Time stamps in SleepyHead with change to DST?

[Snork]

Same thing here. Hopefully there will be a fix? Great software otherwise!

[Pugsy]

For anyone with ResMed machines..(unsure if Respironics was affected)...SleepyHead Beta version seems to have gone a bit whacko when we changed to DST for a good many of us.
I am getting double sleep sessions reported..so double time asleep and the event count is doubled but the events themselves are shown correctly individually.
It may also have messed with my EPAP pressure reporting. Time stamps are off too.

Older version of SleepyHead (I have 0.8.8 still available on my computer) seems to work fine and ResScan is reporting same as older version of SH.
So it appears to be SH Beta version bug for some reason and possibly related to DST change. Mark is aware of it.
Initially we thought creating a new user profile would eliminate the issue but it didn't for me and neither did removing SH Beta and starting over (did it 3 times)..so for now don't worry about those time stamp things being off nor any double sessions you might see. I have played with various clock settings on the machine as well as the DST toggle in SH and nothing makes it right. It is going to need to be fixed on the coding end of things.

So we just need to wait till Mark gets around to fixing those little bugs. The basics of the data reported...AHI..Event times and calculations, leaks...etc. All seem to match up to ResScan..so other than slightly annoying to see that I am reported to have slept 15 hours last night...it isn't a critical issue at this time.

[NateS]

Pugsy, where do you see these errors? I'm flipping back and forth between SH/Mac and SH/Windows and I can't find any of the errors you mention in either version.

Are the errors in the graphs or in the numbers on the left panel?

Could it be related to whether the starting time is before or after midnight? All of the nights I checked seem to look right, but I almost never hit the sack until after midnight so my starting time is always in the AM.

Nate

I need to qualify the above by saying that I have never understood the numbers under Sleep, Wake and Hours in the left panel?

[Pugsy]

Nate,

You may not be affected... Not everyone appears to be and I have no idea why but it isn't relate to time going to bed.
This below is what I see.
Doubled events..note the total in each category in ( ) and the list with time stamp.


Note Sleep times here..
Sleep...22:55 time machine came on....Wake.. 06:38 time machine turned off....Hours...13.26 means 13 hours (and a few minutes) asleep using the machine


Double sessions...Along with EPAP reported as 4 Min...it is 9..
ResScan says the SD card data is correct as does SH older version..so it is only this Beta Versions acting up.

[NateS]

Okay, I guess I never looked all the way down at Session Information - I see mine are doubled there too.

But only there - my Events are not doubled like yours are. Not in any category and neither in Windows nor in Mac in SleepyHead.

But what are Sleep, Wake and Hours in the left panel in the Detail tab? I don't understand what they correspond to. Some nights it lists only 1.30 under Sleep when I know I've slept 8 hours or more. I thought the machines don't know whether you are awake or asleep. And the first two entries never add up to the third one.

Nate

[portiemom]

Mine are like yours Pugsey, I never changed the time on my machine, because I didn't want to mess up worse, so I just left it alone. Remember when we were posting that thing didn't add up? I'm confused enough about the time thingy anyway. Why don't all the graphs have the same time across the bottom?

[DrPepper00]

Okay, I guess I never looked all the way down at Session Information - I see mine are doubled there too.

But only there - my Events are not doubled like yours are. Not in any category and neither in Windows nor in Mac in SleepyHead.

But what are Sleep, Wake and Hours in the left panel in the Detail tab? I don't understand what they correspond to. Some nights it lists only 1.30 under Sleep when I know I've slept 8 hours or more. I thought the machines don't know whether you are awake or asleep. And the first two entries never add up to the third one.

Nate

Sleep is what time you went to sleep. Wake is the time you woke up. Hours is the amount of time you had the mask on. It is wakeup time minus go to sleep time.

[NateS]

But what are Sleep, Wake and Hours in the left panel in the Detail tab? I don't understand what they correspond to. Some nights it lists only 1.30 under Sleep when I know I've slept 8 hours or more. I thought the machines don't know whether you are awake or asleep. And the first two entries never add up to the third one.

Nate

Sleep is what time you went to sleep. Wake is the time you woke up. Hours is the amount of time you had the mask on. It is wakeup time minus go to sleep time.

