Issues with TV program download, One of Us

artisticforge . artisticforge at gmail.com
Sun Sep 11 18:59:19 PDT 2016


Hello Vangelis;

How goes it?

It is nice to have some who is able to see the trees in the forrest. ;-)

It did not occur to me that the beeb may be lying. ;-)

to be honest i download episode 1 several time on various computers
all with the same results.
it should have dawned on me that the beeb were lying.

I will plead Programming under the influence of morphine.
Been in a morphine fog all weekend.

there are things worse than death.


On Sun, Sep 11, 2016 at 8:35 PM, Vangelis forthnet
<northmedia1 at the.forthnet.gr> wrote:
> On Sun Sep 11 14:33:31 BST 2016, artisticforge . wrote:
>
>> attempting to fetch One of Us.;
>> http://www.bbc.co.uk/programmes/b07r26py
>> (snip)
>> Recording:  1365.44MB / 2278.88MB  3400kbps  59.9% 00:36:40 remaining
>> Recording:  1367.72MB / 2278.88MB  3400kbps  60.0% 00:36:35 remaining
>>
>> INFO: Recorded: 1367.82MB in 00:54:55 at  3401kbps to
>> ( snip) /One_of_Us_-_3._Glenarvon_Loch_b07v6kv1_original.partial.mp4.ts
>> (snip)
>> I checked the episodes and it appears that they are in fact complete.
>> (snip)
>> One_of_Us_-_3._Glenarvon_Loch_b07v6kv1_original.mp4:
>> Track: 1
>> Type: video
>> Info: H264 High at 3.2, 3469.600 secs, 3056 kbps, 1280x720 @ 50.000000
>> (snip)
>> ../get_iplayer-2.96/get_iplayer --force --verbose --type=tv --mode=best
>> --pid=b07v6kv1 2>&1  | tee -a
>
>
> On Sun Sep 11 19:01:45 BST 2016, Alan Milewczyk wrote:
>
>> I hadn't noticed that all three programmes
>> seem to be around 1300MB rather than the 2 Gigs that one
>> would expect from one hour recordings at 1280x720 50 fps.
>> However, as you say all programmes are complete.
>> (snip)
>> Puzzled!
>
>
> Many greetings to you both Terry and Alan!
>
> I shall stick to data only pertaining to Episode 3 of the series,
> as the same applies to the previous two episodes...
>
> What the original post of terry's documents is in fact
> caused by a discrepancy between the file size reported
> by the beeb's stream metadata (prior to downloading)
> and the actual size of the file extant on the corresponding
> CDN server.
>
> GiP uses the stream metadata to calculate a rough
> estimate of the file size to be fetched; terry is using
> GiP 2.96 and --modes=best for a TV prog, this
> defaults to --tvmode=hvfhd to be tried first for the
> download.
>
> If you perform a verbose info query for pid=b07v6kv1:
>
> perl get_iplayer-296w.pl --pid=b07v6kv1 -i -v > VerboseInfo.txt
>
> and you then scroll down the log to the "Found mode"
> section for --versions=original (the second set of modes),
> you get:
>
> INFO: Found mode hvfhd1: (gip_hvf_iplayer_5380) hls h264 1280x720 5380kbps
> stream (CDN: mf_limelight_uk_hls/2)
> INFO: Found mode hvfhd2: (gip_hvf_iplayer_5380) hls h264 1280x720 5380kbps
> stream (CDN: mf_akamai_uk_hls/1)
>
> As you can see, GiP expects the bitrate of the
> hvfhd file to be 5380kbps, yet the probe
> of the downloaded file, as performed by terry,
> reveals the actual bitrate to be:
>
> V=3056kbps + A=128kbps ~ 3185kbps
>
> 3185 ~ 60% of 5380, hence GiP reports
> the download to be complete at 60% of
> the initial size estimate.
>
> There you have it! Nothing to be puzzled by!
>
