[qsxcv]

Quote:

Originally Posted by CorruptBE

The thing is, with Synapse this movement also feels "disconnected". Like constantly changing direction left/right but only by like (on average) a pixel feels really "uncontrollable" unless I do it really slow.

I also get the same feeling with every other 3.5 G mouse I've tried in Synapse (Deathadders).

can you save and post logs of a medium speed swipe and also of really slow motion

 

[CorruptBE]

Derp... hang on this might take a while, you have to keep in mind I have to uninstall/reinstall with reboots for every test run xD

(thank god for SSD's)

EDIT:

Here's a bunch of logs,

Synapse_vs_Clean_Logs.zip 143k .zip file


Graphs also look alot different above 4 meter / second. Included, fast swipes, slow, long, tiny movements, ...

All done on the same hardware, same USB port (Abyssus 2014).

 

[Nilizum]

Quote:

Originally Posted by CorruptBE

Puretrak Talent. Also SooperSekretz Firmwarez Edition, si si.

Would be nice to have some serious feedback on that circular movement. If it really is bad behavior... Time to spam their Twitter/Facebook and RMA... it's what got us that firmware update last time xD

Next up in Synapse: "Tournament DPI Scaling - More PRECISION (less choices, just halving steps xD)"

(I honestly would ditch my FK1 if I could use this Abyssus at a decent 800 or 900 step)

Have you tried Povohat's mouse filter? I like it a lot with Abyssus 2014.

Also the uncontrollable left/right is the jitter the sensor has. I have a PureTrak Talent and I know exactly what you're talking about. It's present even with a mouse filter, just less jagged and smoothed out of course. I mean, I'm usually the guy that doesn't blame sensor performance/architect before anything else, but the sensor in the Abyssus is really something awful that REALLY does affect precision/cursor control.

 

[qsxcv]

Quote:

Originally Posted by CorruptBE

Here's a bunch of logs,

Synapse_vs_Clean_Logs.zip 143k .zip file


Graphs also look alot different above 4 meter / second. Included, fast swipes, slow, long, tiny movements, ...

All done on the same hardware, same USB port (Abyssus 2014).

from your synapse fast swipe:

dude what??
i dont even


the presence of 0,0 motion reports strongly indicates that there's either mcu or driver count manipulation (roughly, counts from sensor =/= mousetester data). this is probably related to why you feel small pixels movements to be different.

(you also get 0,0 when you click a button/scroll while the mouse isn't moving. could be why the nosynapse file had 0,0's at the end)

pretty sure what's going on is that when you have synapse, the sensor is actually set to a higher dpi and the mcu and/or driver scales it down. this would also explain the flat line in the highspeed synapse plot

 

[CorruptBE]

Well yeah, I was somewhat hinting towards their software doing something wrong with the scaling ages ago. Never had these problems on older DA's when I used the legacy software (which didn't allow 100 steps, only "halving" or "quarters", 450, 900, 1800, 3500).

It least now we're getting somewhere (fact-wise) to point it out... and perhaps even show Razer that it's messed up.

I still wonder though if the same behavior happens on their 4G or other new mice (Here I'm hinting towards them obviously optimizing the new software for their newer mice and forgetting about their older generations). Though I have no intention of buying a new DA just for testing.

 

[MaximilianKohler]

Quote:

Originally Posted by qsxcv

dont you agree with what popups said though?
i.e. you prefer finalmouse over logitech's 3366mice even though every metric of performance that we have so far and also comments from insiders seem to indicate otherwise

I don't think there's anywhere near enough information to make that claim.

I see it as you guys essentially playing with various small parts of a car without anyone having an idea of which parts of the car are the most important to its ultimate performance. A good example is that mousetester - which is currently the best tool we have - shows the DA4g and the 3366 to be mostly identical. But they feel extremely different ingame. IMO this is the most meaningful difference we need to be able to measure and account for. Until then we have nowhere near the full picture.

Comments from insiders don't mean much to me considering the fact that virtually every gaming mouse released in the past 10 years has been terrible in one way or another, and hasn't been able to match, much less surpass, 10+ year old mice. They're obviously missing something very important.

 

[CorruptBE]

MaximilianKohler

Make yourself useful and Install Synapse, do some mousetester runs with your 4G DA and then uninstall Synapse and rerun the tests

 

[MaximilianKohler]

I returned the mice already.

 

[CorruptBE]

Couldn't help myself and dropped the ball:

https://twitter.com/i/web/status/622795173219729408

 

[MaximilianKohler]

You can link directly to posts. The post number is a link to the post.

 

[CorruptBE]

Tried that before, for some reason it redirects to the page instead of the post (Firefox).

Still would be nice if someone tested this on a newer 4G DA though... any candidates

 

[MaximilianKohler]

I use firefox and it works for me. Maybe it's some addon you have.

