Analog vs Digital video storage - which is smaller?

roxics

roxics

Veteran
I would have assumed digital video storage is smaller than analog video, mostly because of all the compression options you get with digital. But even uncompressed digital, I assumed would be more efficient since you're dealing with samples of an analog signal that contains a more gradient signal. But a couple things got me wondering. First is the Hi8 versus Digital8 formats. You can store two hours of Hi8 video on a Hi8 tape, but only one hour of Digital8. Despite Digital8 using DV compression at 8bit 4:1:1.

The next was a fascinating (to me at least) discussion I found in a newsgroup from 2006 discussing the potential digital capacity of a Laserdisc. The answers vary, but several conclude somewhere in the ballpark of a DVD. Which would mean the analog video on a laserdisc is about the same size as an MPEG2 compressed digital file in the typical DVD bitrate range.
https://groups.google.com/g/alt.vide.../c/NgDSeloDTGo

So I'm curious what the accurate answer is. If anyone here knows.
 
roxics said:
So I'm curious what the accurate answer is. If anyone here knows.
Click to expand...

if you are talking about physical size and/or weight, there is no question that digital is far smaller. I have a couple of these 2TB Samsung T7 SSD drives that are not much bigger than a matchbook ($219), and they are super fast (1000 MB/s). The amount of footage can you store on a 2TB drive depends on the codec, but if you're talking about Hi8 formats, I'll bet a single T7 will store hundreds, of hours of footage. Maybe even thousands. I'll bet you can't beat that with any analog storage device, and there's probably even smaller digital storage than the T7 that I am not aware of.

https://www.bhphotovideo.com/c/produ...table_ssd.html
 
Doug Jensen said:
if you are talking about physical size and/or weight, there is no question that digital is far smaller. I have a couple of these 2TB Samsung T7 SSD drives that are not much bigger than a matchbook ($219), and they are super fast (1000 MB/s). The amount of footage can you store on a 2TB drive depends on the codec, but if you're talking about Hi8 formats, I'll bet a single T7 will store hundreds, of hours of footage. Maybe even thousands. I'll bet you can't beat that with any analog storage device, and there's probably even smaller digital storage than the T7 that I am not aware of.

https://www.bhphotovideo.com/c/produ...table_ssd.html
Click to expand...

I guess your answer points out how strange of question it is that I'm asking. Since I don't know how you would even try to store analog video on such a device as that. I guess what I'm asking is when in comparison to two physical mediums that can store both, like a tape or a disc of some type. Such as the examples I gave.

I don't completely understand analog video and the ways it can be reduced or compressed or whatever the call it. I know composite video and component video are different, with the later requiring more bandwidth as it carries a separation of the chroma and luma signals with separate red and blue. I know analog video can use different signal bandwidth and I don't pretend to know how MUSE works as a way of reducing things. So I can't really say for sure what kind of analog video I'm asking to compare here and what it equals as a digital counterpart. But I imagine there has to be some kind of comparison that can be made between some kind of top quality analog video signal of a given resolution and its uncompressed 12bit or 16bit 4:4:4 counterpart. And at that point I guess I'm asking, which would take up more magnetic tape or require a bigger disc?
 
DLD said:
H264 was roughly a 6-time reduction in file sizes vs. analog.
Click to expand...

Can you clarify that? Reduction from what kind of analog video and converted to what in digital terms? Also reduced 6 times how, you mean like cut in half 6 times or divided by 6 or...?
 
An analog TV channel was 6 Mb/s (Standard Definition). Digital could put six in its place (or a combination of one HD and 2-3 SD's)
 
DLD said:
An analog TV channel was 6 Mb/s (Standard Definition). Digital could put six in its place (or a combination of one HD and 2-3 SD's)
Click to expand...

Oh I see what you're saying. Ok. But ATSC here in the US at least uses Mpeg2 for digital broadcasting. That's going to chance with ATSC 3. But also you could compress it more to fit more channels. Unless there is a certain quality threshold you're meeting, which I assume there was, in which case how was that determined? On what metric does it compare to analog?
 
roxics said:
ATSC 1.0 was Mpeg2. So that's the one with six channels in a 6MHz signal.
Click to expand...

Six SD/480p capacity. Most stations took either a 1080i or a 720p option for the main channel, which left two sub-channels for other things

And, I had an HD tuner then and an antenna for the OTA. And a 4:3 27" Advanced Definition TV for the analog cable.
 
Okay, here's my take.

Analog takes up less room than uncompressed digital, given the same mechanism.

Let us simplify the question and consider only audio, not video, and again only uncompressed.

To record audio, you must store two things: frequency and loudness. All audio is a mixture a various frequencies at various amplitudes. In other words, waves of varying skinniness and various heights.

Let us first consider the analog way of storing it. To store a wave of sound, you store a wave of something else, analogously. So for example, let us consider just 1 second of a 20 kHz sine wave. You need to somehow etch a wave that zigzags 20,000 times, say in a vinyl disk. Furthermore, to store the amplitude, you need to etch it in a way that deepens and rises to the same degree. For example, for a sound that gets 3 times louder than its softest point, I would imagine that you would need a groove that varies also by 3 times, say from 1 micrometer to 3 micrometers, for that 1 second sample. So now you have a groove in a plastic disk that zigzags 20,000 times from side to side while also ebbing and flowing between 1 and 3 micrometers. In other words, you have etched the 2 dimensions of the sound (frequency and amplifude) in 2 dimensions in a material (horizontal zigzags and vertical hills).

