Time Sleuth Lag Tester

The place for all discussion on gaming hardware
strayan
Posts: 671
Joined: Sun Mar 19, 2017 8:33 pm

Re: Time Sleuth Lag Tester

Post by strayan »

Xyga wrote: PS: question to the general, do the TS readings vary a few ms with the display's brightness ? (like the LB for instance)
It does! The lower the brightness the lower the lag! I can get it down to 22ms at the lowest brightness setting.

Also interesting and highly unexpected is 24.97ms of lag when displaying 1080i content.

I love this thing. So many unexpected results!
User avatar
Xyga
Posts: 7181
Joined: Tue Nov 05, 2013 8:22 pm
Location: block

Re: Time Sleuth Lag Tester

Post by Xyga »

Oh, the LB is the opposite, at lower brightness it reads a couple more ms.
Strikers1945guy wrote:"Do we....eat chicken balls?!"
BONKERS
Posts: 417
Joined: Thu Dec 11, 2014 10:41 am

Re: Time Sleuth Lag Tester

Post by BONKERS »

Picked up one of these for my brother as a present this year.
https://www.bestbuy.com/site/sharp-40-c ... 3415654968
Took a gamble that it probably had good input lag based on other Roku based TVs.

It does but, it has a bug in the firmware similar to RTing's review of the 2019 TCL 325 1080p series where Game Mode doesn't engage or disengage properly. If you switch resolutions/inputs or restart the TV with Game Mode enabled then it works and you get 15ms input lag readings. If you go back to the home menu and back into the HDMI input, it returns to non-game mode even though it is enabled. And you will get 30ms of lag instead. 75ms with interlaced resolutions with the bug. Using an HDMI switch and cycling back to your desired input seems to fix it.

Worst case scenario 30ms is still perfectly fine, and 45 with interlaced resolutions is really good (every time I tried to get the bug to appear with interlace resolutions it wasn't repeatable like with progressive resolutions. )



Has anyone else had luck with TVs with CCFL backlights yet though? I still can't get any readings from the two CCFL LCDs I have. I can get the blue like to flash in sync like it's supposed to, but I consistently get like 0.20ms latency readings. Which can't be right.
strayan
Posts: 671
Joined: Sun Mar 19, 2017 8:33 pm

Re: Time Sleuth Lag Tester

Post by strayan »

BONKERS wrote: Has anyone else had luck with TVs with CCFL backlights yet though? I still can't get any readings from the two CCFL LCDs I have.
Unfortunately, no.
User avatar
xeos
Posts: 167
Joined: Thu Oct 20, 2016 10:38 pm
Location: San Diego, California

Re: Time Sleuth Lag Tester

Post by xeos »

For those of you who would like to measure lag but find the price of the timeslueth too high, how about free instead? I have a project that uses a raspberry pi zero to achieve the same thing entirely using software, no soldering required. If you don't already have a raspberry pi the zero model goes for about $5USD.

you can read about it here.

https://alantechreview.blogspot.com/202 ... berry.html
User avatar
orange808
Posts: 3196
Joined: Sat Aug 20, 2016 5:43 am

Re: Time Sleuth Lag Tester

Post by orange808 »

xeos wrote:For those of you who would like to measure lag but find the price of the timeslueth too high, how about free instead? I have a project that uses a raspberry pi zero to achieve the same thing entirely using software, no soldering required. If you don't already have a raspberry pi the zero model goes for about $5USD.

you can read about it here.

https://alantechreview.blogspot.com/202 ... berry.html
Nice. Not as user friendly as my Time Sleuth, though.

Watch out for rolling shutter. :) Ideally, users will want to mount the pi so it partially blocks the display to photograph the bars and led simultaneously.

Probably better to announce it in its own thread, though.
We apologise for the inconvenience
User avatar
xeos
Posts: 167
Joined: Thu Oct 20, 2016 10:38 pm
Location: San Diego, California

Re: Time Sleuth Lag Tester

Post by xeos »

orange808 wrote:
xeos wrote:For those of you who would like to measure lag but find the price of the timeslueth too high, how about free instead? I have a project that uses a raspberry pi zero to achieve the same thing entirely using software, no soldering required. If you don't already have a raspberry pi the zero model goes for about $5USD.

you can read about it here.

https://alantechreview.blogspot.com/202 ... berry.html
Nice. Not as user friendly as my Time Sleuth, though.

