Perception speed versus reaction speed

[Vzearr]

The problem​

-----------------------
I've noticed, over the short course of me being a calc group member, that I've had to reject calculations that use this study as a perception speed timeframe, and I've recently begun to ask myself "why?", why do we treat perception blitz timeframes as being the same as reaction blitz timeframes, it's quite well known that human's have a reaction speed varying from 100 ms to 250 ms, with a much higher perception speed. For example, human's are capable of perceiving motorcycles moving at their top speed, which is above subsonic, this particular study suggests that humans can perceive objects moving at 76 m/s (over the course of a meter), so my question is, why don't we use it?​

 

[Vzearr]

Tagging calc group members (sorry): @DontTalkDT @Executor_N0 @Therefir @Ugarik @DMUA @Damage3245 @TheRustyOne @DemonGodMitchAubin @Jasonsith @Wokistan @Armorchompy @Migue79 @Psychomaster35 @Dark-Carioca @AbaddonTheDisappointment @Aguywhodoesthings @Agnaa @Dalesean027 @DemiiPowa @SeijiSetto @SunDaGamer

 

[Vzearr]

After searching, I see this has already been discussed, however, I do believe this should still be discussed again.

Tagging relevant staff that participated in the previous threads: @KingTempest @LephyrTheRevanchist @DarkDragonMedeus @GrathOfLux.

 

[DarkDragonMedeus]

Perception is merely the ability to see speed, but reaction requires actual physical reaction relative to a projectile and what not.

 

[Vzearr]

Perception is merely the ability to see speed, but reaction requires actual physical reaction relative to a projectile and what not.

Your point?

 

[GrathOfLux]

I do remember a thread being made regarding perception speed blitzing in the past that I was involved in.

I'll see to re-reading the thread shortly to refresh myself on the information, and I'll also look into any changes in the area. I recall it was a long time ago, and if I was to support or reject any specific claim in this regard, I would like to be sure that we are working off of the most up-to-date information. But I will note, plainly, that I thought at the time the above issue of perception speed blitzing being conflated with reaction blitz timeframes was a matter of incredulity. There is a substantial amount of empirical research into the time that it takes for visual information to be processed subjectively under numerous conditions that could be used for such a standard, and I remember reading up on this body of research in detail at the time, with the help of my university's academic resources - I also vividly recall it was rejected on the basis of decidedly non-empirical, non-peer reviewed resources, and these resources were preferred for no better reason I could discern than the fact that the more credible method gets results that feel too high. I only ended up leaving the thread in the end because I was busy with my studies and exhausted at how long the replies were/how many were being made each day.

It's not a standard that should be handled carelessly - it should be clearly delineated when something should qualify for a perception speed blitz over reaction speed blitzing, and I don't think that would always be an easy task - but I certainly think both the evidence and basic experience of the world can affirm that there's no good reason to conflate the two if that is how we still treat it.

 

[DontTalkDT]

Because human vision is complicated and you logically can't just slap a timeframe on perception as it obviously is dependend on various factors. See here and here.
TL;DR We know no really valid way to do perception blitzing stuff, as studies that just use flickering or "can you see the difference of FPS" don't really measure it. They measure something, but it isn't our ability to see fast moving objects. You would need a study that is actually on how fast an object a human can see (as in: based on actual objects that move fast) in order to establish that properly, so we are waiting on someone to find that. Until then we pretty much default to lower values.

 

[GrathOfLux]

Because human vision is complicated and you logically can't just slap a timeframe on perception as it obviously is dependend on various factors. See here and here.
TL;DR We know no really valid way to do perception blitzing stuff, as studies that just use flickering or "can you see the difference of FPS" don't really measure it. They measure something, but it isn't our ability to see fast moving objects. You would need a study that is actually on how fast an object a human can see (as in: based on actual objects that move fast) in order to establish that properly, so we are waiting on someone to find that. Until then we pretty much default to lower values.

You have claimed this to be the case in the past discussions on the topic, but it's still just a claim, and not one I believe you have made a compelling argument for.

These studies did not just use, say, flickering lights. They also tested on natural images - in other words, photographs of things in the actual world - and tested to see how quickly a change from one image to another could be detected subjectively, which is where numerous values mentioned in the prior discussions came from. Do you not see what the correlation is between the timeframe of our capacity to perceive something changing in front of us and our capacity to be perception blitzed? And even if you do continue off of the assumption that an extra variable here does introduce a discrepancy in real-world conditions (whatever it might be), why would you assert that simply conflating such feats with reaction speed blitzing would be more accurate to the real value of the feat, when it results in such hugely awkward margins of error? There is clearly a difference between how fast someone perceives a change subjectively and how quickly they can react to it - there is not such a clear-cut difference between how fast someone perceives a change subjectively and the flicker fusion threshold.