Now let us consider the digital way. To store the same 1 second of a 20 kHz sine way, you need to store 40,000 samples. Each sample must have 16 bits (for CD quality sound, at least). Each bit, if using the same mechanism, would also be etched onto a plastic disc and then read by a needle (I guess we're talking about a digital vinyl record, which as far as I know never existed, but it suffices for this example). In the case of digital, you don't need varying depths, as with analog. You can standardize on one depth to represent a 0, and another to represent a 1. Regardless you will need to stuff 32 of them in the same space that in analog you stored one zig, or one zag. But this likely isn't possible. Remember, we are saying that we will use the same physical technology to store the digital as we did for the analog. It's no fair jumping ahead into the future, and saying, well in the future we will make smaller and smaller needles, which can etch tighter and tigher zigzags in the record --- or how 'bout lasers!!! Because you would have to turn around and grant the same technology to the analog in our example --- since that is the very question of this thread.

I said that digital doesn't need the varying fineness of depth that analog does. But that doesn't buy us any density, since what I was trying to get to is that, for any stage of physical technology, the limitiing factor likely is just how tight it can etch those zigzags. Even though you need to etch bumps over only 2 arbitrary depths, instead of a nice wide dynamic range for the analog record, doesn't mean that the needle can pack them any more tightly together on the horizontal plane. And so in fact, the surface area needed for digital is vastly greater than that for the analog version. Again, it is 32 bumps for every 1 bump that the analog version needed.

Analog exploits the thickness of the plastic, to record information vertically as well as horizontally, while digital records its stuff mostly horizontally, more or less oblivious to the potential of the medium's vertical dimension. The same principle extends, I think, to optical and magnetic media. So that is why perhaps that Digital 8 was only 1 hour, even using compression.

(I could be totally wrong. This is really a question for electrical engineers, who do not frequent dvxuser.com.)
 
Last edited:
combatentropy said:
Okay, here's my take.

Analog takes up less room than uncompressed digital, given the same mechanism.

Let us simplify the question and consider only audio, not video, and again only uncompressed.

To record audio, you must store two things: frequency and loudness. All audio is a mixture a various frequencies at various amplitudes. In other words, waves of varying skinniness and various heights.

Let us first consider the analog way of storing it. To store a wave of sound, you store a wave of something else, analogously. So for example, let us consider just 1 second of a 20 kHz sine wave. You need to somehow etch a wave that zigzags 20,000 times, say in a vinyl disk. Furthermore, to store the amplitude, you need to etch it in a way that deepens and rises to the same degree. For example, for a sound that gets 3 times louder than its softest point, I would imagine that you would need a groove that varies also by 3 times, say from 1 micrometer to 3 micrometers, for that 1 second sample. So now you have a groove in a plastic disk that zigzags 20,000 times from side to side while also ebbing and flowing between 1 and 3 micrometers. In other words, you have etched the 2 dimensions of the sound (frequency and amplifude) in 2 dimensions in a material (horizontal zigzags and vertical hills).

Now let us consider the digital way. To store the same 1 second of a 20 kHz sine way, you need to store 40,000 samples. Each sample must have 16 bits (for CD quality sound, at least). Each bit, if using the same mechanism, would also be etched onto a plastic disc and then read by a needle (I guess we're talking about a digital vinyl record, which as far as I know never existed, but it suffices for this example). In the case of digital, you don't need varying depths, as with analog. You can standardize on one depth to represent a 0, and another to represent a 1. Regardless you will need to stuff 32 of them in the same space that in analog you stored one zig, or one zag. But this likely isn't possible. Remember, we are saying that we will use the same physical technology to store the digital as we did for the analog. It's no fair jumping ahead into the future, and saying, well in the future we will make smaller and smaller needles, which can etch tighter and tigher zigzags in the record --- or how 'bout lasers!!! Because you would have to turn around and grant the same technology to the analog in our example --- since that is the very question of this thread.

I said that digital doesn't need the varying fineness of depth that analog does. But that doesn't buy us any density, since what I was trying to get to is that, for any stage of physical technology, the limitiing factor likely is just how tight it can etch those zigzags. Even though you need to etch bumps over only 2 arbitrary depths, instead of a nice wide dynamic range for the analog record, doesn't mean that the needle can pack them any more tightly together on the horizontal plane. And so in fact, the surface area needed for digital is vastly greater than that for the analog version. Again, it is 32 bumps for every 1 bump that the analog version needed.

Analog exploits the thickness of the plastic, to record information vertically as well as horizontally, while digital records its stuff mostly horizontally, more or less oblivious to the potential of the medium's vertical dimension. The same principle extends, I think, to optical and magnetic media. So that is why perhaps that Digital 8 was only 1 hour, even using compression.

(I could be totally wrong. This is really a question for electrical engineers, who do not frequent dvxuser.com.)
Click to expand...

It's a fine enough guess that makes perfect sense. Thank you for the example. I wonder how big that record would need to be to put digital on it. Likewise i wonder how much analog audio you could put on a CD, if that's even possible considering CDs also use error correction.
 
You must log in or register to reply here.
Back
Top