Communication speed of wireless to wired signals? (eg mice)

[MadAd]

Some of our gaming brethren seem to think that using remote wireless mice causes their gaming reactions to lag. Personally I cant see how it would be any more than 1ms, if that. So my question is, how fast does a remote mouse communicate compared to a wired version?

The answer to this would probably be along the lines of the speed of radio waves vs the speed of electrons in a wire and thats when I thought that this question would be an easy one for some of you knowlegable guys here.

An area of guesswork is how much overhead would the convertion from radio to wired at the tranciever make on the overall latency (since all remote mice transmit to a base unit that plugs into a ps2 or usb socket) which would have an effect on any answer.

Thanks in advance!

Ad

 

[MadAd]

hmmm - ok so it seems that measuring the speed of an electron in a wire is not that simple - no wonder no-ones replied, lol.

The easy bit: It seems that radio waves travel close to C at 299,792.5 Km/s (which is 299,792,500 m/s). If my basic arithmetic is right then thats 0.000000003335640485 s/m, or about 3.3ns per meter. Assuming that the mouse is 2 meters from the base station then thats 6.6ns for the signal to travel for that 2m segment.


The hard bits (help!): lets assume the wire from the base station to the pc is 3 meters of copper wire, and that the non wireless mouse (control comparison) is 5 meters of the same wire from mouse to pc. Fortunatly my (logitech) base station has a rating stamped on it of 5V 100mA so we can plug those straight into the equation. Taking the wire diameter to be 0.3mm (0.03cm) we get;

                  0.1
----------------------------------------------   =  0.00156 cm/s = 5.616 cm/hour
8.5x10e22 x 1.6x10e-19 x 0.015 x pi 

(equation taken from here)


Now this is where im stuck. #1. the above source suggests the electron drift is very slow (slower than maple syrup) and 5.616 cm/hour certainly bears that out (assuming no calculation error) but the speed of the energy transfer is very fast, since we are measuring only a miniscule input that creates an almost instantanious output at the other end of the wire, how do i calculate that from what i have already?

A source generally suggests that the actual speed is near C (similar to the radio wave) but i cant figure how to calculate how close, nor if I should just define that based on a 3 or 5 meter length of wire just like the radio wave (@3.3ns/m)? The limiting factor would be the ability of the base station to regulate the output it seems.

#2. The base station is an unknown quantity to me. All i know about them is the rating, and that the air side operates between 27.045-27.195Mhz. How can I estimate the time it takes for the signal to be recieved, translated and output to the fixed wire? Anyone got any good guesses?

#2b. Would the granularity of the 27mhz band make any changes to the 3.3ns/m calculated in the initial section?



Finally, whats being ignored: 1. mouse cables are rarely 5m but for simplicities sake I will ignore any db loss of an extension connector.

2. any difference in the conversion of the signal to air or cable at the mouse end, for simplicities sake we can assume that the conversion from physical movement to signal is the same in both cases (anyone see any reason why not?).

3. There may be a variation in cable thickness between the wired mouse and the base to pc portion of the tranciever however we can assume there is no difference as it will be small.


Can someone please make any more sense of this?

 

[MadAd]

If the electricity in a wire did move at the speed of light, then would the correct equation not be C minus the electron drift for any given length?

 

[pyrojunkie]

Originally posted by: MadAd
2. any difference in the conversion of the signal to air or cable at the mouse end, for simplicities sake we can assume that the conversion from physical movement to signal is the same in both cases (anyone see any reason why not?).

A signal by wire is sent straight to the computer. A wireless signal is converted to air and then sent to the computer. Because of such "lag", a wired connection will always be speedier.

