girlfriend is driving.