Philips Respironics System One REMstar Pro file format

I am looking for any information and/or ideas people may have about the file formats of the files created by the Philips Respironics System One REMstar Pro C-Flex+ CPAP machine.

I have recently been diagnosed with Sleep Apnea and have started using a REMstar Pro C-Flex+ machine. I looked into getting the software (Encore) for analyzing the therapy data that the machine records during operation, but found it to be too expensive. I also feel that the software should come with the machine and don't feel that it would be right to pay for it on top of the price of the machine. I found a free program called MyEncore but it requires the Encore software to already be present, apparently because it just visualizes data that Encore has generated after analyzing the raw files created by the machine.

I decided to take a look at the files produced by the machine, so I read the contents of the SD card onto a computer (the SD card used by the machine has a standard UDF filesystem on it). I found the following structure:

/P-Series/last.txt
/P-Series/<serial number>/properties.txt
/P-Series/<serial number>/p0/0000000001.001
/P-Series/<serial number>/p0/0000000001.002
/P-Series/<serial number>/p0/0000000001.005
/P-Series/<serial number>/p0/0000000002.001
/P-Series/<serial number>/p0/0000000002.002
/P-Series/<serial number>/p0/0000000002.005
/P-Series/<serial number>/p0/...

The content of some of these files is easy to determine:

<serial number>: The letter "P" followed by a 12 digit hex number.

last.txt: Pointer to the <serial number> folder where current records are stored.

propterties.txt: Human readable list of unit specific values: SerialNumber, ModelNumber, ProductType, DataFormat, DataFormatVersion, Family, FamilyVersion, SoftwareVersion, FirstDate, LastDate=, BasicKey, DetailsKey, ErrorKey, PatientFolderNum, PatientFileNum, EquipFileNum, ValidCheck

.001: ?

.002: ?

.005: A series of 1-byte signed integer samples.

Each set of files consists of a .001, a .002 and a .005 file. I presume that each set consists of what the machine considers a therapy session. That is, from usage starts and until it stops or until the Ramp button is pushed.

The main bulk of data is stored in the .005 files. A quick look at these files in a hex editor showed what looked like a 24 byte header followed by a series of 1-byte signed integers. This file can be imported into audio editors such as Audacity using the raw import function and setting the sample size to 1-byte signed integers. The resulting curve shows was looks like breathing patterns, so presumably the samples are from the pressure in the machine.

I have not looked closely at the header data in the .005 files or the data in the .001 and .002 files, but this data looks like it will be much harder to reverse engineer than the sample data in the .005 files. Since the sample data is unencrypted, I don't think any of the other data is encrypted either.

Any feedback and speculation on the contents of these files is welcome. Some questions are:

- Since metrics such as Periodic Breathing and AHI is provided in the user interface, the machine must do some of the analysis of the raw data on board. Is this analysis data also stored on the card and if so, what is the format?

- What, if any, analysis is not performed on board, but by the Encore software?

- Which other data does the machine store on the card?

- What are the algorithms that the machine uses for turning raw sample data into therapy metrics such as Periodic Breathing and AHI?

$99.00 is too much to see how your therapy is working? If you think you're going to reinvent the wheel, you'll be spending far more time than what you'd pay for the software.

https://www.cpap.com/productpage/respir ... chine.html

That's half the price that it used to be and I thought it was a helluva bargain at that price......better than going back to doctors and having them ask "How do you feel?".

The MyEncore software only worked with the 1.4 thru 1.6 versions. The Encore Pro Analyzer worked up to the 1.8 versions.
Without those kinds of diagnostic programs that used an MSSQL database, there is no more reason to use the Encore Pro version.

But, yeah, I've always thought it should be included with the machine........except that there are now too many PC operating system variations for the manufacturers to write that many versions of the software.
On the other hand, the Encore Pro software IS available for a VERY affordable price ......if you know where to look.


Den

The P-Series files are encrypted, methinks, or use variable length binary records. Not worth the effort, IMHO. Just buy the EncoreViewer 2.0 product and browse the literature for the explanation of the indices. Sadly, I could not google the calculations for the indices.

Regards.

Though I am sure both of the followup posters mean well, I really have to take them to task.

The OP asked a question that is germane, well explained and specific, neither answered it
or moved the discussion forward. Yes, it is a good thing that Encore Viewer 2.0 is now $99,
but it still has a lot of limitations.