 

[f95toli]

First of all: The speed of an electron is not relevant for calculating the "speed" of a signal. The speed is of course c for wireless and about 60-90% of c in a cable (in copper it is about 90%). So the speed of the tranmission is not relevant at all. I guess the extra electronics a wireless mouse (transmitter etc) might slow things down a bit but not much.
AFAIK there is no technical reason why a wireless mouse should be slower than an ordinary mouse. I think that any difference could be attributed to a) bad drivers and b) the fact that in order to save power (battery) most wireless mice might have somewhat slower electronics than an ordinary mouse,

 

[silverpig]

Electrons don't actually "flow" like water down a pipe. It's more of a wavey-sloshy type of thing. You can send an electrical signal down a wire and not have the electrons at your end ever reach my end.

Basically though it's not a big deal. Those guys say it's laggy to cover their lame asses when they get fragged.

 

[Shalmanese]

I think the primary reasons a wireless mouse lags more is that te wireless signal is more complex due there being more noise in the wireless medium. This means the computer has to do more decoding of the signal in order to get a usable vector out of it so it induces a slight lag.

 

[IaPuP]

Originally posted by: Shalmanese
I think the primary reasons a wireless mouse lags more is that te wireless signal is more complex due there being more noise in the wireless medium. This means the computer has to do more decoding of the signal in order to get a usable vector out of it so it induces a slight lag.

First of all, I highly doubt there is a significant lag, second, the computer has NO IDEA whether the mouse is wired or wireless and therefore shouldn't have to bother with "decoding the signal".

I am quite sure that the Radio Frequency wireless mice use a simple FSK (frequency shift key) modulation, which is very easily decoded by a $0.10 IC chip with no delay at all.

If (and this is a big IF) there is any delay, I would think it might be because of power-saving features. They could save a ton of power by transmitting in 'bursts' every few milliseconds, however; the sample rate of "standard" USB mice is only about 100Hz anyway as far as I'm aware, so unless the bursts were slower than that, I find it VERY hard to believe that it would cause problems.

Now, if there was substantial interference, I could see some of the signals getting dropped, but that shouldn't be a problem if you're not moving too far away with the wireless...

Eric

 

[Pulsar]

Speaking as a hardcore gamer AND someone who has gone both mouse routes, I can expound on the issue.

Many wireless mice BEFORE the last generation - MX700 (Logitech) and the revision 2 microsoft optical had severe lag.

The newer mice, like the logitech MX700, the logitech dual optical, and rev 2 microsoft optical wireless, etc, have latency no greater than that of a "wired" mouse.

A large part of what you see is the sampling rate. Older optical mice did not have a high enough sampling rate - their optical eye speed was the limit. Whereas an older ball mouse directly connected to the computer could sample up to 200 hz if set by the program PS2 rate, which gave you performance that was smooth as SILK.

The newer mice have fixed the latency issue created by the optical eye and the slowdown caused by poor A/D coversion and reception of the wireless units. Now, you'll actually get better precision from the dual optical models that some of the companies put out.

I will highly recommend the logitech MX700 as a mouse that is near perfect.

 

[MadAd]

Thank you all for a fantastic response - after some thought perhaps I should focus on the bottleneck and work up.


IaPuP: What do you mean by 'no delay at all' please? there must be a little delay? Also, how would the fact that it works in the 27Mhz band affect this? Would any transmission conversion from air to wired be limited to 27,000 cycles per second? (excuse my lack of knowledge here).


LsDPulsar: Ahh, now that does make sense re: Prior generations, however as above, even tho ps2 works at up to 200hz, would the fact that it transmits on 27Mhz cause that to be the bottleneck?


f95toli: 90% of C you say? do you have any equations where I can work that out more exactly please?