Ahah! I knew it was going to be something that would make me feel stupid!

But thank you DrPepper, sincerely for the explanation. Now I know! And of course that makes sense!

Best regards, Nate

[Pugsy]

Why don't all the graphs have the same time across the bottom?

Do you go to bed at exactly the same time every night? That is the only way the graphs will have same time every night.

[NateS]

Today, SleepyHead says my AHI last night was 3.89, yet ResScan says it was only 2.3!

I do not recall encountering a large discrepancy in AHI reporting like this before.

In fact, I do not recall any previous discrepancies in AHI reporting other than rounding.

Nate

[Snork]

Not everyone appears to be and I have no idea why but it isn't relate to time going to bed.

Last night shows up correctly in my SleepyHead. Only thing changed is that I went to bed after midnight... The other nights when I went to bed before midnight still show up wrong.

Of course this isn't the cause of the problem, but I thought it could be useful input for correcting the problem...

[Pugsy]

I too went to bed after midnight (first time in a long time) last night and the time asleep was correct this morning.
It did show 2 sessions though...I don't think the machine went off in the middle of the night but I suppose it could have. Events totals are no longer double...but...the report shows I went to bed 2 hours before I did. Go figure that one. I had changed the clock on the machine to accommodate DST as I had previously changed it to regular time when all this stuff started in an attempt to sort out the issues.

Apparently SleepyHead and ResMed machines don't like DST for some reason.

Just now compared SH AH to ResScan AHI.. Pretty much spot on for last night's data.
SleepyHead.... 0.99 AHI and ResScan 0.9 AHI.

ResScan shows I went to bed as same time SH does....roughly 10 20 PM (thought it might be when I changed the clock and I may have inadvertently started the blower but I see an unknown event flagged at 11 50 PM... I definitely went to bed at 12:20 AM...I fell asleep on the couch sitting up and woke up right after midnight.

Maybe I will just change the machine clock back to regular time...at least then I am only one hour off.

Actually the time stuff is just a minor annoyance. I had same thing with PR S1 machine when DST hit.

[jedimark]

Hi Guys..

I'm gonna have a look into this today and see if I can get to the bottom of it.

I thought this DST stuff was only happening on ResMed.. But if it's happening on PRS1 too, it indicates a nasty bug in Qt's date-time handling stuff (I try to offload all that stuff on these libraries to avoid these kind of problems, it could still be a snafu in my code though.)

In SleepyHead I use milliseconds since UNIX epoch (Jan 1st, 1970) for all internal stuff. The PRS1 uses seconds since epoch internally, so I didn't expect it to have problems. ResMed needs conversion from LocalTime to UTC, then to ms since epoch, which are all conversions handled by the Qt libraries.

With the S9, I use seconds since epoch as the SessionID, which is supposed to be unique, however something is happening that gives a different result for the same date, depending on the date the function is run, resulting in a different SessionID, and tricking SleepyHead into importing that session twice.

The problem is replicating this is rather difficult. If it is a Qt bug as I suspect, some other programmer has bound to come across it and there will be stuff on Google.. I've just gotta keep digging. :/

Edit: Is anyone who's experiencing this bug building from source code?

/Mark

[portiemom]

Maybe I wasn't clear in my previous post about the times on my graphs. No I don't go to bed at the same time every night, but for instance the Event graph begins at 22:00 which I am guessing is 11:00 pm the graph directly below the Flow Rate graph begins at the same time but the AHI graph sometimes doesn't show the same time, it will begin with 00:30 which I am guessing is 12:30am so where is that hour and half? Most times the graphs match up, but then there are days it doesn't the times are whacky. I never changed the time on my machine, so Pugsey changing it may not make a difference.

[Pugsy]

Flow Rate graph begins at the same time but the AHI graph sometimes doesn't show the same time, it will begin with 00:30 which I am guessing is 12:30am so where is that hour and half?

Ahhh, that graph... okay...disregard what I said. I have no clue why it is whacko. I don't normally bother looking at that graph. I just checked mine. Scrolled through a half a dozen....Mine all match with the times of the other graphs. Don't have any idea why yours is off.
Aliens....