Keep in mind it is a viewer, not a system which can continually collect and manage the data to
assist you in adjusting settings. Can you introduce additional data sets and correlate to your
sleep experience? Imagine being able to compare sleep events with relative humidity and room
temperature. Imagine begin able to link energy usage to this to fine tune your power needs
(I live where power fails periodically so I run my setup on a UPS, I like to camp ad bring a
small battery in summer when it is humid and a large setup for the winter when it is dry and cold.)

Right now I manage all of that with spreadsheets, I'd rather have a program that incorporates the
data direct from the device.

Imagine if you could adjust the machine by changing half the variables each night through a
reducing binary sort. A program will tease out the significant factors quicker than changing one
variable at a time.

I will say the good thing about those well-meaning but ultimately unhelpful replies is keeping the
good-for-some program and new pricing in peoples minds, and it is probably needed in the confirmation
of any third party software written, but ultimately discovering the data format is what is really wanted.

The irony and possibly best example of the need to reverse engineer the data is the software that
this cpaptalk (and most internet) forums are written in, not a manufacturer provided pay software such
as Microsoft Sharepoint but phpBB a freeware GPLed piece of software.

Sorry to sound harsh as that is not my intent, just to open ones eyes to the possibilities. Just like many MDs
have downplayed our requests for studies and answers regarding OSA, lets not downplay a valid computer
data question just because ones own needs seem to be met by the Encore Viewer.

That said, it looks like I may have pop this idea higher on the roundtoit list. I have been waiting and wishing
someone else would do it.

Has anyone made any headway on this issue? I've also started looking into the fileformat a bit. If anyone else is still looking at this, it may be useful to share ideas (and datafiles). Please post!

Continuing with rdrd's work, I'm also able to get the presumable graph of pressures from the 005 files.
Sample zoomed:


However, the full plot seems to indicate a few reference points/lines included. Three seeming constants above and one linear below:


The mass of red is due to the horizontal squishing of a lot of datapoints. However, the four "artificial" lines I mention above are quite clearly visible. Anyone know what these lines might indicate?

EDIT: I think the sloping line is a clock function. It appears to run consistently at a slope of approx every 1300 samples throughout all datafiles, modulated at range of 256 and centered on the origin.

EDIT2: Sorry, I'm a little slow. Had a couple beers in me. Of course it's the header (that rdrd identified) repeated every so often. Not 1300 but 1280 (=2^8*5). The header is, at least partially, a timestamp. It's also used in the 002 and 001 files (bytes 12-15 = timestamp).

Thank you for looking into the file format and data. This has been very interesting to me. I hope you folks with more technical skill keep digging.

I noticed Audacity seemed to have the first five minutes of my waveforms when reading the .005 file, but the EncorePro waveform from that day did not -- the printout started after that time span. The end of the sleep time looked the same on the EncorePro printout and in Audacity. I wonder if the software leaves out that first five minutes on purpose. I can tell my breathing hasn't settled in during that initial settling in period. (I don't use the ramp feature, so don't know if that further shifts reporting).

I am also pondering the slight difference in numbers shown on the LCD each day, compared to the EncoreViewer or EncorePro reports. My AHI numbers may be off in comparison to each other sometimes, maybe by .1 Perhaps that is related to the truncation of 5 minutes? I reset (zero out) the memory each day, to see the night's data rather than 7 & 30 day averages. One morning my AHI (on LCD) was 2.2 and Periodic Breathing was 6%. The printout from the Encore software showed AHI 2.1 and PB 5.4%

I'm using EV 2.0 with Win XP running on an Intel Mac and Boot Camp.

Would someone enlighten me as to how to find the ".005 File" ?

M.D. Hosehead asked:

Would someone enlighten me as to how to find the ".005 File" ?

When I put the SD card in my Mac, I look at the folder on the card "P-Series", then there is another folder inside that which is labeled with the serial number of my machine, then within that a p0 folder. Each day's data is in that folder. Each run of the machine seems to produce files which are named by run number (so your 99th run would have lots of zeros, then a 99) followed by a .001 and another file the same except followed by .002, and the wave form file which is .005.

For safety, I don't dig into the card with Audacity -- I copy the whole P-Series folder to my Mac and play with that.

SleepyCPAP wrote:M.D. Hosehead asked:

Would someone enlighten me as to how to find the ".005 File" ?

