I've been wondering for some time now about signals and tuning.

First off, I'm not an electrical engineer, nor do I know much about cable signals.

I understand the concept of tuning to a specific channel on cable before viewing or recording, but is it theoretically possible to record the untuned signal to an analog device? So that later, one could replay the signal and tune to specific channels from that?

Example: cable signal -> analog recording -> tv tuner -> tv.

I know it's not something that's done, and it probably can't be done, but it's struck me as one of those things that may be possible but no one has bothered to ask.

The other question that's come to mind, just in the last few days, is until tivo, replay, and MythTV came along, people basically had single tuners per device. Only relatively recently has there been a use for multiple tuners on one card.

VCRs never really needed them, since they couldn't record two shows at once. If you wanted to watch one show and record a different one, you'd just use your tv's tuner.

But these days, there's more and more demand for multiple tuners using fewer resources.

I can see some company such as Hauppage or NVidia developing, maybe even in development now, a single card that can offer 16 or even 100 mpeg or xvid encodings at once. The only limit being the bandwidth to the MB itself.

Obviously something like that would have to be digitially tuned rather than using an analog tuner, but am I wrong in thinking it's possible to use a digital chip to read an analog signal?

Feel free to laugh at me or dismiss my questions. I'm just rambling anyway.

Hi,

Actually it is an interesting thought to simply record a whole band of frequencies at once and then later go back to pick just the desired channels to watch or listen. However, I don't know of any technology that is capable of doing that at this stage in life. I know of no way or material for storing a chunk of rf band directly and as with all things, there are limits.

Typically what a tuner does is to selectively look at a window in that band of rf spectrum that is all around us in the form of radio, pagers, cell phones, police radar, weather radar, outdoor thermometers, and the endless list.

To try and capture even one channel of rf space would be resource intensive. The rf of a channel is a carrier of "intelligence" and depending on the method of adding the intellegence to the carrier also has a factor of effecency. A pause in speech (a form of intelligence) still requires full transmitter signal so to record that silent moment will use just as much storeage space as a MaMa Cass high note.

Then comes the storage aspect, just the digital information for a hd decoded stream consumes space at a rate of at least 8 gig per hour. That is already processed into a digital state ready for direct storage. To capture the full channel of that same 8 gig of data would mean that you would also have to save eveything that is not data. To do that would mean that one would have to sample that window into the rf world, and "digitize" it with suffecient detail to be able to reconstruct that moment.

Take a cd as it is a sample of audio (20htz to 20khtz?) and that is tested at about 44khz rate. So that would mean that a single wave of audio is tested 44,000 per second to obtain a digital value to represent that moment. Pick any tv channel (radio, etc) use the same basic pattern and to grab a sample you must use a clock of at least 2x to be able to get enough samples to rebuild the wave form. Starting at the low end ch2, is around 54mhz so 2x would equal 108mhz for a clock to get a sample... oh my,that is the end of the fm radio band..... ignoring that wrinkle, that would leave us with two data words to represent ONE cycle of the 54mhz. Now if the sample has 256 bit level sensitivity that would mean two bytes. There are still 53.999,999 more samples needed to be obtained in that one second snapshot of time. Maybe that gives a idea of complexity and actually, quite wastefull usage of storage space even if it could be done.

The tuner assembly selectively "looks" at a space in the rf band and and demodulates the carrier to obtain the "intelligence" That is what is sampled and at a much lower rate, usually 3 or 4x chroma subcarrier (~3.58mhz) for professional stuff so with a video signal of a maximum of 4.5mhz of video, we can obtain a very good data stream to rebuild that signal later. Storage for your single channel of demodulated video is ~2.2gig per hour as the KM default. I don't know what the actual Hauppauge clock is and may be as low as 2x the 4.5 upper limit or 9mhz. Long ways away from 108mhz.

Many tv's come with two tuners, that would be needed for picture in picture.

Sorry if it is kind of a rough explaination, but hopefully gives you a basic idea behind what is done and what may be limiting factors. Once you have a real signal, some processing can be added that will see that moment of silence and make it into a data steam that says it was 3 seconds of silence and be able to save that as fewer data words. (compression)

Mike

Hi Mike,

thanks for responding.

As I said, I'm just rambling here. But one point you made about space...I'm really focusing on analog storage. Vinyl recordings, for example, store much more of the sound than digital recordings such as CD.

I hear what you are saying though. Trying to record the entire spectrum of signal coming in from cable would be including each channel as well as the noise between channels. It would be like trying to see every color of the rainbow at once.

Kidding.

I know it would, at the very least, require a much higher resolution analog system than video tape, for example. I'm not sure if one could even rip the tuner out of a vcr and have the vcr head even attempt to write the entire incoming signal. Even if the head wrote to the tape, the head isn't fine enough to record it accurately enough.

One would need a much finer stylus and media, something akin to the needle used for playing records, but with the capacity of tape.

Even then it probably wouldn't work.

This really isn't all that useful, even if it were feasible, since one wouldn't really *want* to record most of what's on at any particular time.

As you might guess, I dreamed up this inane idea when I was stuck with a single vcr and frustrated over having to pick which show to record.



True. I'd forgotten about that. However, I'm not sure there's ever been a tv with 3 tuners in it, unless maybe it had picture in picture and a built-in vcr. Even then, adding the extra tuner wasn't a big deal, making it unnecessary for anyone to develop a 16 tuner tv card back then.