Watch out for rolling shutter. :) Ideally, users will want to mount the pi so it partially blocks the display to photograph the bars and led simultaneously.

Probably better to announce it in its own thread, though.
For sure my method is more work - it's a question of whether you want to spend $80-100 OR $5-10 and your time. Though, there are some advantages to my approach - since the camera handles exposure you can operate over a wider range of display intensities. Some folks were having issues with the timeslueth being too(?) sensitive.

Rolling shutter is def. an issue - but the back light flicker of most displays allows you to figure out which way your sensor is read out, so you just make sure the pi and the target probe are on the same readout line. I discuss this issue on the "how to use" page for the project.

FYI I don't think the pi can do 240p, but it can output 480i/480p and on up. It might be able to run at >60hz too, though I haven't the display to test that out properly.
User avatar
citrus3000psi
Posts: 668
Joined: Wed Dec 25, 2013 11:56 pm
Location: Indiana

Re: Time Sleuth Lag Tester

Post by citrus3000psi »

There might be a new TS in the works at some point. I found a new hdmi transmitter that can do 4k30 w/ deepcolor.
We also picked up dev kit to test with. The new TS would also have a ADC built in. This would fix the current "issues" with the time sleuth. It would list display Lag, Black to Full white, and Black to Black.
User avatar
xeos
Posts: 167
Joined: Thu Oct 20, 2016 10:38 pm
Location: San Diego, California

Re: Time Sleuth Lag Tester

Post by xeos »

citrus3000psi wrote:There might be a new TS in the works at some point. I found a new hdmi transmitter that can do 4k30 w/ deepcolor.
We also picked up dev kit to test with. The new TS would also have a ADC built in. This would fix the current "issues" with the time sleuth. It would list display Lag, Black to Full white, and Black to Black.
Ah, so the current time slueth uses the same kind of binary threshold that the ossc lag tester mod uses?
User avatar
TrantaLocked_
Posts: 46
Joined: Wed Jun 24, 2020 3:13 pm

Re: Time Sleuth Lag Tester

Post by TrantaLocked_ »

Isn't the Time Sleuth supposed to ship with a firmware that can output different resolutions? My Time Sleuth only outputs 1080p at all knob positions, tested on multiple displays.
User avatar
Fudoh
Posts: 13015
Joined: Mon Mar 06, 2006 3:29 am
Location: Germany
Contact:

Re: Time Sleuth Lag Tester

Post by Fudoh »

I guess you have to ask the seller.

I bought mine from videogameperfection and both came with a 240p-enabled firmware (replacing 1080i) as requested.
User avatar
TrantaLocked_
Posts: 46
Joined: Wed Jun 24, 2020 3:13 pm

Re: Time Sleuth Lag Tester

Post by TrantaLocked_ »

I don't know what I'm doing wrong here. Quartus 18.1 is installed, proper driver is installed for the Altera USB blaster, JTAG cable connected, Time Sleuth powered on, and the programmer can't detect the JTAG chain. I added the FPGA manually just to see, and of course the programming fails. However, the debugger is able to detect "UNKNOWN_NO_JTAG_ID" for devices 1, 2 and 5, "UNKNOWN_FFFFFDFF" for device 3 and "UNKNOWN_FFFFFE55" for device 4.

Update
Wow I'm really surprised by what just happened. I just got the programming to work and I'm pretty sure it was a bad JTAG cable. Mine had good continuity and was still having issues with Quartus, so I pulled off the header plastic on both ends so I could move the cable off the metal crimp pins that pierce through the cable, made new crimp holes a bit down the cable length in case the originals were bad, and that initially didn't work. But then, as I had all devices connected with Quartus programmer open, I kept attempting to flash while slightly moving the cable up and down the exposed crimp pins to see if I could get a better connection, and eventually one of my movements did it somehow. This JTAG cable came with one of those really cheap knockoff Altera USB Blasters.

