Page 2 of 4 FirstFirst 1234 LastLast
Results 21 to 40 of 65

Thread: using the transient (dy/dx) filters

  1. #21
    Senior Tuner Frost's Avatar
    Join Date
    Jan 2006
    Location
    Richmond VA
    Posts
    1,950
    I'm waiting up tonight for the answer
    Steve Williams
    TunedbyFrost.com


  2. #22
    Advanced Tuner
    Join Date
    Dec 2005
    Posts
    214
    Quote Originally Posted by Frost View Post
    I'm waiting up tonight for the answer
    Me to!!

  3. #23
    Senior Tuner LSxpwrdZ's Avatar
    Join Date
    Dec 2008
    Location
    on the Dyno
    Posts
    1,825
    I've not had the chance to tinker with this yet. Hopefully I can help test and give some insight on it.
    James Short - [email protected]
    Located in Central Kentucky
    ShorTuning
    2020 Camaro 2SS | BTR 230 | GPI CNC Heads | MSD Intake | Rotofab | 2" LT's | Flex Fuel | 638rwhp / 540rwtq
    2002 Camaro | LSX 427 | CID LS7's | Twin GT5088's | Haltech Nexus R5 | RPM TH400

  4. #24
    Senior Tuner Frost's Avatar
    Join Date
    Jan 2006
    Location
    Richmond VA
    Posts
    1,950
    ahhhhh I need to know the answer to this!!
    Steve Williams
    TunedbyFrost.com


  5. #25
    Senior Tuner
    Join Date
    Dec 2004
    Location
    Laurel, MD
    Posts
    1,020
    do we get an option for how many frames/seconds our 'window of change' is?

  6. #26
    Senior Tuner S2H's Avatar
    Join Date
    Apr 2005
    Location
    Schexnayder Racing - Arnaudville LA
    Posts
    4,387
    per a talk we had in the moderator section...
    should be dx/dt

    use could be something as simple as [send.90.dx/dt]
    it reports the difference in TPS +/- from the last frame..
    currently no option to change the amount of frames it uses...
    -Scott -

  7. #27
    Senior Tuner Frost's Avatar
    Join Date
    Jan 2006
    Location
    Richmond VA
    Posts
    1,950
    abs([PID.17.dy/dx])
    or
    [PID.17.dy/dx]

    Returns actual TPS with two decimals. This is not what is expected.

    edit

    you mean literally that dx/dt need to be used...
    Steve Williams
    TunedbyFrost.com


  8. #28
    Senior Tuner Frost's Avatar
    Join Date
    Jan 2006
    Location
    Richmond VA
    Posts
    1,950
    No, using dx/dt returns the same (actual value) as dy/dx with or without (obviously but just mentioning) ABS.
    Steve Williams
    TunedbyFrost.com


  9. #29
    Senior Tuner DSteck's Avatar
    Join Date
    Mar 2008
    Location
    St. Louis, MO
    Posts
    4,856
    Quote Originally Posted by soundengineer View Post
    per a talk we had in the moderator section...
    should be dx/dt

    use could be something as simple as [send.90.dx/dt]
    it reports the difference in TPS +/- from the last frame..
    currently no option to change the amount of frames it uses...
    The change since the last frame would be ideal. It'd be the easiest for people to understand, too.

    DSX Tuning - Authorized HP Tuners Dealer
    http://www.dsxtuning.com
    http://www.facebook.com/dsx.tuning
    Just say no to bull s***.
    IF YOU WANT HELP, POST A FILE!

  10. #30
    Senior Tuner
    Join Date
    Dec 2004
    Location
    Laurel, MD
    Posts
    1,020
    ideal is not the word you're looking for, more like 'inflexible.' some things change quickly, others change slowly, so you have to regulate the duration of the change, cant just rely on fixed and arbitrary intervals.

  11. #31
    Senior Tuner DSteck's Avatar
    Join Date
    Mar 2008
    Location
    St. Louis, MO
    Posts
    4,856
    I'll rephrase... for what I want to do, the change since the last frame would be ideal.

    DSX Tuning - Authorized HP Tuners Dealer
    http://www.dsxtuning.com
    http://www.facebook.com/dsx.tuning
    Just say no to bull s***.
    IF YOU WANT HELP, POST A FILE!

  12. #32
    Quote Originally Posted by Frost View Post
    abs([PID.17.dy/dx])
    or
    [PID.17.dy/dx]

    Returns actual TPS with two decimals. This is not what is expected.

    edit

    you mean literally that dx/dt need to be used...
    i need to check with Keith, sounds like something is broken. you need to use dy/dx when you spec the formula. I have discussed with Keith changing it to read dx/dt in future as this is what it *should* be.
    I count sheep in hex...

  13. #33
    Senior Tuner DSteck's Avatar
    Join Date
    Mar 2008
    Location
    St. Louis, MO
    Posts
    4,856
    Quote Originally Posted by Chris@HPTuners View Post
    i need to check with Keith, sounds like something is broken. you need to use dy/dx when you spec the formula. I have discussed with Keith changing it to read dx/dt in future as this is what it *should* be.
    So it's definitely change in the parameter divided by the time since the last frame... at least, that's how it's supposed to be?

