Signal strength should effect transmit time, not quality.

[natsirt721]

In the current system, any science transmissions are multiplied by your signal strength, reducing the science return for your experiments.  Before 1.2, transmitting science only incurred a proportional penalty for transmission versus recovery, but now this proportional penalty is again divided by signal strength.  This seems non-intuitive - If I collect n bits for return, I should be able to return all of n of those bits, and a weaker signal just means that the transmission will have to have more redundancy to prevent data loss.  It seems silly that my data is just lost to the ether because of a poor signal-to-noise ratio, when I should be able to constantly re-broadcast the data until I have 100% of it accumulated on Kerbin.  The Shannon-Hartley theorem seems to verify this, but I also don't know much about information theory and this could have exactly 0 bearing on the situation that I'm discussing.

 

I propose that instead of a proportional science penalty, there would be a proportional time penalty, incurred by requiring more bits to be transmitted as a function of signal strength.  At full strength the required bits to transmit would be that listed for the experiment, and with decreasing signal strength, a logarithmic or pseudo-logarithmic function to determine how many redundant bits should be transmitted.  If the Shannon-Hartley theorem is valid here, the S/N term would be a function of signal strength, and the bandwidth the advertised bandwidth of the antenna doing the transmitting.  This would require some interaction between the science dialogue window and the antenna system, as the dialogue window would require knowledge of the primary antenna onboard the spacecraft.

 

I do realize that this would greatly reduce the penalty for having poor probe connections, however if the signal strength to S/N conversion function were curved steeply enough, it could offset this by requiring an extreme amount of time and EC to transmit data.  Something that could be useful here would be a manual bandwidth limiter to ensure the probe didn't deplete its batteries right away in the event of an hours or days long transmission.  At this point I'm just spitballing, but let me know what you think of this idea.

EDIT: whoops, wrong forum.  I need to not make posts when I'm tired.

Edited October 24, 2016 by natsirt721
wrong forum correciton

[sal_vager]

Pushed to dev discussion.

[SpacePilotMax]

Seems good, but what do you mean by "transmit time"? Is it like you have to keep an antenna running for a longer period of time and use more EC?

[tater]

Yeah, this got said by me in the 1.2-pre forum, but it's not a thing. 

 

11 minutes ago, PilotMax said:

Seems good, but what do you mean by "transmit time"? Is it like you have to keep an antenna running for a longer period of time and use more EC?

Data transmission is all about S/N. At great distances, this results in a lower effective data rate. Look at New Horizons. Took a ton of data in a few hours, it's taking like a year to send it all back.

Transmission should be dealt with like the MPL, or ISRU. You set it running, then go do something else. At some point you check the craft again, and the games "catches up" the data that has been sent based on the elapsed time.

[p1t1o]

I think this is a gameplay allowance. Given that we have the ability in-game to accelerate time up to arbitrarily fast values, any temporal penalty is meaningless, the only meaningful penalty is quality of data.

If one is that fussed about the realistic effects of signal degradation and bandwidth, then the way "science" and research has been reduced to a simple numerical count must send one into conniptions!

[Wjolcz]

I have a feeling this science transmission "tweak" was made to simply make the science system more meaningful. Except it doesn't (because science points, amirtie?).

I'm not against the idea, but if that would become a thing people would simply warp throught the whole transmission time. Maybe the probe could lose control until the whole science was fully trasmitted, or something? So all you would have to do is just put the probe in a stable orbit and make sure it has power until all the science is transmitted. Could be a cool time-based mechanic only if the probes had the ability to collect science automagically. That wouldn't differ much from how the MPL works, except they would run out of science eventually if they were kept in the same orbit.

Edited October 24, 2016 by Veeltch

[natsirt721]

1 hour ago, p1t1o said:

I think this is a gameplay allowance. Given that we have the ability in-game to accelerate time up to arbitrarily fast values, any temporal penalty is meaningless, the only meaningful penalty is quality of data.

Well sure, we do have time warp, but for things like the MPL you don't just start in and fast-forward until it is finished.  Or maybe you do, and that's a gameplay choice.  I think the 'infinite time warp' argument doesn't hold for a lot of things in KSP, primarily ISRU (but also transfer windows and travel times) like tater said.  

1 hour ago, Veeltch said:

I have a feeling this science transmission "tweak" was made to simply make the science system more meaningful. Except it doesn't (because science points, amirtie?).