 

[qsxcv]

Quote:

A good example is that mousetester - which is currently the best tool we have - shows the DA4g and the 3366 to be mostly identical. But they feel extremely different ingame. IMO this is the most meaningful difference we need to be able to measure and account for. Until then we have nowhere near the full picture.

we've been through this before... you really don't see much if you just look at a zoomed out plot of a fast swipe. likewise if you draw circles in mousetester
Quote:

Comments from insiders don't mean much to me considering the fact that virtually every gaming mouse released in the past 10 years has been terrible in one way or another, and hasn't been able to match, much less surpass, 10+ year old mice. They're obviously missing something very important.

i dont mean comments as in "sensor x" is better than "sensor y".
i mean stuff like how the 3988/3310 has more variance than other sensors, which sensors/srom revisions have smoothing (not the feel you have, the actual antijitter algorithm), etc...
Quote:

Originally Posted by MaximilianKohler

I see it as you guys essentially playing with various small parts of a car without anyone having an idea of which parts of the car are the most important to its ultimate performance.

well here's an exhaustive list of stuff that affects tracking and perception of tracking

which sensor and srom (also variations between the same sensor)
> sensor configuration
> power quality from motherboard
> electrical things/pcb design
> illumination and lens quality
> lens/sensor vertical positioning and alignment
> mcu firmware, especially the timing of spi communications, not buffering the usb motion data
> mcu and/or driver count processing
preconceived notions/biases
shape/weight/weight distribution
sensor placement, some people are more sensitive to this than others though

for all the ones with >, it's really a case of not screwing up rather than doing something special that improves performance.

 

[MaximilianKohler]

So you think you can identify differences between the 3366 and the DA4g with mousetester and/or your other devices? And you are confident these differences are the main factor in why they feel/perform so differently?

 

[uaokkkkkkkk]

Quote:

Originally Posted by MaximilianKohler

We already established that "newer != better" when it comes to mice. There are certain things like higher malfunction speed and other technical things that newer mice are good at. But there are plenty of very important things they are much worse at than very old mice.

The copperhead was one of the best mice I've ever used. Only its extremely low malfunction speed made it unusable. Other than that it felt like an aimbot.

And since the DA4g feels very similar, it seems likely you'd get similar results.

You brought up a good point. I need to do the tests with the very first/earliest firmwares for the Copperhead.
Quote:

Originally Posted by MaximilianKohler

So you think you can identify differences between the 3366 and the DA4g with mousetester and/or your other devices? And you are confident these differences are the main factor in why they feel/perform so differently?

Going in circles there.

 

[qsxcv]

Quote:

Originally Posted by MaximilianKohler

So you think you can identify differences between the 3366 and the DA4g with mousetester and/or your other devices? And you are confident these differences are the main factor in why they feel/perform so differently?

not sure what you're trying to get at... 3366 and da4g's 3988 are different sensors. you can distinguish between the two just by looking at the sensor lens

but for instance if i had two 3310 mice, i could figure out most of the things they're doing differently.

 

[uaokkkkkkkk]

My god there is 20 different firmwares for the Copperhead. I just had to share this.

6.30


6.14e


It's a little broken...

 

[wareya]

>sensor placement, some people are more sensitive to this than others though

I would say that forwards/backwards sensor placement makes a huge difference. It gives wrist styles a different horizontal dpi than arm styles. A sensor should be placed between the point where people "feel" like they're tracking when they grip it, and slightly towards whichever left/right direction they tend to tilt it under stress (I tilt it rightwards). This is one of the only legitimate arguments in favor of forced grip style mice out there.

One of the reasons I like the FM so much is because the grip pulls my thumb and pinky backwards, so even though the sensor is in a similar place to the DA, it feels less "swamp cursor"-like to my fingertip grip and wrist aiming style, because the DA pushes my fingers forwards from the center of gravity of the mouse because of its harsh taper. The DPI doesn't seem to change when I change aiming styles, but with the DA, wrist movements feel less sensitive.

The KPM isn't perfect in terms of grip but still an improvement for my grip over the DA because the sensor is in a better place (towards the right side of the mouse, and a couple mm further forwards from where my fingers rest).

Weight and center of gravity can affect this because of shifting your grip to compensate for the different force requirements etc.

http://i.imgur.com/Ucu5Ft5.png
http://www.hardwareheaven.com/wp-content/uploads/2014/08/roccat-kone-pure-camo-base.jpg
http://www.custompcreview.com/wp-content/uploads/2013/03/razer-deathadder-2013-4g-optical-gaming-mouse-custom-pc-review-12.jpg

https://imgrush.com/dECsImgeDPNb
https://imgrush.com/iuozX7eRgKj4
https://imgrush.com/bmBKT2LcNgYx

 

[Nilizum]

Everyone should be able to notice the difference between sensor placement. Those that don't probably don't calibrate accordingly for a proper comparison, but it's there. Given a similar shape, sensor placement +- 3-5mm makes a big difference to cursor control.

 

[Nilizum]

Quote:

Originally Posted by qsxcv

not sure what you're trying to get at... 3366 and da4g's 3988 are different sensors. you can distinguish between the two just by looking at the sensor lens

but for instance if i had two 3310 mice, i could figure out most of the things they're doing differently.

I think he means in an actual game environment. Honestly I agree with Kohler. I would not be able to feel the difference between s3988, 3310, or 3366. I would feel a difference if there was heavy prediction, or crazy jitter like the cheap razer abyssus(es). I'm gonna reiterate my point that any differences felt would be sensor position, and mousefeet friction or cable drag.