GRITTY 1.2 essai iw1ayd DRAFT V 004#2.pages
Contents
1. any group of bits that make a character or a symbol of the Baudot International Telegraph Alphabet No 2 ITA2 we use Leave out the symbol definition for data communication by now But those decoder PAGE 2 OF 7 DRAFT V 004 doesn t work out the of very simple to very so phisticated bit recognition technique Usually we look at the printing cursors more than one decoder is printing at a time and de cide what is the real strings that we need catching it in the way as we prefer or we could If there are fragments we have to use our hands to write down what s the results whatever we think it is Then we have other ways when the relevant data is already inside the field we de sire to valorize Let s say we are looking for a call sign one decoder print RTVWIUP the other print FSA2UR We in this case already fo cused well into all that is on the screen and separated those two fragment of strings focus ing a sliding windows Than we correlate these strings from different sources and decide for VA2UP Consider that we have A2 and UP frag ments in the received strings Place that ex tracted callsigns into the proper filed and look for the checkout result made by screening VA2UP string into databases of known callsign It is there and valid Some of us perhaps the more experienced fastest or all together check also that the time of the day and the aiming of aerials to have more clue By example a 18 00 UTC on 40m with a2. low dipole it s much easy that the call we try to recognize is RA2ZUP and not VA2UP Than a check with known call DB and discover that RA2UP is not a valid well known call Now we have at least other two ways to get a clue but those are out of the scope by now What we have done Correlation Correlations are useful because they can indicate a predic tive relationship that can be exploited in prac tice We placed together scavenged different things that have nothing of clever Translating garbage to mangled data and extracting others the more relevant for us data We have applied our knowledge of calls and others skills we have to help ourselves getting out informations from a stream of useless characters we had at the start All this in a very short time hopefully With callsigns it s pretty easy It s not that easy with exchanges do you have never seen that NR NR PSE NR ONLY A very frenetical re quest Yes we are able also to correlate sets of characters that contain numbers But with INTEGRATING WITH NIMM GRITTY 1 2 IW1AYD those we don t have any clue or any further skill to apply when in needs 599 0030EQ13P out to make shorter and clever that retry request and compare again with more data A data base of common exchange doesn t have any common sense as numbers are a volatile part off the ex change Still we should apply easy rules We see 1 one time may it s a 001 Then it depends from u
3. solved BTW GRITTY is 87 aimed to contester not ca sual user It s code is designed to give the most valuable functions when contesting General note GRITTY have a new technology and aim to make it deployable at user level into our machines It is developed around it s parent application Making it resembling or perhaps only reusing some aspects But it s aimed to a new group of potential users Some parts of the new tech nology are more prominent maybe as well known by some Others living just under the hood of its brand new approach to ours world It is well mixed and somewhat promising At tempting to outreach the pure filters and bits correlations limits with data and the meanings of those data correlations That s a new ap proach leaving old fields and coming inside new ones It represent a great set of changes and could improve a significant slice of our hobby RTTY Contesting Nonetheless it is a new one and our habits are much well oriented elsewhere By now we use INTEGRATING WITH NIMM GRITTY 1 2 IW1AYD as real time running decoders two other appli cations MMTTY and 2Tone Somebody of us also trust well known pieces of HW also with some FW well hidden inside Spacing with those last from pure decoding to don t read simply Lissajous waveforms check Every time we face a new decoder we also face ideas on how to check those Comparing older and newest as they could be a full soluti