 

[DontTalkDT]

You have claimed this to be the case in the past discussions on the topic, but it's still just a claim, and not one I believe you have made a compelling argument for.

These studies did not just use, say, flickering lights. They also tested on natural images - in other words, photographs of things in the actual world - and tested to see how quickly a change from one image to another could be detected subjectively, which is where numerous values mentioned in the prior discussions came from. Do you not see what the correlation is between the timeframe of our capacity to perceive something changing in front of us and our capacity to be perception blitzed? And even if you do continue off of the assumption that an extra variable here does introduce a discrepancy in real-world conditions (whatever it might be),

I already explained in extremely great and lengthy detail why the test results don't work that easily. Address my actual existing arguments from the past threads in their entirety if you have doubts, so that we don't have to have this same debate from zero again, please.

why would you assert that simply conflating such feats with reaction speed blitzing would be more accurate to the real value of the feat, when it results in such hugely awkward margins of error? There is clearly a difference between how fast someone perceives a change subjectively and how quickly they can react to it - there is not such a clear-cut difference between how fast someone perceives a change subjectively and the flicker fusion threshold.

I'm not sure what gives you the idea that I think reaction speed is an accurate measure. It's simply that we don't use values that we don't have proper evidence to actually work and hence default to different methods of quantification.

 

[Vzearr]

We can perceive cars at their top speed, which move faster than 70 m/s.

 

[DontTalkDT]

We can perceive cars at their top speed, which move faster than 70 m/s.

We can also perceive hypersonic rockets flying by, assuming the light conditions and distance are appropriate.
We can perceive the moon which flies relative to us at over 1000 km/s.

Turns out neither an absolute time nor an absolute speed value work as a quantification. Size and distance quite obviously play a role as well, as do other factors.
And look at that, we are back at the last threads, which I would like you to read as they basically already went through all of this.

 

[Vzearr]

We can also perceive hypersonic rockets flying by, assuming the light conditions and distance are appropriate.
We can perceive the moon which flies relative to us at over 1000 km/s.

Turns out neither an absolute time nor an absolute speed value work as a quantification. Size and distance quite obviously play a role as well.
And look at that, we are back at the last threads, which I would like you to read as they basically already went through all of this.

We can perceive the humans in said cars, most reaction blitzes occur with humans moving FTE, if we can perceive a human moving faster than 70 m/s over the course of 1 meter, which I am willing to bet is possible, then our reaction speed as perception speed timeframes are bunk.

 

[DontTalkDT]

We can perceive the humans in said cars, most reaction blitzes occur with humans moving FTE, if we can perceive a human moving faster than 70 m/s over the course of 1 meter, which I am willing to bet is possible, then our reaction speed as perception speed timeframes are bunk.

Never said they weren't faster. In fact, I don't recall us saying they are 70 m/s, unless I miss some note on some instruction page. I only recall us saying that we vaguely rank such feats, under appropriate conditions, as Subsonic, which goes up all the way to 171.5 m/s (617 km/h) which is way faster than any car. Fairly sure I argued in past threads about how people on motorcycles should work as a decent example to establish a low end, given comparable conditions.

 

[Vzearr]

Never said they weren't faster. In fact, I don't recall us saying they are 70 m/s, unless I miss some note on some instruction page. I only recall us saying that we vaguely rank such feats, under appropriate conditions, as Subsonic, which goes up all the way to 171.5 m/s (617 km/h) which is way faster than any car. Fairly sure I argued in past threads about how people on motorcycles should work as a decent example to establish an end, given comparable conditions.

Sorry, why don't we use 0.0292s as a timeframe?

 

[DontTalkDT]

Sorry, why don't we use 0.0292s as a timeframe?

Many reasons, starting with the fact that obviously a single timeframe can't possibly work, as an object doesn't actually have to leave the field of view to become imperceivable at all.
See last thread for many many more details.

 

[Vzearr]

Starting with the fact that obviously a single timeframe can't possibly work, as an object doesn't actually have to leave the field of view to become imperceivable at all.

