nicesplice - joining produces short output

cdmackay

Active Member
I've just started using nicesplice (and how wonderful it is, thank you).

I'm having problems joining two recordings: the result is much shorter than it should be.

To test it out, I used a pair of recordings I had, of the same programme (two transmissions on the same channel on consecutive days). The first recording was missing the beginning, the second missing the end. Creating a single recording that covered the whole programme seemed like a good test.

I used bookmarks to trim both recordings; that left me with a 1 minute recording that contained the start, and an almost hour long recording that had the rest.

I then used the cmdline to join them, but the result is only 2 minutes long:

Code:
humax1# nicesplice -in The\ Fantastic\ Mr\ Feynman\ 1  -in The\ Fantastic\ Mr\ F
eynman\ 2  -out joined
progLen = 63s, 0 bookmarks, HD = 0
Read 1599 entries from nts
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Read 88048 entries from nts
progLen = 3523s, 0 bookmarks, HD = 0
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Wrote 89647 entries to joined. Stripped 0 packets (0k) of EPG data
New Program Length = 145s
now shrunk

The odd thing is that the file sizes looked correct: the joined file was roughly the sum of the sizes of the other 2 files. But when playing the joined file, I get the 1st minute from the first file, then there's an obvious join, but it jumps to the last minute of the second file.

If I play both individual files, they look entirely correct. Neither file has any bookmarks.

I'm happy to supply whatever further information might be required, to shed any light on this, but the obvious oddness in the extract above (New Program Length) seems to sum it up.

Any ideas, please? thanks much...

Web interface version: 1.0.18
Custom firmware version: 2.23 (build 2035)
Humax Version: 1.03.12 (kernel HDR_1.02.32)
 
Here's the details of the 2 files:

Code:
humax1# hmt The\ Fantastic\ Mr\ Feynman\ 1.hmt
Format:SD
Title:The Fantastic Mr Feynman 1
ITitle:The Fantastic Mr Feynman 1
Channel:19 (Yesterday)
Folder:/mnt/hd2/My Video/Calum/Docs/science/keep/
Filename:TThe Fantastic Mr Feynman 1
Genre:News & Factual (32)
EPG:Family and friends contribute to this touching tribute to prominent scientist Richard Feynman, who helped develop the atomic bomb and solved the riddle of the Challenger disaster. [AD,S]

Flags: SD,Unlimited Copies,Shrunk,
Copy count:0

Scheduled start:1402315200 (Mon Jun  9 13:00:00 2014)
Scheduled duration:3600
Recording start:1402315195 (Mon Jun  9 12:59:55 2014)
Recording end:1402315258 (Mon Jun  9 13:00:58 2014)
Duration:63
Stored duration: 3599
Play resumes at: 0 seconds in.

Service ID (SID):25792
Event ID:43862
Transport Stream ID (TSID):24640
Originating Network ID (ONID):9018
Programme Map Table PID (PMTPID):1003
Video PID:301
Audio PID:302
Bookmarks:0 =

humax1# hmt The\ Fantastic\ Mr\ Feynman\ 2.hmt
Format:SD
Title:The Fantastic Mr Feynman 2
ITitle:The Fantastic Mr Feynman 2
Channel:19 (Yesterday)
Folder:/mnt/hd2/My Video/Calum/Docs/science/keep/
Filename:TThe Fantastic Mr Feynman 2
Genre:News & Factual (32)
EPG:Family and friends contribute to this touching tribute to prominent scientist Richard Feynman, who helped develop the atomic bomb and solved the riddle of the Challenger disaster. [S]

Flags: SD,Unlimited Copies,Shrunk,
Copy count:0

Scheduled start:1402257600 (Sun Jun  8 21:00:00 2014)
Scheduled duration:3600
Recording start:1402257603 (Sun Jun  8 21:00:03 2014)
Recording end:1402261126 (Sun Jun  8 21:58:46 2014)
Duration:3523
Stored duration: 3599
Play resumes at: 0 seconds in.

Service ID (SID):25792
Event ID:43847
Transport Stream ID (TSID):24640
Originating Network ID (ONID):9018
Programme Map Table PID (PMTPID):1003
Video PID:301
Audio PID:302
Bookmarks:0 =
 
And the joined result:

Code:
humax1# nicesplice -in The\ Fantastic\ Mr\ Feynman\ 1  -in The\ Fantastic\ Mr\ F
eynman\ 2  -out joined
progLen = 63s, 0 bookmarks, HD = 0
Read 1599 entries from nts
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Read 88048 entries from nts
progLen = 3523s, 0 bookmarks, HD = 0
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Wrote 89647 entries to joined. Stripped 0 packets (0k) of EPG data
New Program Length = 145s
now shrunk

humax1# hmt joined.hmt
Format:SD
Title:The Fantastic Mr Feynman 1
ITitle:The Fantastic Mr Feynman 1
Channel:19 (Yesterday)
Folder:/mnt/hd2/My Video/Calum/Docs/science/keep/
Filename:TThe Fantastic Mr Feynman 1
Genre:News & Factual (32)
EPG:Family and friends contribute to this touching tribute to prominent scientist Richard Feynman, who helped develop the atomic bomb and solved the riddle of the Challenger disaster. [AD,S]