Nowadays though, unless broadcasters decide to stop broadcasting, home media centers are going to be growing the way that the internet started to boom back in the 90's. No matter which segment ends up taking the biggest slice of the pie (home brew like mythtv, cable provider pvrs, or tivo systems), there will be a huge and increasing demand for lots of tuners in a single quiet box.

With the growing backlash against drm-heavy corporations, it could be anyone's game.

As I understand it this is something like what some of the radio telescope signal analysis projects do.
With the progression to DTV, the necessity for something like this becomes less relevant as there is only going to be a few muxes (transports) transmitted within an area. For example to capture every UK current mux, you would need 6 tuners, so if you were able to demux the particular programme stream, you would be able to capture every programme.
I remember the BBC developing an experimental box having multiple tuners and a humongous amount of storage so that you were able to capture every mux, for a week, so you have access to a whole weeks worth of programming.

Bruce S.

_________________
Updated 2019/10/26: AthlonII X2 265 Gigabyte GA-970A-DS3P
16Gb PC 1866 DDR3, 500GB+2TB+4TB SATA HDD,
SATA DVD-RW Asus DRW-24D5MT , NVIDIA GeForce GT1080
Hauppauage Nova-T 500, Nova-T LinHes R8.6.1

Hi,

It is a only a thought process for now, however one never knows what future talents and engineers will design and build. It wasn't very long ago when audio was being stored on a hand cranked machine that had a ~2 inch roll covered with a hard wax, then came the acetate, and not far after vinyl. Also mixed in there was wire and tape recorders which most ended for the consumer era with the digital world and cd's. cd's are followed by memory sticks.... next a memory stick with a holographic projector so you can also see the show? And still fit into the pocket?

Maybe a direct recording of a portion of the rf spectrum is possible or may be come possible some day, the key is if there is a need or benefit to doing it.

Mean while, KM puts most historical methods for capturing a moment in time rather far in the dust and I can be comfortable doing it one channel at a time. If I mis a show tonight, a simple check of the guide and record it next time it airs, and you don't have to turn a crank to do it.

Mike

Yes. Theoretically.

A few hours of recording might exceed the entire storage on this planet, though

(actually, the new Solaris ZFS filesystem supposedly can address more storage than there are atoms in our universe, or some such sillyness)

_________________
| Nigel Pearson, nigel.pearson.au@gmail.com
| "Things you own end up owning you" - Tyler, Fight Club

Well, I have some experience with this 'kinda' thing. I have worked on digital recording of a 500 MHz bandwidth at, lets call it, 10 GHz. However it was only recording a few micro-seconds of time at that bandwidth. So I do understand the principles and requirements. To record the entire cable tv bandwidth is a bit larger than 500 MHz and so you would be talking about a storage requirement that is, by comparison to a single HDTV stream, excessive, to say the least. You might, for example, consider it to be in the neighborhood of 100 times the requirement of a single HDTV channel. Do you have a machine that you can store at the rate of 800 MB per second?

I could tell you, but then I'd have to shoot you.

Ah, but the original spec was as analog... that's the one that bites

The high tech 2" video tape was doing some reaching to obtain the 4.5mhz band needed and that was with spinning heads, ~several thousand rpm so I think as analog recording..... not likely with anything current to reach frequences 54mhz> <1ghz.

The question I think, is there any value to even attempting such a goal? Although, an interesting idea.

Mike

A more practical approach would be to provide a bank of tuners to select every single frequency used in a given area, then demodulate and store (for digital signals) or convert (analog) and store the programs.

For example, suppose you have 100 viewable channels in a given area. You could imagine a box that has 100 tuners a lot more easily than one which stores the entire spectrum. Plus, its easier for playback as well.

My house is in a development of 52 homes. We each have a fiber which links our homes to a common point where we have another fiber that links us to the Internet. The point is, we have 100M connections between each home, since we're all on a common switch. (And it could be a 1000M connection if we update a $50 box in each house.)

What is the largest network of MythTv BE systems you've ever heard of? Wouldn't it be cool to have 52 homeowners all sharing shows via KnoppMyth? If each one had at least two tuners, we could record 100 channels simultaneously with such a system...

_________________
Do you code to live, or live to code?
Search LinHES forum through Google

Do it!

_________________
Currently Running:
Too lazy to update this with my current hardware, I'll redo it during my next install =)

I would LOVE to be a part of such a neighborhood. Do we also get to share pr0n videos also?

Liv2cod has the right idea. If you lived in a neighborhood that had its own intranet, I could see it happening. Hmm.. haven't seen a neighborhood that fits this description, but I could see it happening. Not sure what the economic driver would be though.

I am not aware of any analog device that has the BW to capture the (cable) bah-jigga-hertz of analog broadcasting out there. Although this **may** be theoretically possible, technology (device physics) does not have the capability to do this (today). If this could be done, you would still need to demodulate the channel and theoretically the information content would still be the same: you would need amount of space as if you demoded all channels at the same time.

It would be 1000X more cost effective to have 100 tuners and PCs than to try to implement the device physics for a super BW analaog recorder.

_________________
R7.3: 0.22.20091023-1, Hauppauge PVR-500 (Philips FQ1236A MK4), Gigabyte Gigabyte EG45M-UD2H, E5200 2.4Ghz, 2GB RAM, NVIDIA GEFORCE 256MB