High Definition via Internet

While I do mostly agree with this article from Gizmodo explaining why High-Definition is still a long way from what we want it to be, I can't stop wondering why they insist so much on the bitrate issue.
 Sure, you can't have high quality content without enough bitrate - but... how much is enough? Do we really need those 40Mbits+ to have an enoyable HD internet experience?

I don't think so...

While I do agree that companies severey throttle bitrate so they can push some extra channels down the tube (and this is happening in almost every transmission media) there's more to quality than just the bitrate itself.

Here's a chart showcasing the bitrates used in different media:

Well, the issue isn't exactly what the bitrate is, but instead we'll focus on how it's used!

Just imagine for a second, what would happen if I took one of my old and dusty VHS tapes (assuming they would still work - and assuming I would still have some old VCR around to play it) and convert it to Bluray using its maximum bitrate?
Do you think you would have a "movie experience" with those 40Mbits of advertised bitrate?

It's all about the source material and how much we can compress it so it looks "nearly" the same.

There's a huge community of movie fans that excels at compressing movies at highest quality and the least used space, squeezing every bit of perfection out of the available codecs. These are the people that know when enough is enough.
It's not uncommon to see them arguing like:
"Why didn't you make it fit a DVD DL perfectly (8GB)?"
- "I won't make it use more space if there's no point in it. If it fits in 5GB and it won't look any better if it used 8GB; what's the point in making it larger?"

This is exactly what compression is all about: achieving  transparency.


Iimagine you have just an image consisting of just two dots (pixels) in black and white.
You could store this image in a computer using just 2 bits: one "0" for the black pixel, one "1" for the white pixel.
But hey - if you were trying to sell it, wouldn't it sound better if you said you have a 32bit color image?

So we'll convert this image that used 1bit per pixel to another format using 32bits per pixel.
Instead of the original "01" image using 2 bits of storage space you'll now have:
00000000 00000000 00000000 00000000 (that's the first 32bits for the black pixel)
plus
00000000 00000000 00000000 00000001 (that's the second pixel, with the "1" white color)

You went from 2 bits to 64 bits (a 32x increase in used space) - without gaining anything in image quality.


Of course, thing get a lot more complicated when we're talking about "real images" and motion, but the key point about transparency is compressing the original image as much as possible without loosing any details.
For a very dark - or completely black - image, that may mean a couple of bytes, for a very complex scene it might mean Megabytes.

Taking the 2 pixel example from above, if you had a 64 bit image of those 2 pixels, you could compress it down to just 2 bits without loosing any detail.


You've probably noticed this already when you're listening to MP3 music.

As you know, you can have different quality setting for your MP3, some using 64Kbits/s others using 256Kbit/s. Most people also know that the higher the bitrate, the better it usually sounds.
But if you can usually tell (hear) the difference between a 64Kbit/s MP3 and one encoded at 256Kbits - distinguishing between a 256Kb and 384Kb MP3 is a lot harder. And what about if you had a 512Kb/s MP3 vs a 1024Kb/s one? Could you really tell the difference?
Not really, because like I said above, the extra bandwidth wouldn't really be needed for it.


In High-Definition movies things are a bit different because we're talking about an insane amount of data compared to CD audio.

Consider this: a single uncompressed 1080p frame (that's 1920x1080 x 24bits) takes about 6Mbytes. But, you have 30 frames per second showing in your TV screen, so that adds up to about  178MB per second --  10GBytes per minute - 625Gbytes per hour!

Just try and fit that into your "high-tech" Bluray disc which takes only 50GB and holds 3 or 4 hours of film.

That's where codecs like MPEG come in - and in the case Bluray it will reduce the original bitrate from 1400Mbits/s down to 40Mbits -- that means cutting about 97% of the original information.

Then we have different codecs, each with its own efficiency. While MPEG2 (the codec used in DVDs - and some Bluray discs still) was efficient, the newer H.264 is considered to be 4x better.
That's why you can easily rip and encode your DVD movies to your media center, and have the movie shrunk down to 1/4 its original size without noticing any quality loss.
And the same thing happens to Bluray movies and any other HD stream (provided they're not already using the H264 codec.)
That's why you can have great looking HD movies in a quarter of the size used by Bluray disks.

Even if the movie is already using H.264 and taking 40GB of space, you can still compress it a bit more and get it to around 10GB with "some" acceptable quality loss.

Unless you compare it side by side, the differences won't be that visible for most people.

The point is, it's not all about the bitrate...
More is better, yes. But just until it is "enough".

I don't want to reach a point where people look blindly at a bitrate sticker to access a movie quality.

If not, companies might just start encoding movies so they fill up the disc completely whether that's the best way to do it or not.

Back to the topic, even though the available offerings are a clear rip-off, offering so-called HD streams worst than your regular DVD, it is indeed possible to get HD via the Internet. They just need to get closer to the 5-10Mbits bitrates for the HD streamsand people will be able to see the difference (for the better.)

Considering most broadband connections are now advertising 20+ Mbits (not considering those countries where 100+ is already the norm) that's not so far off.