What exactly do you mean by that? It can only really be imperceivable if it's camouflaged or if it's extremely small, if we're talking about a perception blitz in fiction, 99% of the time, it is done by a human. I'm wondering why, I have to reject a feat that uses the timeframe it takes for a small ball to disappear, for a human to disappear, the other 1% can be figured out later, but we need to focus on the majority of perception calcs being bunk due to current standards.

 

[DontTalkDT]

What exactly do you mean by that? It can only really be imperceivable if it's camouflaged or if it's extremely small, if we're talking about a perception blitz in fiction, 99% of the time, it is done by a human. I'm wondering why, I have to reject a feat that uses the timeframe it takes for a small ball to disappear, for a human to disappear, the other 1% can be figured out later, but we need to focus on the majority of perception calcs being bunk due to current standards.

Thread 'VSBATTLE WIKI REVISION: PERCEPTION BLITZING'

Jul 17, 2023

I had a small discussion with Agnaa, DMUA, and Armorchompy to see why certain ways of calcing speed blitzing speed feats or speed blitzing time frames were unusable. I realized that virtually most, if not all, ways of calcing speed blitzing feats were not quite valid. We even have sections that forbid the use of time frames from sources where humans have perceived extremely fast objects.

So I was forced to find a new way.

Now the new method I came up with is starting to get used for speed blitzing calcs so I want all of us to discuss the guidelines on how to properly use this...

• Arnoldstone18
• Replies: 117
• Forum: Calc Group Discussion

• 

 

[Vzearr]

Thread 'VSBATTLE WIKI REVISION: PERCEPTION BLITZING'

Jul 17, 2023

I had a small discussion with Agnaa, DMUA, and Armorchompy to see why certain ways of calcing speed blitzing speed feats or speed blitzing time frames were unusable. I realized that virtually most, if not all, ways of calcing speed blitzing feats were not quite valid. We even have sections that forbid the use of time frames from sources where humans have perceived extremely fast objects.

So I was forced to find a new way.

Now the new method I came up with is starting to get used for speed blitzing calcs so I want all of us to discuss the guidelines on how to properly use this...

• Arnoldstone18
• Replies: 117
• Forum: Calc Group Discussion

• 

This thread targets mainly targets FPS/FFT. Can you link where you talk about what I'm talking about as of current.

 

[DontTalkDT]

This thread targets mainly targets FPS. Can you link where you talk about what I'm talking about as of current.

Ah, sorry, maybe I have misunderstood where you took that value from. Which exact part of the article resulted in 0.0292s?

 

[Vzearr]

Ah, sorry, maybe I have misunderstood where you took that value from. Which exact part of the article resulted in 0.0292s?

0.0292 seconds comes from subsonic reactions on the reactions page. Since that's the max we say humans can perceive, why not use it?

 

[DontTalkDT]

0.0292 seconds comes from subsonic reactions on the reactions page. Since that's the max we say humans can perceive, why not use it?

Because that has no basis. IIRC the perception rating are entirely created by the logic of "1 / speed value of the corresponding speed rating = perception time rating".
Like, it is 0.0292 seconds for the sole reason that we start Subsonic at 34.3 m/s and 1/34.3 = 0.0292.

Superhuman perception doesn't mean perception above what a human has, it means perception corresponding to the superhuman speed rating.

 

[Vzearr]

Because that has no basis. IIRC the perception rating are entirely created by the logic of "1 / speed value of the corresponding speed rating = perception time rating".
Like, it is 0.0292 seconds for the sole reason that we start Subsonic at 34.3 m/s and 1/34.3 = 0.0292.

Yeah, but in general, it isn't absurd to assume we could perceive a human moving at 34.3 m/s over the course of a meter, considering we can do that for a motorcycle.

 

[DontTalkDT]

Yeah, but in general, it isn't absurd to assume we could perceive a human moving at 34.3 m/s over the course of a meter, considering we can do that for a motorcycle.

Yes... which is why we would accept the argument that someone moving FTE moves faster than 34.3 m/s.

I suppose what you are trying to argue is that, if we can perceive someone on a motorcylce in the timeframe they need to drive past us, then we should also be able to perceive literally any other movement as long as it's done by a human-like character and in a timeframe at least that long.
However, that again runs into the issue of it being possible for something to not leave our FoV at all, yet move fast enough to be invisible. I.e. timeframe alone doesn't solve the issue.