Thanks a lot all, keep em coming

**edit** updated misquoted Mhz to Hz

 

[ZeroNine8]

Just sort of a sanity check:

On average, humans have a reaction time of 0.25 seconds to a visual stimulus, 0.17 for a audio stimulus, and 0.15 seconds for a touch stimulus

So even if you consider your reaction times to be the pinnacle of human evolution, it is still relatively slow compared to machine times. The USB specification allows a maximum delay between signals of 26 nanoseconds (~40MHz if I calculated it right), while the PS2 Rate maximum of 200Hz (not MHz) is "smooth" for our limited perceptions. Considering in a best case scenario, your reactions are around 10Hz and your vision is in the 60-80Hz range, a 27MHz wireless band can transmit. Looking at what old PS2 sampling times were set at, being able to detect 'lag' is certainly reasonable, especially in fast-paced gaming where reflexes are finely tuned. In all reality, above 200Hz there is no discernable 'lag' due to input device update frequency. I suspect that the lag on older wireless mice was due to bursting the data rather than continuous 27MHz updating of the position that the device is theoretically capable of (which is far far higher than necessary). Just for some back of the envelope calculations,

lets say 200Hz is the minimum update frequency we want to tolerate (generally accepted as being smooth for even fast-paced gaming)

200Hz -> 0.005s between position updates

0.0000000066s for the wireless signal to travel (6.6ns)
which leaves 0.004999993s for encoding and decoding, assuming encoding and decoding take roughly the same amount of time, 0.002499997s for each of those operations, meaning that the bare minimum (I'm not incredibly familiar with DSP) is in the ~400Hz range, several orders of magnitude below the update frequency of the transmitters. Regardless of how many bits must be sent to fully communicate the mouse information (as long as it's less than about 67K ) at each update, you still are able to achieve performance on par with a 200Hz PS2 mouse with the hardware.

I think the lag issues you are referring to are more than likely due to bursting of the signal to save power or possibly software/drivers, rather than continuously updating as fast as the hardware can.

From the looks of it, the theoretical hardware bottlenecks are nowhere even close to the detectable threshold of humans.

 

[blahblah99]

The amount of time it takes to process the data into a form that can be sent wirelessly and the amount of time it takes to decode that data will be far more than the amount of time the data travels through the air wirelessly..

Now, with that in mind, most wireless mouse/keyboards run off AA batteries or AAA batteries, which to me, hints that the clock speed of those devices will be less than 1Mhz, or even less than 100Khz to conserve battery power.

Depending on the SAMPLING RATE AND RESOLUTION of the mouse, it has to packetize X amount of data, send it wirelessly (almost next to nothing delay compared to other functions), and the receiver has to depacketize it and send it to the computer.

Low sampling rate = jumpy mouse cursor when moving it across the screen quickly. Low resolution (say, 8-bit versus 14-bit for example) = blockly movements when trying to draw an image.

Anyhow, I personally think that the delay of wireless mouse is much much less than the reaction time of a human being. Say the fastest reaction time of humans is 1/10th of a second - and that is FAST. If the delay through the wireless mouse is 1/3rd of that, that means from the time you see a change on the video screen to the time you react, 0.100 + 0.033 = 0.133 seconds would have passed.

In reality, I doubt anyone has reaction times of 0.1 seconds. Even if someone did have that reaction time, 0.133 seconds means nothing when latency to a server is +20ms in the case of online gaming.

 

[silverpig]

Originally posted by: MadAd
Thank you all for a fantastic response - after some thought perhaps I should focus on the bottleneck and work up.


IaPuP: What do you mean by 'no delay at all' please? there must be a little delay? Also, how would the fact that it works in the 27Mhz band affect this? Would any transmission conversion from air to wired be limited to 27,000 cycles per second? (excuse my lack of knowledge here). Certainly that is slower than 100Mhz (your quoted usb freq).

LsDPulsar: Ahh, now that does make sense re: Prior generations, however as above, even tho ps2 works at up to 200Mhz, would the fact that it transmits on 27Mhz cause that to be the bottleneck?


f95toli: 90% of C you say? do you have any equations where I can work that out more exactly please?


Thanks a lot all, keep em coming

If you want some equations you'll have to look up a book on condensed matter physics. It's not easy stuff.

 

[Pulsar]

>>LsDPulsar: Ahh, now that does make sense re: Prior generations, however as above, even tho ps2 works at up
>>to 200Mhz, would the fact that it transmits on 27Mhz cause that to be the bottleneck?