I'm not against the idea, but if that would become a thing people would simply warp throught the whole transmission time. Maybe the probe could lose control until the whole science was fully trasmitted, or something? So all you would have to do is just put the probe in a stable orbit and make sure it has power until all the science is transmitted. Could be a cool time-based mechanic only if the probes had the ability to collect science automagically. That wouldn't differ much from how the MPL works, except they would run out of science eventually if they were kept in the same orbit.

The idea here is that you shouldn't be penalized in terms of science return for having poor signal.  Either you have a connection and can send the data, or you don't and can't, but being able to half-ass the transmission and only get a partial return is nonsensical.  That being said, there should be some penalty - and seeing as how KSP is a game based on real world science, why not look to real data transmission for a better solution.  Sure, you can warp through if you want, but the transmission system could be tweaked to make this a less viable option. Personally, I feel like the current system is a little funky and would gladly accept an overhaul to make TX/RX more realistic, if more challenging.  Probe design would also have to be tweaked to account for constant power draw, hence the manual bandwidth limiter on antennae.  Your signal strength and antenna would set a maximum bandwidth, and you could further lower that if you didn't have the power requirements to do it in one go.

The probe should definitely still be controllable during data uplinks, I think that would just be frustrating more than anything else.

[Tyko]

Since time penalties aren't really game-friendly, another approach would be to increase the EC cost as signal degrades and THEN remove the ability to send partial transmissions - make it all or nothing. This would require that you included enough batteries to send the entire transmission.

That would seem to be a decent way to show impact of low signal strength without reducing the end quality of the data sent.

 

[eddiew]

4 hours ago, tater said:

Transmission should be dealt with like the MPL, or ISRU. You set it running, then go do something else. At some point you check the craft again, and the games "catches up" the data that has been sent based on the elapsed time.

Best answer, IMHO. Lower the EC/t required (or remove it) and have it take N hours/days/weeks/months to transmit back home. Yes, yes, we can timewarp - but we can timewarp x100 and cheese the EC/t too. If the time was really meaningful (like 200 days) many players would prefer to do other things meantime.

[Tyko]

3 hours ago, eddiew said:

Best answer, IMHO. Lower the EC/t required (or remove it) and have it take N hours/days/weeks/months to transmit back home. Yes, yes, we can timewarp - but we can timewarp x100 and cheese the EC/t too. If the time was really meaningful (like 200 days) many players would prefer to do other things meantime.

The problem is that to make the time "meaningful" it has to be so long as to require time warp. Why not increase power requirements? This ties back to core concepts in the game of building ships to meet your mission goals. The option of trading off weight for science gain is pretty straightforward and doesn't require time warp to work.

[tater]

7 hours ago, p1t1o said:

I think this is a gameplay allowance. Given that we have the ability in-game to accelerate time up to arbitrarily fast values, any temporal penalty is meaningless, the only meaningful penalty is quality of data.

If one is that fussed about the realistic effects of signal degradation and bandwidth, then the way "science" and research has been reduced to a simple numerical count must send one into conniptions!

It's not meaningless at all. It makes time move forward. Right now in career, you can unlock everything, and start sending things everywhere inside of a year after "discovering" rocketry. That's absurd. The game needs time warp to happen specifically to make time pass---just to make time pass. This also makes declines actually meaningful.

[StahnAileron]

@Prisma You quoted the thread title; did you even read the actual OP? (Or any other post in this thread, for that matter...) You immediately dug into the PHYSICAL meanings of the title, rather than the GAMEPLAY mechanics the OP is actually talking about.

The OP is taking about the amount of time it takes to transmit a given amount of science for a given connection quality ("Transmit Time") vs. the current system where you lose science based on the connection quality ("Signal quality"). The argument is that the amount of science you return should NOT be penalized because of the signal quality, but that it should just take more time to transmit. (e.g. Instead of 5 seconds at 100% quality, it'd take 10 seconds at 50% quality.)

So you first topic of discussion (quality) is a tangent and your second (signal delay) has no bearing on the topic at hand. (That being, "What penalties a player should have for a poor connection in regards to transmitting science.")

[natsirt721]

26 minutes ago, Prisma said:

Okay okay okay okay...

Two things:

1. You're telling us that no atmo, no problem.

Okay.

No atmosphere, no signal weakening. 

Don't forget plasma blackouts though...