When it comes to detecting the JTAG device chain in Quartus Programmer, a start would be if the debugger at least shows some unknown devices. Better would be if auto-detect actually works and shows you a 10M02SC (even better 10M02SCE144 which in the hardware device list is closest to the actual FPGA in the Time Sleuth). I think it's supposed to show just 10M02SC, but I don't really know. What do you guys see when you auto-detect? Anyway, if there's anything else someone wants to know about what I did I will be happy to help.
User avatar
BuckoA51
Posts: 3357
Joined: Sat Oct 02, 2010 10:08 am
Location: Ireland
Contact:

Re: Time Sleuth Lag Tester

Post by BuckoA51 »

I had a lot of issues and weird intermittent errors programming a TS using a cheapy knockoff JTAG programmer, but it always works for me using a genuine USB blaster.
OSSC Forums - http://www.videogameperfection.com/forums
Please check the Wiki before posting about Morph, OSSC, XRGB Mini or XRGB3 - http://junkerhq.net/xrgb/index.php/Main_Page
User avatar
TrantaLocked_
Posts: 46
Joined: Wed Jun 24, 2020 3:13 pm

Re: Time Sleuth Lag Tester

Post by TrantaLocked_ »

I really don't want to spend the $200+ for a real one. Unless I can get an authentic one for cheap?

Also, was the "Time Sleuth" in the upper right corner removed in recent firmwares? (edit: Yes, it was a necessary removal for the custom firmware creator due to a storage limitation). And I've been having an issue with weird rolling lag cycles on my 2014 Sharp HDTV. The 1080p cycle is 22ms - 38ms increasing over a 10 second period. 960p, 720p, etc, usually start at a seemingly random number and very slowly decrease (1ms per 2 or 3 minutes) over a long time period, for something like a 20-30 minute cycle. This doesn't happen to the same degree on my Samsung 1080p monitor over DVI in, which is either stable at 2.5ms at 100% brightness, or for any lower brightness cycles between 2.5ms and 3ms over a 2 second period (at all Time Sleuth resolutions).
chriz2600
Posts: 21
Joined: Sun Feb 07, 2016 12:23 pm
Location: Cologne, Germany
Contact:

Re: Time Sleuth Lag Tester

Post by chriz2600 »

TrantaLocked_ wrote:I really don't want to spend the $200+ for a real one. Unless I can get an authentic one for cheap?
There are some other option between the really cheap one and the original:

- There is one from Terasic: https://www.terasic.com.tw/cgi-bin/page ... yNo=&No=46
- And a Waveshare model: https://www.waveshare.com/usb-blaster-v2.htm
TrantaLocked_ wrote: And I've been having an issue with weird rolling lag cycles on my 2014 Sharp HDTV. The 1080p cycle is 22ms - 38ms increasing over a 10 second period. 960p, 720p, etc, usually start at a seemingly random number and very slowly decrease (1ms per 2 or 3 minutes) over a long time period, for something like a 20-30 minute cycle. This doesn't happen to the same degree on my Samsung 1080p monitor over DVI in, which is either stable at 2.5ms at 100% brightness, or for any lower brightness cycles between 2.5ms and 3ms over a 2 second period (at all Time Sleuth resolutions).
Does the decrease stop at some point? Is it also showing up in the "average" ("avg")? It's pretty normal to see jumping values on displays with PWM controlled backlight. This was also the reason to add the "average" and "min/max" ("m/m") rows.

This is from the README.md (https://github.com/chriz2600/time-sleut ... /README.md):
  • min/max (m/m)
    Minimum and maximum value within the last averaging period. These values are reset, when a new averaging period starts.

    average (avg)
    Average value of the last 16 measurements. Most LED backlit LCD screens are using pulse width modulation for brightness adjustment and the PWM duty cycle of the backlight is often not 100% even if brightness is. So the current readings are often jumping, so average gives you the mean lag.
    To avoid the flicker pattern to be "in sync" with the PWM frequency, the time between two fields is cyclically changed by one frame up to 8 frames.
User avatar
BuckoA51
Posts: 3357
Joined: Sat Oct 02, 2010 10:08 am
Location: Ireland
Contact:

Re: Time Sleuth Lag Tester

Post by BuckoA51 »

Terasic is the one I use and it's rock solid.
OSSC Forums - http://www.videogameperfection.com/forums
Please check the Wiki before posting about Morph, OSSC, XRGB Mini or XRGB3 - http://junkerhq.net/xrgb/index.php/Main_Page
User avatar
orange808
Posts: 3196
Joined: Sat Aug 20, 2016 5:43 am