> The iPlayer team encoded that particular
> series (and possibly other programmes,
> as hinted by Alan) at a lower bitrare than usual,
> however they have not updated the
> relevant stream metadata to reflect
> that reduced bitrate - no fault of GiP's!
>
> =============================
> Note to the developer:
>
> Examining the above verbose log for the
> audiodescribed version, vpid=p0471vhc,
> reveals not less than six (!) hvfhigh tvmodes:
>
> hvfhigh1: (gip_hvf_iplayer_956) hls h264 704x396 956kbps stream (CDN:
> mf_limelight_uk_hls/2)
> hvfhigh2: (gip_hvf_iplayer_956) hls h264 704x396 956kbps stream (CDN:
> mf_akamai_uk_hls/1)
> hvfhigh3: (gip_hvf_iplayer_989) hls h264 704x396 989kbps stream (CDN:
> mf_limelight_uk_hls/1)
> hvfhigh4: (gip_hvf_iplayer_1757) hls h264 704x396 1757kbps stream (CDN:
> mf_limelight_uk_hls/1)
> hvfhigh5: (gip_hvf_iplayer_956) hls h264 704x396 956kbps stream (CDN:
> mf_akamai_uk_hls/1)
> hvfhigh6: (gip_hvf_iplayer_956) hls h264 704x396 956kbps stream (CDN:
> mf_limelight_uk_hls/2)
>
> Yes they're all at 704x396 resolution,
> but at 3 different bitrates, 1757kbps (possibly 50fps),
> 989kbps and 956kbps.
>
> Loading
> http://open.live.bbc.co.uk/mediaselector/5/select/version/2.0/mediaset/iptv-all/vpid/p0471vhc
> http://open.live.bbc.co.uk/mediaselector/5/select/version/2.0/mediaset/apple-ipad-hls/vpid/p0471vhc
>
> it appears that the 1757kbps bitrate is indeed at 50fps
> with 128kbps audio and is dervided from the "iptv-all" mediaset:
>
> #EXT-X-STREAM-INF:BANDWIDTH=1800000,CODECS="mp4a.40.2,avc1.64001F",RESOLUTION=704x396,FRAME-RATE=50
> vf_b07v6hy4_64e1fd28-440a-42d0-9f97-08792f7cff6f.ism.hlsv2-audio_1=128000-video=1570000.m3u8,
>
> The 989kbps bitrate is at 25fps with 128kbps audio, also
> derived from the iptv-all mediaset:
>
> #EXT-X-STREAM-INF:BANDWIDTH=1013000,CODECS="mp4a.40.2,avc1.4D401E",RESOLUTION=704x396,FRAME-RATE=25
> vf_b07v6hy4_64e1fd28-440a-42d0-9f97-08792f7cff6f.ism.hlsv2-audio_1=128000-video=827000.m3u8
>
> while the 956kbps bitrate (25fps) with 96kbps audio (HE-AAC)
> are all derived from the "apple-ipad-hls" mediaset:
>
> #EXT-X-STREAM-INF:BANDWIDTH=979000,CODECS="mp4a.40.5,avc1.4D401E",RESOLUTION=704x396,FRAME-RATE=25,AUDIO="audio-aach-96",CLOSED-CAPTIONS=NONE
> vf_b07v6hy4_64e1fd28-440a-42d0-9f97-08792f7cff6f-audio=96000-video=827000.m3u8
>
> I haven't tried to actually download any of those hvfhigh modes.
> I have a gut feeling the above stream detection is bugged;
> hvfhigh tvmodes should not have been derived from
> iptv-all mediaset, only from the apple-ipad-hls one;
> I think what breaks GiP here is the existence
> of supplier="mf_bidi_uk_hls" in the available CDNs for iptv-all too!
>
> And the duplication of (CDN: mf_akamai_uk_hls/1)
> and (CDN: mf_limelight_uk_hls/2) (four 956kbps
> modes detected) can also be put down to CDN bidi
> (more likely unsuccesful removal of the unsupported
> "mf_bidi_uk_hls" streams from the audiodescribed
> version).
>
> Please have a look...
> =============================
>
> Regards,
> Vangelis.
>
> _______________________________________________
> get_iplayer mailing list
> get_iplayer at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/get_iplayer



-- 
terry l. ridder ><>



More information about the get_iplayer mailing list