Note - that's 200 hz, not Mhz. Your standard ps/2 style connection normally only samples at 40 hz, but can be changed to 200. And the difference between the two is huge. Download ps2rate and try it for yourself.

I'm not sure if 200hz is the holy grail or not - it's just the highest a ps/2 style mouse connection can physically sample.

 

[chorner]

There is definitly a perceivable difference in performance between the old and current gen. wireless mice produced by both Microsoft and Logitech.

I know for a fact -as I have been in direct contact with an ASIC designer at MS - that the issues with the older gen. Microsoft mice in specific were due to battery longevity issues within the circuitry. They just couldn't get a high and steady enough refresh rate in order to maintain smooth onscreen cursor movement without severely reducing battery life. This caused Microsoft to resort to a burstable design. With the Microsoft Wireless Bluetooth mouse for eg. if you were to monitor the signal processing rate in Hz you would find that it peaks at 800Hz very briefly when the mouse "awakes", down to about 90-120Hz in spike intervals during 'normal' usage, and down to as low as 20Hz with a low level (think :speed) of mouse input.

As for your mathematical theories as to which is faster : direct wiring or wireless technology. The obvious answer ( and can be proven by math, or even just listening to your local radio station ) direct wiring will always be faster - from point to point - that is. As for the distances obtained between your wireless base, and your mouse or keyboard .. from my experience I don't think I've seen a wireless base that has been set any further than a meter distance at max from the periferals. This of course will tell you that there should be very little difference between cabled and "wireless" mice; but of course the cabled mouse will always be faster.

On another note ... using a wireless device you are of course more likely to run into transmission errors, which may also be a contributing factor to the feeling of lag.

 

[MadAd]

ZeroNine8 & LsDPulsar Oops, my bad, yes 200Hz is not 200Mhz, and a lot slower than 27 meg - ty.

Yes - Im aware of the wholly inefficient human response time. Occasionally in gamers forums someone will post a 'test your reaction time' test, and in all cases where i see people doing this, I can say that a seasoned gamer has an eye-to-click response averaging 220ms (some faster, some slower, but thats the average i see people post) however I was interested in the actual physical speed of the devices themselves, being the curious person I am Its as much a journey of learning and discovery for me, as well as answering the question.

Anyway, good angle ZeroNine8 & blahblah99, I see where your coming from, ty - the point is that the hardware has to work quicker or we cry lagg lagg, so as before, its back to the 'how quick' question.

silverpig After reading many head warping threads on quantumn physics and the like in this forum, a bit of condensed matter should be a walk in the park, hehehe

LsDPulsar Yes, 40hz (25ms) was indeed the windows 9x/me (and logitech) default for ps2, i should have remembered that from my win9x days (I even have an old copy of ps2rate). It appears that w2k has a rate of 60hz, and after a reg snoop (plus testing with ps2rate) XP with a logitech optical returns a default of 100hz(10ms).

Looking a bit further, apparently usb mice run at 125hz (8ms) default so either way, usb or ps2, they must be the limiting factor in any response time critical path.

chornerThats interesting, thank you for that- so originally even when it was motionless it was sampling at full rate and struggling but now if it lays idle then the rate can go down as low as 20hz. I guess even at 800 then its wasted if the port max is 200? Anyway, yes overall a wired will always be faster than a cordless however i initially wanted to find what the difference was - but thats now immaterial if the best the port can do is 200hz (5ms).


Ok, thank you all very much, hopefully I have enough information (estimated or raw) to put together a sum for the envelope of each device.

 

[xts3]

I think you're all missing an important factor here: Human adaptation and compensation for 'lag', I've seen people on 30ms connections get absolutely assraped by people playing @ 120ms. The difference between a wired and new generation wireless mouse are negligable. If you really want to find out if it makes a difference in your gaming you have to game with seperate mice over a period of weeks for each mouse then compare your frags per hour, deaths, shot accuracy, etc. And only then you might have the right to say "I suck because of mouse 'lag'" justification!!!!