Re: Time Sleuth Lag Tester

Post by orange808 »

BuckoA51 wrote:Terasic is the one I use and it's rock solid.
+1 on the Terasic.
We apologise for the inconvenience
User avatar
TrantaLocked_
Posts: 46
Joined: Wed Jun 24, 2020 3:13 pm

Re: Time Sleuth Lag Tester

Post by TrantaLocked_ »

Ok I have testing results:

Test setup:
-Time Sleuth 1.2a direct over HDMI
-Sharp 1080p HDTV (Sharp LC-39LE551U)
-Time Sleuth firmware presets #1-#5: 1080p/960p/720p/480p/480i

Notes:
-No difference between different HDMI ports
-No difference between HDMI cables
-No difference between power source (TV USB or wall adapter USB)
-All measurements for this test were taken from the first (top) field
-In all instances, the decrease or increase in lag is linear over a time period. For example, in a lag decreasing state, the time between 29.55ms and 29.50ms would be the same as between 29.50ms and 29.45ms.
-TV backlight is at 65%, other settings at "stock" with post processing off and game mode on. Most of these values have little affect on lag, but brightness and contrast have an effect.

Test Results:

1080p
-The 1080p lag cycle always measures consistently between 21ms-38ms over a ~10-second period. The cycle starts at 21ms, increases linearly over this time period to 38ms, then restarts back at 21ms.

Other resolutions
-The starting lag measurements for the other resolutions tend to differ depending on which knob position I'm approaching from and how long I have had the TV turned on, or how long I've stayed on non-1080p resolutions while the Time Sleuth and TV are on. Switching resolutions many times seems to bring up the starting value for non-1080p resolutions. Resetting the TV also brings starting times up, usually between 30ms and 35ms for non-1080p resolutions.
-Most of the time, the lag will either increase or decrease over a long period of time until it settles around 20ms-23ms.
-In essence, after first initially turning on the TV and Time Sleuth, the non-1080p resolutions tend to start at a value above 30ms and either go up or down over the course of an hour and eventually settle at a value between 20ms and 23ms. 480i is a bit of an exception in that it tends to be higher and can settle in the high 30s.
-Once one resolution has been allowed to settle, it usually means that when I switch to another non-1080p resolution, the value for that next resolution will already be close-ish to the 22ms mark, however the more I switch resolutions the higher the starting value becomes.
-Turning off the TV completely resets the starting value for a non-1080p resolution, so even if I wait for 720p to settle at 22ms, if I restart the TV the starting value goes back to 30ms-35ms.

720p
-My full test run for 720p had a starting lag value of 30.56ms.
-The test ended after about 50 minutes when the lag value finally settled at 22.56ms and appeared it would stay there indefinitely.
-After reaching a stable 22.56ms value, I switched the knob to 960p and to my surprise the starting value for 960p was 22.38ms, stable.
-After that, I switched back to 720p and the starting point became 24.91ms and it was again decreasing slowly from this value.
-I then switched to 480p and the starting value was 23.20ms, again slowly decreasing.
-After all of this, 1080p still cycled between 22ms-38ms.

480p
-There was an instance where the starting value for 480p was 34.99ms, and it slowly increased for about 5 minutes, passing 38ms and then cycling back to settle at a stable 20.79ms.

480i
-I had one instance of 480i starting at 50.90ms and slowly increasing. I did not wait to see what it would stop at, if at all.
-I had another instance of 480i starting at and being stable at 39.30ms

Conclusion:
I have no idea.
Last edited by TrantaLocked_ on Mon Jul 20, 2020 9:36 am, edited 1 time in total.
User avatar
orange808
Posts: 3196
Joined: Sat Aug 20, 2016 5:43 am

Re: Time Sleuth Lag Tester

Post by orange808 »

An electrical engineer posted some conclusions earlier in the thread--in direct reply to your experience. :)
We apologise for the inconvenience
eatnumber1
Posts: 13
Joined: Fri May 11, 2018 5:13 pm

Re: Time Sleuth Lag Tester

Post by eatnumber1 »

Hey folks, just got a Time Sleuth, running into a different issue that hopefully someone can advise on.