4. 4 WAE 0 1 Exchange1 599 nnn lt nnn gt 0 9 Exchange3 5NN J nnn lt nnn gt 0 1 Some vocabulary to interpret this is inside the Con tests ini also for Patt3Ch Ist that collaborate for call INTEGRATING WITH NIMM GRITTY 1 2 IW1AYD sign structures checkout Also remember the RTTY Skimmer manual Then apply statistical analysis using the pattern recognition as a sliding window over the test incoming and already received As a result it could post fetch and correct data already re ceived as garbage applying such correlation schema Almost as any of us after having some experience could do What about the differ ences in between us and GRITTY 1 GRITTY even if it is sweet have only a pre pared set of elements to be correlated 2 GRITTY doesn t become tired or lazy it couldn t each correlation schema it has into will be applied from 00 00 UTC until the end of the contest the same way Point number 1 says also that GRITTY have a finite but modifiable set of patterns We have to make choices about witch contest will be in use We have to trust the patterns that are already defined for that contest Or we have to find ourselves new pattern s for new contest not already into is table of recognition Point 2 wow But still we still need to stay bot tom in the chair believe it There are already several contests patterns include obviously into a file used by GRITTY The short CHM help say Contest Ini that reside by i
5. DRAFT V 004 GRITTY 1 2 iw layd Salvatore Salvo Irato 18 Apr 2015 GRITTY 1 2 Integrating with NIMM gene ral usage report Why this paper We are gaining new decoding tools and we must thanks who is working so hard to make all us able to run it There was a launch of GRITTY 1 lt 2 and then the announce of the latest now GRITTY V_ 1 2 This latest message had given also a very interesting Demo page link with audio utility pointers wav recorded files and more stuffs on how to test it off line Some more informations are into this docu ment on the same site RTTY Skimmer Server V 1 0 User manual PDF It relay deeply descriptions of the RTTY Skim mer Server that are part of GRITTY itself As in herited by this last from the big brother I E the Contest ini format and tokens hereafter just summarily presented and shown more than described All this is great also considering that it is a pure new technology for us RTTYers GRITTY will be further integrated into NIMM as a Haddon RX only decoder What is better It received some critics on the same email list and the writer responded promptly Well done not cherishing anyone This is an attempt to connect what s new and highly needed with today aspects all the aspects from roots of what we use This reason let me use such Italian heavy metrics to drill down onto showed aiming of GRITTY e the way as it look It s my personal view fell free to co
6. GE 3 OF 7 DRAFT V 004 GRITTY aiming GRITTY have not only those aspects here por trayed It has much more It attempt to logically find and discover from late symbols what was the previous set of symbols and thus the right word since it was transmitted and already gone It try to receive symbols recognize patterns and back writing what was already received Thus it rewrite a word that where mistakenly printed actual signal decoding capabilities with a word well recognized by its vocabulary of patterns To make this it use a description of exchange for any known contest and a list of call signs That s out of any alchemy already done with DSP filters and bit correlation It is the first as known by me attempt to post fetch data for our world Last but not least it is really the core engine of the pattern recognition GRITTY apply statistical concepts to make its choice about words or string that it think we must read in stead of the mangled ones There is a page into the CHM help file for MS Windows that ex plain very quickly but effectively it all It doesn t relay only on whatever we know was under the hood of our machine A lot of care must be exercised looking at those new fea tures But at least what we need is every time the same IW1AYD 599 384 384 IW1AYD GRITTY search into received data with patterns vocabulary like this IW1 IW1 And this WAE CQ1 CQ 0 3 CQ2 TEST 0 3 CQ3 QRZ 0 3 CQ
7. Microham devices DKII MKII MK2R Keyboard s and mouse stuffs via USB KVM Further considerations and close Going a little bit out of the main topic this is a view that could be shared isn t What should be really needed as MMVARI is making it a lot but could be made better is a multi stream RTTY decoder into the 3 KHz of usual not SDR band pass But not alone a IW1AYD s backgrounds GRITTY 1 2 IW1AYD things like this need some more integration may be also from the logger side A pure dream is a DXEpedition machine a SDR connected to the MF output of a traditional FSK transceiver Leaving to the operator of that time the abili ties to receive a relevant slice of the pile up AGC and sub filtering well achieved controlling with no delays the whole train having in view the most of the generated pile up Lags in be tween the subsidiary not that much SDR re ceiver and the transceiver should be the lowest possible i e not RS 232 replication black holes But this is just a dream as several other per sonally have Maybe GRITTY is not the one but his big parent the RTTY skimmer is to be used mod for All this is not needed for RTTY contesting What we would need is measured often into a 300 500 Hz bandwidth and thus the ability to catch on signals slightly out of tracks 73 de iwlayd Salvo screen real estate is a scarce if not lacking resource usually Well defined placement of windows