 

[Vzearr]

Objects moving faster than a human can react to will generally appear as a blur. If a motorcycle moving at 70 m/s can be perceived as a slight blur over the course of a meter, so can a similarly sized human moving at that speed, which means a timeframe of 1 meter / 70 m/s would be fair. Your claim of "However, that again runs into the issue of it being possible for something to not leave our FoV at all, yet move fast enough to be invisible" Contradicts how human vision works. Complete invisibility without much speed is highly unlikely unless the object lacks sufficient size to leave a noticeable sight on the human eye, this does not however affect humans moving fte as they attain a fair size and would have to move faster then the frame rate humans perceive to not be registered as a blur, but registered as invisible.

 

[DontTalkDT]

Objects moving faster than a human can react to will generally appear as a blur. If a motorcycle moving at 70 m/s can be perceived as a slight blur over the course of a meter, so can a similarly sized human moving at that speed, which means a timeframe of 1 meter / 70 m/s would be fair. Your claim of "However, that again runs into the issue of it being possible for something to not leave our FoV at all, yet move fast enough to be invisible" Contradicts how human vision works. Complete invisibility without much speed is highly unlikely unless the object lacks sufficient size to leave a noticeable sight on the human eye, this does not however affect humans moving fte as they attain a fair size and would have to move faster then the frame rate humans perceive to not be registered as a blur, but registered as invisible.

I didn't say complete invisibility without speed. I said invisibility to which no timeframe applies that would limit the duration of the movement. That invisibility would be the result of speed.

The problem is that there must be limits to both. There is an limit to how short the timeframe of a movement may be so that a human can still see it, but there should also be a limit to how fast the movement may be. And the latter is independent of the former, in the sense that even if the movement is way longer than the before-mentioned timeframe, as long as it's above the maximum speed it will not be visible.

So if you have a character move 1000m without someone being able to see them, then you can't be sure that it happened in a timeframe of 1 meter / 70 m/s = 0.014s. (With a consequent speed of over Mach 200 for the 1000m movement)
It could also be that the movement took a longer time, but the speed is the one that was above the threshold of what could be seen. Like, it could easily be that the character moved at only Mach 100, but Mach 100 movement under those circumstances is too fast to see, despite the timeframe then being 0.028s.

 

[Vzearr]

If the objects movement exceeds 0.0292 seconds the human eye will perceive it. Therefore an object of relatively fair size must move a distance within 0.0292 seconds to surpass human perception, there isn't another way to do it without speed if its a fairly sized object.

 

[DontTalkDT]

If the objects movement exceeds 0.0292 seconds the human eye will perceive it.

No, that is not the case (and, like, would need a source if it were). The human eye isn't like a super-camera that takes one perfect image every 0.0292s. Something staying in the field of view of a person and remaining invisible for extended durations of time is possible given sufficient speed. Now you can really refer back to the old thread, since that was a point plenty debated.

 

[DarkDragonMedeus]

Another thing that's quite common is that perception of certain velocities does tend to become weaker at closer distances; for example, plenty of people have tracked the movements of jets moving at mach speeds when they look up at the skies. Or even meteors that travel 1.1 * 10^4 m/s are trackable by the human eye at great distances, but we do not grant Supersonic or High Hypersonic perception respectively for those. Especially if a human sized character was simply running through a road in front of you, it would be a mere blur or you would not even notice.

 

[Arkenis]

Instead of following the one study, is there not a way to measure a character's general perception speed level through their previous feats or statements? Not saying to use calcs where their speed is gotten, but just blatant feats like characters with special perception like Sharingan users or those who perceive at a slower rate than others? Two examples from JJK would be Gojo and Choso: Gojo perceives time and processes information much faster than others. Choso directly increases his dynamic visual acuity to perceive Naoya's Subsonic speeds.

 

[Agnaa]

why do we treat perception blitz timeframes as being the same as reaction blitz timeframes

Mainly because we don't have great support for a different number, as explained by DT.

We can perceive the humans in said cars, most reaction blitzes occur with humans moving FTE, if we can perceive a human moving faster than 70 m/s over the course of 1 meter, which I am willing to bet is possible, then our reaction speed as perception speed timeframes are bunk.

And most blitzes occur at closer distances than we'd typically track motorcycles, making them more difficult to see.

Instead of following the one study, is there not a way to measure a character's general perception speed level through their previous feats or statements? Not saying to use calcs where their speed is gotten, but just blatant feats like characters with special perception like Sharingan users or those who perceive at a slower rate than others? Two examples from JJK would be Gojo and Choso: Gojo perceives time and processes information much faster than others. Choso directly increases his dynamic visual acuity to perceive Naoya's Subsonic speeds.