It appears that each time I plug in the Time Sleuth (power or HDMI), the lag changes, but remains very stable until I unplug & re-plug a cable. I'm getting measurements in the range 43ms - 59ms, and as I mentioned, each lag measurement is extremely stable until I restart the Time Sleuth or unplug / plug back in the HDMI cable.

I'm measuring my LG 55EF9500 TV. Each measurement I take is against the top of the screen. Each time I do it, I let it run for a while, and every measurement value (cur, m/m, and avg) all settle out to the exact same value.

Here's the measurements, taken while unplugging the power cable to the TS:
48.05ms
56.57ms
43.85ms
50.49ms
55.37ms

And here's more measurements, this time unplugging the HDMI cable:
47.44ms
50.00ms
54.39ms
59.75ms
54.74ms
56.54ms

Does anyone know what's going on here? Why am I getting such varying lag times?
User avatar
maxtherabbit
Posts: 1763
Joined: Mon Mar 05, 2018 4:03 pm

Re: Time Sleuth Lag Tester

Post by maxtherabbit »

it's called variable latency, totally normal for some tv processors
eatnumber1
Posts: 13
Joined: Fri May 11, 2018 5:13 pm

Re: Time Sleuth Lag Tester

Post by eatnumber1 »

Can you elaborate? I can’t seem to find any information on variable latency for TVs other than the variable refresh rate feature that my TV doesn’t have.

Furthermore, if the latency was going to vary, why does it “lock in” on each connection and remain stable?

Also other folks have done lag testing on this TV and don’t report anything weird with it. For instance, this website reports an input lag of 53.5ms for 1080p60. On this page they talk about the tool they use, and do discuss variation in measurement due to what is essentially tearing when measuring with vsync turned off, but that seems like it should then vary over time, not just each time I connect it.
User avatar
orange808
Posts: 3196
Joined: Sat Aug 20, 2016 5:43 am

Re: Time Sleuth Lag Tester

Post by orange808 »

Different issue? Are you sure? Sounds like the PWM again to me.
We apologise for the inconvenience
eatnumber1
Posts: 13
Joined: Fri May 11, 2018 5:13 pm

Re: Time Sleuth Lag Tester

Post by eatnumber1 »

If I understand the PWM issue correctly, cur should vary over time. Right? That’s not happening here. Every metric: cur, average, min/max are all the same after a few seconds.
User avatar
orange808
Posts: 3196
Joined: Sat Aug 20, 2016 5:43 am

Re: Time Sleuth Lag Tester

Post by orange808 »

eatnumber1 wrote:If I understand the PWM issue correctly, cur should vary over time. Right? That’s not happening here. Every metric: cur, average, min/max are all the same after a few seconds.
Why would flicker always vary? There's no regulations regarding cadence.

Also, you have an average reading that's fine. (By "fine", I mean the Time Sleuth gave you an estimate of the latency. The actual display lag of that particular LG panel isn't fine; three frames of lag is a bad score.)
We apologise for the inconvenience
eatnumber1
Posts: 13
Joined: Fri May 11, 2018 5:13 pm

Re: Time Sleuth Lag Tester

Post by eatnumber1 »

orange808 wrote:
eatnumber1 wrote:If I understand the PWM issue correctly, cur should vary over time. Right? That’s not happening here. Every metric: cur, average, min/max are all the same after a few seconds.
Why would flicker always vary? There's no regulations regarding cadence.

Also, you have an average reading that's fine. (By "fine", I mean the Time Sleuth gave you an estimate of the latency. The actual display lag of that particular LG panel isn't fine; three frames of lag is a bad score.)
Okay, saying this differently. My TV is an OLED display. It doesn't have a backlight, so can't be prone to the problem as it's been described.

Thinking this through more carefully though, if we pretend my TV did have a backlight, if the backlight is flickering at frequency fd, where fd > the refresh rate, and the TS samples at its own frequency ft, where ft also > the refresh rate, if fd is an even multiple of ft, then depending on the phase of ft relative to fd, ft will latch into sampling fd at one particular point in its duty cycle. This seems like it would produce a consistent latency measurement which changes each time the fd or ft clocks are restarted, but between runs the behavior would only vary within a full clock period of ft (assuming it's greater than fd), meaning that the latency test may show very small variance between runs, but not anywhere near as large as a full 16ms of variance, which is what I'm seeing. If a full 16ms of variance were possible due to this explanation, it would mean that the backlight only flickers once every frame.