Flags: SD,Unlimited Copies,Shrunk,
Copy count:0

Scheduled start:1402315200 (Mon Jun  9 13:00:00 2014)
Scheduled duration:3600
Recording start:1402315195 (Mon Jun  9 12:59:55 2014)
Recording end:1402315340 (Mon Jun  9 13:02:20 2014)
Duration:145
Stored duration: 3599
Play resumes at: 0 seconds in.

Service ID (SID):25792
Event ID:43862
Transport Stream ID (TSID):24640
Originating Network ID (ONID):9018
Programme Map Table PID (PMTPID):1003
Video PID:301
Audio PID:302
Bookmarks:0 =
 
Can you crop the second file? Sometimes you get a glitch in a recording that makes cropping fail: if this is the case you will get an error message and a partially cropped file. If this is the problem, try auto-shrinking both sections first before joining. This has worked for me in the past.
 
thanks...

I do seem able to crop the 2nd recording:

Code:
Processing The Fantastic Mr Feynman 2 (inverted: 0)
Moving recording to /media/My Video/Calum/Docs/science/keep/_original
  The Fantastic Mr Feynman 2.ts
  The Fantastic Mr Feynman 2.nts
  The Fantastic Mr Feynman 2.hmt
  The Fantastic Mr Feynman 2.thm
CMD: /mod/bin/nicesplice -in {/media/My Video/Calum/Docs/science/keep/_original/The Fantastic Mr Feynman 2} -out {/media/My Video/Calum/Docs/science/keep/The Fantastic Mr Feynman 2} -cutBookMarks
Found bookmark at - 304
[QUOTE="MontysEvilTwin, post: 75988, member: 4074"]Can you crop the second file? Sometimes you get a glitch in a recording that makes cropping fail: if this is the case you will get an error message and a partially cropped file. If this is the problem, try auto-shrinking both sections first before joining. This has worked for me in the past.[/QUOTE]

progLen = 3523s, 1 bookmarks, HD = 0
Read 88048 entries from nts
cut at nan seconds = frame 7637 (304928) -------+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Wrote 80394 entries to /media/My Video/Calum/Docs/science/keep/The Fantastic Mr Feynman 2. Stripped 0 packets (0k) of EPG data
New Program Length = 3217s
now shrunk
Renaming file group to The Fantastic Mr Feynman 2-1418353040
Time taken: 83.808

and other than that "nan" oddity, it looks OK?
 
In case it's a problem in the index files, you could regenerate them using raydon's new sidecar package. As the joined file size looks correct, this may fix it.
 
@cdmackay: I was expecting the crop operation to fail partway into the 2nd file and highlight an error, but it looks OK. I would try af123's solution: I think my suggestion also works by fixing problems in the sidecars, but would take longer to run.
 
arg, thanks both. I was following the sidecar development with interest, so I should have thought of that too - I had assumed it was a metadata issue, as the file size does look correct.

Although oddly, playing the joined file doesn't work, in that it jumps from near the beginning to near the end. I'd have expected sidecar issues to affect skipping around, and elapsed/remaining time displays, but not to actually disturb play.

I'll test more tonight, thanks.
 
Although oddly, playing the joined file doesn't work, in that it jumps from near the beginning to near the end. I'd have expected sidecar issues to affect skipping around, and elapsed/remaining time displays, but not to actually disturb play.
That's exactly what happens with a bad .nts file.
 
You can prove the .ts data by deleting the sidecars (or disassociating them) and then seeing if it plays.
 
One point of caution: I believe nicesplice will work on undecrypted recordings, but deleting the .hmt file for an undecrypted (ie still encrypted) recording will render it useless.
 
I suspect it may be due to the first clip having time stamps significantly after the second. I have a feeling I saw this once, and didn't get round to looking into it. Does it work if you attempt to join them the other way round? I'll try some experiments here...
 
A quick bit of experimentation seems to be showing the problem is only with two recordings that have already been through nicesplice being joined in the wrong order (date wise). May be worth trying the opperation by joining the two original recordings first, then doing the cropping. Its going to need some debugging, but I've recreated it here, so its just finding the time. Maybe sometime over Christmas when everyone else is asleep in front of the TV...
 
A quick bit of experimentation seems to be showing the problem is only with two recordings that have already been through nicesplice being joined in the wrong order (date wise). May be worth trying the opperation by joining the two original recordings first, then doing the cropping.

thanks much indeed for looking at this.

Being short of space, once I'd verified that the initial crops were good I deleted the originals before trying to join the cropped files; sorry.
 
Back
Top