8. cking i e a prefer ence menu This seems an early stage of the whole thing there isn t a central preference menu where to get forth and back clicking buttons so to en force or not rules or where to enter text as to use new checking schemas From one side this could mean no user setup out of hands on operations on files From the other side it means that those operations aren t necessarily to be done It s hard to say on V 1 2 Speculating on the subject could give dif ferent answers but it will be not useful by now Logs Some attempts to use simple strings com mands as explained by the help chm file though the TCP port went on successfully but without any effectiveness Effectiveness is mea sured having on hands the old but still unique and effective MMTTY log file Out of this it seems that in a slightly real time ambient like SO2R or Mx contesting using the network for a simple task like this should be considered heavy As usual not by only itself but considering all the other dressing we have to use and what s typically under the hood of our PC So we should wait until some good soul will come in help of us making anything of useful to write a file with incoming data hope that any body will recognize the importance to have an automated and clever log output That s not only for cheating purposes it seems that PAGE 5 OF 7 DRAFT V 004 everybody write software is preventing logging as a cheating
9. code environment as to the carer for that By the way native installation of NIMM aren t also portable from one PC to another That s it just taking care to fully install It on every PC then making on one PC the whole setup lastly move one single directory in one single shot onto the second or more PC But this lack is easily overcome just taking care of some basic rules So definitively NIMM will all the setups PAGE 1 OF 7 DRAFT V 004 and the needed basic now those are to be considered basic decoders and ancillary appli cations files could be moved around the whole group of operating posts GRITTY is still not able to write and correct data into NIMM DI window Yes it fetch back cor rected data for know and configured described tokens as callsigns and exchanges But it seems that the ability to have a back and forward ca pability to write into a window owned by an other process is not reachable now MS devel opment kit shortcomings It could be nobody since now do that in short burst backing the Cursor and rewriting one or more characters May be it is not a provisioned callable routine from inside the MS development kit of the sea son GRITTY is not plain user configurable at least for the GUI GRITTY still is a very new and promising cake it is young and some of those issues and maybe other that didn t seen or could imagine may give use a very productive weapon for our con testing arsenal when
10. countermeasure A log a com plete log with TX and RX timestamps is needed also for DXing and on any mode for QSLing often ended out a fragment of MMTTY log to say sorry you are not in the log as received this One or more times confirmed with LOTW QSO not in my log for the contest but there callsigns 599 only for the QSO itself Having a log is not cheating Thinking that a log is not useful or it must be a hard wired operation to be done by hands is not a good practice Real life unravel testing It was not a goal to test it that deeply Nor to simulate it on the run It will be a task to be ac complished in the future Overall impression From the interesting reading about the subject and testing it on the fly during some QSO BTW low tones it doesn t seem to be used first line now It is something new well working more promising and stable some hours It is made to be a great tool Summary of issues Not all but what was seen until now 1 Integration and portability NIMM could be rendered truly portable at least by the file sys tem standpoint GRITTY will not follow Nothing that could not be solved by some scripting and checks Still having more thing to be done but not to be done the last five minutes before 00 00 UTC or in the middle of a creek KIS 1 a metering is how much are portable MMT TY and 2Tone They don t need not even force for to be sparsely installed into default Win dows
11. directories But could reside read para meters and write logs into user defined places It s easy to tie all those together into well known places 2 N1IMM integration is coming It is only half done right now Clicking a call will place it into the main input window But incoming data aren t showed into the DI place that result not INTEGRATING WITH NIMM GRITTY 1 2 IW1AYD useful by now That is just in case anyone use it BTW judging on the rate and type of incoming symbols GRITTY will need much more printable space onto its sub DI printable window Those data are literally flying into the GRITTY window As a result the incoming string to be interpolat ed or not are prone to flow out of the readable lines to easy 2 a metering is how other decoders work Changes are welcomed but they should have measurable advantages over previous things Still changes have to be aimed to several other factor as i e KIS 3 Preferences and control over the ini file s is to be done with some help by hands Not a lack by itself but a pitfall together with the main issue Places to be reached are sparse into you must remember sub directories names into the standard for Microsoft trees Further to this there is not a GRITTY directory in view where it live as to search it by eyes It is placed down into a AFREET directory you must re member before 4 Logs aren t just a piece of anyone other cake QSL data verif