When I put the SD card in my Mac, I look at the folder on the card "P-Series", then there is another folder inside that which is labeled with the serial number of my machine, then within that a p0 folder. Each day's data is in that folder. Each run of the machine seems to produce files which are named by run number (so your 99th run would have lots of zeros, then a 99) followed by a .001 and another file the same except followed by .002, and the wave form file which is .005.

For safety, I don't dig into the card with Audacity -- I copy the whole P-Series folder to my Mac and play with that.

Wowzers, it works swell

Thanks, SleepyCPAP.

I got the graph.

The index numbers along the top run from 0 to 3.40 for a tracing of about 8 hours. I'd like to zoom in on a specific time to, for example, see how long a hypopnea lasted. I could probably find a particular region of the graph by calculation, for example 3.40=480 minutes (8.0 hrs X 60minutes), therefore 171 minutes falls at 1.211 in the graph. But every night's tracing would require a new set of calculations.

Any thoughts as to how to make the index numbers display a more useful unit of time, such as minutes or hours?

Thanks again.

On the left of the waveform in Audacity (at least on Macs) there is a pull-down menu labeled 0000000 (near the X that closes the waveform box) and the choice at the bottom of that menu is "Set Rate." At the bottom of that "Set Rate" submenu is "other." I entered 18320 Hz and it seems to work that .5 is on the half hour, 1.0 is the first hour, etc. I see it gets off slightly as time progresses, but that Hz seems close enough.
Still not giving minutes at that setting, but perhaps some other Hz would change the numbers to something useful for zooming in on an event and gauging the timing. Let me know if anyone figures that out.

This thread has an interesting mix of sw savvy Respironics and ResMed users.
One potential aspect that interests me: even if the raw data doesn't have the same digitized resolution in time and pressure, is it possible to transfer it from one hardware and sw and computer and operating system to another?
Ideally, as in other signal processing applications, a neutral format has been standardized to simplify this transfer. The opposite is true in industries fighting each other, like the cell phone locking business in the US.

More specifically: let's say that I move to another region, and seek another sleep physician. I was told by a sleep tech that the raw data saved on an SD card from a ResMed S9 Autoset home CPAP (EDF format) can be read into the Encore analysis sw, then track long term trends. Would anyone have a confirmation of this, and have some link to detail specs?

Hi again everyone! Just checked back and glad to see some people exchanging ideas about this!

You guys are right, the data seems to be recorded at a fixed sample rate. Therefore, it should be pretty easy to figure out the actual times. However, the timestamps are also injected at every 1526 bytes inside the 005 file, so this means two things:
1) The timestamp could be used to periodically verify or re-sync the calculated sample times
2) Raw graphs of the 005 files are every so slightly incorrect because they're plotting this timestamp as if it were real data

You can see the timestamps included in the graphs i posted above. The regular repeating patterns...

I wrote a little python script to simply filter out these timestamps. Can share if anyone likes, but the basic equation for the correct bytes is
byte>7 and (byte-7)%1526 > 13

I've made a small web application that will let you visualize data on your .005 files. In order to use, you'll need to be able to retrieve your 005 files (as explained earlier in this topic) from your SD card. There are some examples from my own therapy sessions there, as well.

The application will plot the data and make a simple effort to detect "Events". Please note these events are not defined by the CPAP machine nor do they necessarily have any physiological or medical significance.

I've only tested it with data from my own machine, so if you try it and notice any problems, please let me know (here or via the email link on the application).

Lastly, I'm sensitive to the privacy concerns you may have about submitting your CPAP data to a website. Please read the short privacy statement on the website before you submit any data.

Visit the PRS1 Data Viewer at http://www.hoolehan.com/onkor/

I hope you find this useful!

Wow, cpapdork! Very impressive! Really cool for something that operates as a web app. I don't have the EncoreViewer 2.0 Software for Respironics Machines (nor the $99 for a new copy), but I do have a respironics machine with an SD card, and I have a computer. So, what's missing for me is how close a correlation there is between the output of your web app and the EncoreViewer 2 sw. Do you have encoreviewer? Have you made that correlation? Would you be willing to sell your used copy of it? With all rights of course.

Is there anyone here who would like to sell their copy?

I asked the Apria Home Healthcare rep about reading the SD card and he said I'd need the software and that it costs thousands of dollars. I told him I could probably google my way into the data, and he relented a bit and said, yeah, you probably can. My searching led me here. Lucky me.

Snowed in like a mother here in New Jersey.