Feels way off topic. We already allow this, and this thread is about our baseline assumptions for this sort of thing, and about drawing a distinction between reaction blitzes and perception blitzes.

 

[Arkenis]

Feels way off topic. We already allow this, and this thread is about our baseline assumptions for this sort of thing, and about drawing a distinction between reaction blitzes and perception blitzes.

Yeah I'm saying the character should just be based off its feats instead of assuming a baseline since it isn't attainable.

 

[Agnaa]

The vast majority of characters don't have stated numbers we could use for that sorta thing.

 

[Flashlight237]

I don't have a tachometer on me, so I can't say I can fully hop into this discussion without sounding stupid. There are things out there that appear as a blur without really going 70 m/s, never mind 76.2 m/s according to the OP, most notably ceiling fans and fans (and more recently fidget spinners), which we see as a blur due to angular velocity rather than just plain velocity. I bring this up because I feel things that spin around will provide a more accurate picture than things that have a linear velocity like cars, as blurs are more like when something exceeds how many images our eyes can process than anything else. Keep in mind that we are humans, not reaction-speed gods like speedrunners (ex. SimpleFlips) and that one guy who sliced a BB pellet in half.

 

[CloverDragon03]

I'm seeing talks of how we can see super fast stuff from a distance and... I don't see how that's relevant?

Like, if we're talking perception blitzes, those are portrayed in fiction as happening up close, not with something far away. So talking about how we can see meteors moving from far away seems kinda derail-y tbh

 

[Agnaa]

I'm seeing talks of how we can see super fast stuff from a distance and... I don't see how that's relevant?

Like, if we're talking perception blitzes, those are portrayed in fiction as happening up close, not with something far away. So talking about how we can see meteors moving from far away seems kinda derail-y tbh

Yeah but if we get to close distances, then we run into things like "moving your hand from side to side really quickly", and fans.

 

[CloverDragon03]

Yeah but if we get to close distances, then we run into things like "moving your hand from side to side really quickly", and fans.

But perception blitzes aren't about moving as a blur, they're about moving so fast that you can't even see them. So I don't see how that's relevant

 

[Agnaa]

But perception blitzes aren't about moving as a blur, they're about moving so fast that you can't even see them. So I don't see how that's relevant

At which point we don't have any reliable way of determining the timeframe needed to do that for a human-sized object in decent brightness from within 0.5-3 meters.

 

[CloverDragon03]

I mean you're free to think that, but isn't what Grath's saying literally for handling that? In which case, you're just kinda claiming what you're claiming without substance. And I think that study's a much better thing to go off of than equating reactions to perceptions

 

[Agnaa]

I mean you're free to think that, but isn't what Grath's saying literally for handling that? In which case, you're just kinda claiming what you're claiming without substance. And I think that study's a much better thing to go off of than equating reactions to perceptions

Those are relying on photographs. With most common experimental setups, human-sized objects in those wouldn't take up the same portion of vision as they would if a person was standing within 50-300cm of me.

Also, given the mention of them being photographs that switched between states, it seems like a bit of a different case than someone engaging in movement, which I think would make it plausibly different. The badge on someone's shirt changing colour for 0.02 seconds would presumably be less recognisable than that person moving 1m to the right, then back, over 0.02 seconds.

But perception blitzes aren't about moving as a blur, they're about moving so fast that you can't even see them. So I don't see how that's relevant

(Also, I think most cases don't have them, like, literally being invisible; many of them just have a character shocked at a sudden appearance, an actual statement ruling out them being a blur is uncommon)

 

[CloverDragon03]

Those are relying on photographs. With most common experimental setups, human-sized objects in those wouldn't take up the same portion of vision as they would if a person was standing within 50-300cm of me.

Also, given the mention of them being photographs that switched between states, it seems like a bit of a different case than someone engaging in movement, which I think would make it plausibly different. The badge on someone's shirt changing colour for 0.02 seconds would presumably be less recognisable than that person moving 1m to the right, then back, over 0.02 seconds.

The thing is that while it's not perfect, I don't see why conflating perception speed with reaction speed is a better approach. It honestly kinda downplays perception blitzes because there's a stark difference between the two

(Also, I think most cases don't have them, like, literally being invisible; many of them just have a character shocked at a sudden appearance, an actual statement ruling out them being a blur is uncommon)

Kinda, but also not really. There are also multiple examples of characters being like "I didn't even see a thing!" or something like that. Here's one off the top of my head.