If instead, as what seems much more likely, fd is not an even multiple of ft, the phase difference should shift over time, which I think should produce varying latency measurements without restarting either clock.

Also FWIW, I want to get a useful measurement at all here because I want to then proceed to test the latency of various components in my HDMI chain, which I can't do until I've got a base latency measurement for the TV itself that I'm confident in.
User avatar
TrantaLocked_
Posts: 46
Joined: Wed Jun 24, 2020 3:13 pm

Re: Time Sleuth Lag Tester

Post by TrantaLocked_ »

Eatnumber1, I think I might be on to something too, with very similar results to your's although my Sharp TV does use PWM and the frequency is probably 120Hz or 240Hz based on reviews of similar TV models (1)(2). In my case the starting value isn't 100% stable like your's, but rather gravitates down at about .01ms / s until hitting a value near 22ms and staying there. I do think it's possible that it's due to refresh rate sync and not PWM in both our cases, with your 120Hz TV naturally having a 8ms range while my 60Hz TV having a 16ms range. I think that it's possible due to what the sync state is when the Time Sleuth pairs with the TV. At time of pairing, if the TS sends the output just after or just before the TV's most recent frame finished drawing (plus base input lag), or somewhere in between, that determines what the remaining time the TS output needs to wait for the current frame to finish drawing for the rest of the session, until power cycle for either TS or the TV. The minimum lag of my Sharp TV is 22ms, so to see a 38ms starting lag value, the first TS output would need to be sent just after ~10ms into the TV's second to last frame drawing before paring, and just before ~10ms for a 22ms starting value.

I've been power cycling the Time Sleuth and the TV in different combinations and have noticed what the starting values tend to be. The range of starting values for 720p is 22ms-38ms, and I believe the starting value is random (as it naturally should be, but certain conditions can lead to a higher likelihood of a particular starting value). In my case, the difference between 720p and 1080p is that for 1080p I see the full cycle play out within 10 seconds, but the other resolutions give me a more stable (but still changing) starting number. Maybe just the TV's PWM being different depending on input resolution. Could chriz2600 help here?
Last edited by TrantaLocked_ on Mon Jul 20, 2020 12:18 pm, edited 2 times in total.
User avatar
Kez
Posts: 818
Joined: Thu Jul 20, 2017 7:09 am

Re: Time Sleuth Lag Tester

Post by Kez »

eatnumber1 wrote:Also FWIW, I want to get a useful measurement at all here because I want to then proceed to test the latency of various components in my HDMI chain, which I can't do until I've got a base latency measurement for the TV itself that I'm confident in.
You'll just have to increase your sample size. Take 10 or 20 measurements of the TV alone, then add your device and take a bunch of measurements again. It should be pretty obvious how much latency is being added to the chain. The larger the sample the more accurate your results.
sofakng
Posts: 218
Joined: Fri Sep 02, 2016 1:30 am

Re: Time Sleuth Lag Tester

Post by sofakng »

Does anybody have a bare PCB they would be willing to sell me?

I'm looking for the newer version with the 5-position switch.

I just ordered and built a Time Sleuth but didn't realize a newer PCB with the 5-position switch was available.
User avatar
xeos
Posts: 167
Joined: Thu Oct 20, 2016 10:38 pm
Location: San Diego, California

Re: Time Sleuth Lag Tester

Post by xeos »

citrus3000psi wrote:There might be a new TS in the works at some point. I found a new hdmi transmitter that can do 4k30 w/ deepcolor.
We also picked up dev kit to test with. The new TS would also have a ADC built in. This would fix the current "issues" with the time sleuth. It would list display Lag, Black to Full white, and Black to Black.
FYI, a little friendly competition: I've made a raspberry pi based lag tester that uses and ADC so that it can provide that kind of stuff (b2w, not yet b2b or g2g). Not to discourage you from making your new device. My real goal is get reliable input lag and response time numbers more generally available. My device isn't stand-alone like the time slueth and requires a bit of command line knowledge so it won't replace the time slueth, just offer a cheaper alternative for somewhat more technically knowledgeable folks.

https://alantechreview.blogspot.com/202 ... -page.html
Post Reply