12. is not optional Doing some PSK JT modes on HF since 2005 Licensed since 1983 Pursuing RTTY contesting as other life aspects permit since 2008 Most of these the contests where done from the IQ1RY Club Station both as part of a team or as a single operator Since 2008 the switch over RTTY was driving also the switch to NIMM Working into IT over several different scenarios and matters all over the years From hardware integration DC care Security FCO amp ECO to OS amp Application prototyping and pursuing deployments On each matter working roles went from Field Service Engineer Tech nical Support for Customers Product Specialist amp Support to Consulting Architect Trainer Nerd and Geek Some thoughts KIS is the dogma KISS is not the same Sustainable maintainability is a must RTFM is a dogma too often forgotten Open Source software need more than windows click ing All the document was produced with Pages INTEGRATING WITH NIMM PAGE 7 OF 7
13. mment and ask to me direc tly INTEGRATING WITH NIMM GRITTY 1 2 IW1AYD It useems a bit too early to write about GRIT TY It is at the first version s DOC is just a short help file However the programmer s inten tions seems defined Thus as an avid RTTY con tester even if with a short experience that doesn t cover the last 10 years and a profes sional IT roles am spending yours and mine time here Summary GRITTY is an interesting and promising piece of software It promise to give inside all of us PC some of the new technologies already in use from the newest RTTY Skimmer application The one before relied on MMVARI integration This late relies on the lessons learned with CW skimmer Unluckily also the installation and data view schema of GRITTY are prone to it s big parent But GRITTY is aimed for standard users by now it lacks real integration with N1IMM logger and worst it is not made to be portable Better it couldn t be made portable Each installation inside each PC must be copied by hands or scripts This is the worst case for MS to MM setups But also for who use 2 PC with two radio MMITTY and 2Tone are portable NIMM is de facto portable even if not supported for this feature GRITTY is not that Don t be fooled like it s complexity of an application would pre vent portability It s just a choice of the pro grammers to have all the items needed de scribed or to relay onto the
14. nstallation into appdata Afreet Products RttySkimServ Contests ini The call signs are interpreted by Y appdata Afreet Reference MASTER DTA That is the master dta well known file We will have to maintain it by hands It is in another directory than Contest ini There are also other files GRITTY ini is into appdata Afreet Products GRITTY and oth ers All well displaced into at least four different places with just appdata Afreet in common between each That s appdata a structural tree for Mss standing for Wherever disk_letter witcheveristheuser Ap pdata Whereveruser data for this Program PAGE 4 OF 7 DRAFT V 004 s_are_stored The Program and MS could im pose its security by Obscurity rules on it This is one of the first issue appdata It usu ally correspond to C users USER NAME OF THE TIME Appdata Roaming And it is a hidden and name variable directory It is changing as change the USER NAME of one of the registered USER On different or even in the same PC More on it will follow But anyway back on analysis just looking at those ideas of statistical recognition leave From a TX data string like IW1AYD 599 004 004 IW1AYD the RX resulting string could be IW1AY 599 034 004 IW1AYD well valid but still needing operator human correlation to be engaged GRITTY aiming II This second aiming is considering to whom GRITTY is aimed to be used My understood is that GRITTY