    DSX Tuning - Authorized HP Tuners Dealer
    http://www.dsxtuning.com
    http://www.facebook.com/dsx.tuning
    Just say no to bull s***.
    IF YOU WANT HELP, POST A FILE!

  14. #34
    Senior Tuner
    Join Date
    Dec 2004
    Location
    Laurel, MD
    Posts
    1,020
    so with the syntax available, how would I express 'change of change?' Laplacian of Gaussian is a common noise reduction technique, which would be hugely useful to us.

  15. #35
    Senior Tuner Frost's Avatar
    Join Date
    Jan 2006
    Location
    Richmond VA
    Posts
    1,950
    Quote Originally Posted by redhardsupra View Post
    so with the syntax available, how would I express 'change of change?' Laplacian of Gaussian is a common noise reduction technique, which would be hugely useful to us.

    I have used LoG playing with images in the now far past...

    I guess you could define the first portion as a transient user PID and make a second user transient PID which examines the first..
    Steve Williams
    TunedbyFrost.com


  16. #36
    i discussed this further internally and i think the best way to go with it in future will be to simply have a reference to the previous frame value of any PID and also the time between frames eg. [PID.17.old] and [PID.17.time] or something similar.

    This way you can implement derivative functions, simple lag filters using any coefficient you want and any other formula you can think of without needing us to code it in.

    eg.

    FastFilteredValue = PID.17.old + (0.9 * [PID.17 - PID.17.old])

    SlowFilteredValue = PID.17.old + (0.1 * [PID.17 - PID.17.old])

    DerivativeValue = (PID.17 - PID.17.old) / PID.17.time
    I count sheep in hex...

  17. #37
    Senior Tuner Frost's Avatar
    Join Date
    Jan 2006
    Location
    Richmond VA
    Posts
    1,950
    Is there any other information on this feature?
    Steve Williams
    TunedbyFrost.com


  18. #38
    Senior Tuner S2H's Avatar
    Join Date
    Apr 2005
    Location
    Schexnayder Racing - Arnaudville LA
    Posts
    4,387
    Quote Originally Posted by Chris@HPTuners View Post
    i discussed this further internally and i think the best way to go with it in future will be to simply have a reference to the previous frame value of any PID and also the time between frames eg. [PID.17.old] and [PID.17.time] or something similar.

    This way you can implement derivative functions, simple lag filters using any coefficient you want and any other formula you can think of without needing us to code it in.

    eg.

    FastFilteredValue = PID.17.old + (0.9 * [PID.17 - PID.17.old])

    SlowFilteredValue = PID.17.old + (0.1 * [PID.17 - PID.17.old])

    DerivativeValue = (PID.17 - PID.17.old) / PID.17.time

    I like the idea to code it as a variable that you can change...
    perhaps a setting for dt somewhere in the main options section of the scanner

    dt= ___ Frames
    then it could be used as pid.17.dx/dt and it would automatically return a value based on the difference from the dt value set in the options
    as an example dt=5 would return the change in value over 5 frames

    we could also use a way to average things..so you could get the average change in real time vs plotting the average in the scanner...
    sometimes it would be better to get a constantly updated average over a given period of time vs the total average for the entire log.

    Maybe an Average over time setting? AOT=600 would return the average over 600 frames...continuously updating...dropping early values ad keeping the latest values
    -Scott -

  19. #39
    Senior Tuner DSteck's Avatar
    Join Date
    Mar 2008
    Location
    St. Louis, MO
    Posts
    4,856
    Quote Originally Posted by soundengineer View Post
    we could also use a way to average things..so you could get the average change in real time vs plotting the average in the scanner...
    sometimes it would be better to get a constantly updated average over a given period of time vs the total average for the entire log.

    Maybe an Average over time setting? AOT=600 would return the average over 600 frames...continuously updating...dropping early values ad keeping the latest values
    This would be awesome. I've often tried to make this possible with what's available, but if it was a simple feature, it'd kick ass.

    DSX Tuning - Authorized HP Tuners Dealer
    http://www.dsxtuning.com
    http://www.facebook.com/dsx.tuning
    Just say no to bull s***.
    IF YOU WANT HELP, POST A FILE!

  20. #40
    Senior Tuner mowton's Avatar
    Join Date
    Mar 2007
    Location
    Florida
    Posts
    3,773
    Are there any updates on how to use this feature? .

    Thanks,

    Ed M
    Last edited by mowton; 11-24-2010 at 12:09 PM.
    2004 Vette Coupe, LS2, MN6, Vararam, ARH/CATs, Ti's, 4:10, Trickflow 215, 30# SVO, Vette Doctors Cam, Fast 90/90, DD McLeod, DTE Brace, Hurst shifter, Bilsteins etc. 480/430

    ERM Performance Tuning -- Interactive Learning ..from tuning software training to custom tunes
    HP Tuners Dealer- VCM Suite (free 2hr training session with purchase), credits and new Version 2.0 turtorial available
    http://www.ermperformancetuning.com
    http://www.facebook.com/ERMPerformanceTuning

    [email protected]