The signal spreads out with distance, and so the signal gets weaker. It happens with antennas, but very little with satellite dishes. So, the signal should get weaker with the red and white antenna and the high gain antenna. 

2. Also, transit time.

Okay. So you want there to be a transit time for signals. Either have crews on you spaceship or invest hours in a video game planning, organizing, programming, revising and sending a science and maneuver sequence, only to find out that you probe already burned up. What's wrong with that picture?

See, the problem is transit time for signals from Kerbin to Duna is about 3 minutes. Waiting three minutes to turn on the frickin' wheels is not acceptable, at least for a video game. That one change would turn Kerbal Space Program into NASA Kids Space Mission Simulator.

Bit of an issue there. But otherwise, the satellite dish issue should be fixed.

 

Sure, no atmosphere no problem.  That's a difficult problem to model, but we can assume that Kerbin's own DSN is reasonably powerful and/or already compensates for atmospheric attenuation.  On the TX end, well, that's a difficult problem to model.  Remember the game uses soft occlusion i.e. signals through solid rock, so I don't think the gaseous atmosphere is going to have any bearing whatsoever on signal strength.  The game also has a difficulty option for plasma blackouts IIRC.

Second, transit time is another issue, but if you want that, go install remotech or something.  I don't think there is anything wrong with the current system as far as communication lag goes.  As someone who has played around with remotech a bit I can tell you that it is very annoying to arrive at your periapse only to find that you don't have a connection and that your Duna lander is now a deep space probe.  But, if you want that, the modding community can deliver.

Perhaps you confused transmit time with transit time?

EDIT: Couldn't have said it better myself, StahnAileron

Edited October 25, 2016 by natsirt721

[Zargg]

I agree with the OP this is a good idea. Transmission protocols that correct data loss are pretty trivial, it just comes at the expense of needing more time (and energy) to ensure all the correct data is transmitted. I think the in-game penalty is not so much time, but having the energy. If your signal is so weak that you are re-transmitting over and over, you risk the chance of running out of energy and only getting the partial science at that point.

[natsirt721]

I realize now that this would require a pretty major overhaul of the communications system, something I'm currently working through conceptually.  Will make another post on this forum when I'm finished.  Main concepts include:

• Signal strength is based on antenna gains and TX/RX power
  • Vessels will have a different uplink and downlink signal strength, and require an uplink for control
• Antennas have a 'nominal' bandwidth, and the effective bitrate is dependent on this bandwidth and and signal strength
• Signal strength is no longer a percent value, but measured in some quantitative unit (dBm/dBw or mW)
• The sun has a massive occlusion zone due to interference
• Tracking station upgrades will affect TX power, antenna size, bandwidth and other characteristics to improve the overall TX/RX of the DSN
• Atmospheric signal attenuation?
• Multiple communication bands?

Still on the drawing board for now.

[Tyko]

I'd consider removing the "partial transmission" option from antennas. Partial Transmission effectively negates any limitations you add regarding power requirements. If you want the balance to be low signal strength = higher power requirements = changes the way you design/fly missions, then forcing players to include enough power to complete the transmission in one shot fits well with that balance.

[StahnAileron]

On ‎10‎/‎27‎/‎2016 at 5:38 PM, natsirt721 said:

I realize now that this would require a pretty major overhaul of the communications system, something I'm currently working through conceptually.  Will make another post on this forum when I'm finished.  Main concepts include:

• Signal strength is based on antenna gains and TX/RX power
  • Vessels will have a different uplink and downlink signal strength, and require an uplink for control
• Antennas have a 'nominal' bandwidth, and the effective bitrate is dependent on this bandwidth and and signal strength
• Signal strength is no longer a percent value, but measured in some quantitative unit (dBm/dBw or mW)
• The sun has a massive occlusion zone due to interference
• Tracking station upgrades will affect TX power, antenna size, bandwidth and other characteristics to improve the overall TX/RX of the DSN
• Atmospheric signal attenuation?
• Multiple communication bands?

Still on the drawing board for now.

Signal strength should just be display in effective bandwidth (mit/s or whatever unit it is that KSP uses), IMHO. I don't really care how strong the signal is in terms of a power rating, I care about how strong it is in terms of how fast I could transmit data. Besides, 5mW between points A & B might not be the same as 5mW between points C & D because of antennae designs. (If you drill down to that much detail, at any rate.)

Actually, I do kinda wish the science reports had a time & power estimate when transmitting science so I'd have an idea of when is a good idea to send it back.