15. on for all our problems when in Contest There is not such a solution more seasoned contester well know this Others sometime forgot this truth However out of a nowhere place for such a so lution we have to establish some criteria and how to meter around those That is my person al view about all this Well before than leaving he place to whoever is much more technical than me would apply my nerd abilities to say about integration ease of use operational capabilities and so on Think about it as how is easy to drive this car GRITTY was presented as to be integrated into N1MM Sided by this was quick and dirty te sting it personally would not have too much interest in a not integrated decoder My focus is where the action is Action components must stay well tied together N1MM Classic and N1MM are the living examples Human Correlation We usually apply some kind of correlation whenever we look at things Some of us better than others Some of us try to train their self each time in any contest It is not that unusual to see at least two decoders windows printing together different words letters symbols or simply garbage That last is not all garbage of ten there are mangled word data vital for our hobby RTTY Contesting just in case Decoders by their self apply different technologies for the tones decode and the most recently devel oped bits correlation or lost bit recovery Aim ing to write the most of
16. s if just one time is enough or note Le s say TIUTI2WER12 and TOO312PQW we have almost a 12 in this case as QW is also 12 That s easy isn t It s much more easy in the contest where ex changes are letters and letters that make sense State Names locators those last at least are signaling two letters and two digits Still this make our RX error prone but not hav ing four letters on hands JNRE is JN35 at least What s GRITTY doing to correlate GRITTY correlate in a differed ways Applying information it has about specific callsign apply ing informations it has about generic callsigns applying information that it has about contests specific exchanges schema The first is the well known master dat binary formatted file The second is a file containing the most rele vant types of callsign The third and maybe the more interesting as it is the driver for all the greatest changes that GRITTY introduces at the user level is a file that contain relevant descriptor signatures for any contest exchange that was taken in account So not all the fancy ways we have seen not for all the contests How think it could work It s working applying the signatures as parame ters with internally hardcoded rules Plain ASCII files may have some sort of user modification How this not human correlation will work Plainly as a finite state machine doing pattern stuffing recognition at your PC speed quite tire less PA
17. will be used by contest enthusiasts early adopter They will have advantages on using it maybe some They will have some way to test different GRITTY behavior not that much Right now for common users with low and scarce knowledge of PC GRITTY is just an other decoder to look at when having the screen real estate to use it The underlining question is would be GRITTY aimed to end user with different characteristics than RTTY Skimmer volunteers Different answers will require substantial changes to GRITTY Also as GRITTY is now free of any charge to use it several other questions may arise from this same points Integration and with OS amp NIMM Too much integrated into the OS usual for MS hate or love it It has all the needed files com ponents into the MS programs sub directories Just the executable and the ini are moveable It is extremely not portable by itself It need care INTEGRATING WITH NIMM GRITTY 1 2 IW1AYD to be moved along together with is dispersed ancillary files into different PC MS M2 MM installations Further attempts to make it portable will follow With NIMM any attempt to have a DI sub window populated by incoming data hadn t worked at all Clicking a single call decoded by GRITTY move that call inside Main Input win dow call sign field OK BTWw FLDIGI suffer the same integration syn drome Configurability Out of the portability issue some parameters are not configurable by cli
18. ying UBN understanding Or the pie would result already sliced and somewhat lacking on slices 4 1 having another networked subsystem un der the hood could increase with high or low tops the lack of HW amp SW OS resources Carv ing all that onto the same environment even if asynchronously driven by OS buffers could be too much for two cores two treads machines 4 a metering could be empirically considered having a look to other inter program communi cations transposed into Inter networking com munications The inter networking communi cation making programming code communi cate each other make sense when prototyping since the UNIX roots of the method or pre paring to slice applications in a Client Server behavior It doesn t make sense for local facili PAGE 6 OF 7 DRAFT V 004 ties even if more easy to be done reusing code already made I E syslogd vs dmesg From the same environment as used for any other SO MO contesting NIMM last rev MTTY double 2Tone 24 1920x1200 to adding one copy and then a second of GRITTY 30 From that start to some more checks 3 hours a scarce dozen of HRD and QSO 1 hour Platform HW SFF Mac Mini Early 1012 17 2 3 4 GHz 16 GB 1 GB HD not SSD OS W 8 1 latest patches as yesterday Other applications are there not to be used with the logger To have low latency is and ongoing task not a simple check when all was installed lt 150uS Radio interfaces are
Download Pdf Manuals
Related Search