ShowAnalyzer Website Gone - Anyone Have the Installer?

marvin-miller

Posts: 221
Joined: Tue Sep 18, 2012 9:50 pm
Location:

HTPC Specs: Show details

#21

Post by marvin-miller » Sat Apr 08, 2017 10:02 pm

marvin-miller wrote:
1.x is the version that destroyed Jere and killed off SA completely. Back in the day everyone went right back to .9x after trying out 1.x Jere himself was pretty ashamed of 1.x and finally gave up on on it publicly stating that he had programmed himself into a box with it. He took a LOT of very public flack for it and it was embarrasing to him.

1.x was very uncomfortable for everyone involved and brought an end to SA as a product. Historical fact there :mrgreen: Jere was never able to recover from it and it was the beginning of the end.

technodevotee

Posts: 963
Joined: Thu Sep 10, 2015 4:10 pm
Location: West Midlands

HTPC Specs: Show details

#22

Post by technodevotee » Sat Apr 08, 2017 10:19 pm

You said that already but what, specifically, is so bad about it?

marvin-miller

Posts: 221
Joined: Tue Sep 18, 2012 9:50 pm
Location:

HTPC Specs: Show details

#23

Post by marvin-miller » Sat Apr 08, 2017 10:41 pm

If you haven't used the pre-1.x release you won't understand what's wrong with 1.x

technodevotee

Posts: 963
Joined: Thu Sep 10, 2015 4:10 pm
Location: West Midlands

HTPC Specs: Show details

#24

Post by technodevotee » Sun Apr 09, 2017 7:11 am

As I said in another thread,
technodevotee wrote:I used SA for years on Windows Vista and it was always very good. When I switched to Windows 7, I couldn't get it to work properly and gave up (mostly due to all the faffing about causing diminishing WAF). Since switching to Windows 10, I've been having a problem with fast forward/rewind during playback or catch-up causing the Media Center video to freeze (posted another thread about that) and started playing with Ad-skipping again.

I think the detection method used by SA is a bit out of date and doesn't know about some of the tricks the broadcasters use to defeat such things but I'm going to persevere tweaking the settings to try and improve it.
Maybe I missed that version.

marvin-miller

Posts: 221
Joined: Tue Sep 18, 2012 9:50 pm
Location:

HTPC Specs: Show details

#25

Post by marvin-miller » Sun Apr 09, 2017 7:25 am

technodevotee wrote:I am currently running SA 1.1.0.825 and, apart from the fact that it seems a bit less accurate at picking out adverts than the older versions I have used, it seems pretty solid to me.

I assume that the stations use some tricks nowadays that confuse it but that it can be tweaked to improve the detection or am I deluding my self?

What is it that makes it so bad in people's eyes?
That's where School House comes in and was one of the key reasons 1.x ended up being the straw that broke the camel's back - no schoolhouse support.

I took one of my most complicated modern recordings from the other day, transcoded it from H.264 to MPEG and ran it through the old Show Analyzer (.97x). It found all but one commerical segment. I opened the file up in Schoolhouse and adjusted two settings and re-ran the detection - it found every one. Support for profiles per channel meant you could use a custom detection scheme on a per-channel basis. All of that led to very good commercial detection and Schoolhouse meant that when a certain broadcaster decided to get 'tricky' you could just open up the recording in Schoolhouse, re-run the analysis, adjust as required and update the profile. Pretty much a 10 minute job - or less.

This is why pre 1.x was so good - Schoolhouse. There were a lot of other things wrong with 1.x but it's really easiest to just put it this way, pre 1.x and everyone was happy other then folks using .wtv and H.264.

1.x was an abysmal failure and ultimately caused Jere to abadon the software altogether. He publicly admitted that 1.x was a failure, that he programmed himself into a box, that the softwarekey system caused a LOT of problems and that basically he would have to do a ground-up re-write of the software to make it work with modern (.wtv/h.264 - as memory serves) DVR's.

So it used to be a choice of Comskip or SA - in it's day SA was way better. I'd still say that pre 1.x is the way to fly. The problem is, in my case, I would have to transcode every recording to MPEG and convert to dvrms in order to run it through SA. I tested this the other day with that one recording, and while commercial detection was flawless it seemed to take as long as the original recording to convert to mpeg.

It seems to all be moot though as I lost my email with the key from back when I bought it.

technodevotee

Posts: 963
Joined: Thu Sep 10, 2015 4:10 pm
Location: West Midlands

HTPC Specs: Show details

#26

Post by technodevotee » Sun Apr 09, 2017 8:51 am

Yes, I remember schoolhouse now - it was very good.

But that was years ago and now I get satellite from all aver Europe and Terrestrial from the UK so there's a lot of stuff that uses h.264.

So, what with that and the fact that everything is in wtv files, the choice is either to continue using the newer version and try to tweak it or to go back to using DVRMS Toolbox to convert everything and go back to the old version.

Not sure which is the most work....

technodevotee

Posts: 963
Joined: Thu Sep 10, 2015 4:10 pm
Location: West Midlands

HTPC Specs: Show details

#27

Post by technodevotee » Mon Apr 10, 2017 8:28 am

technodevotee wrote:Not sure which is the most work....
Update (to recap, I have two spare Windows 7 PCs set up with SA Suite v1.1).

As a test, I removed the v1.1 suite and installed SA with SchoolHouse v0.97 on one of the PCs.

I copied some WTV files containing commercials from my server to each machine. I left the files as WTV on the one with v1.1 and converted them to dvrms on the one with v0.97.

In almost every case, v0.97 had a lower detection rate than the ones processed natively by v1.1. Next, I converted the files on the one with v.1.1 to DVRMS and there was the same difference between the detection rate in WTV files and DVRMS files.

Next, I tested the ability of v0.97 in SchoolHouse mode and it didn't make any difference. In fact, in some cases, it was worse than ever.

So, my quick and dirty test seems to indicate that transcoding the files makes things worse and using v1.1 to natively process WTV files is the way to go.

technodevotee

Posts: 963
Joined: Thu Sep 10, 2015 4:10 pm
Location: West Midlands

HTPC Specs: Show details

#28

Post by technodevotee » Mon Apr 10, 2017 10:29 am

Nah, may have been a bit hasty there.

Some more testing shows that .97 IS better at detection than 1.1 in programmes recorded on certain channels.

Back to the drawing board then.

Maybe a mix and match approach?

technodevotee

Posts: 963
Joined: Thu Sep 10, 2015 4:10 pm
Location: West Midlands

HTPC Specs: Show details

#29

Post by technodevotee » Wed Apr 12, 2017 6:38 am

Continued my ramblings at http://www.thegreenbutton.tv/forums/vie ... =9&t=10803

On a side note, I see that http://www.babgvant.com has finally disappeared.

The repository at Sourceforge https://sourceforge.net/p/babgvant/code/526/tree/ is still available at the moment but basically, DVRMS Toolbox users are completely on their own now.

adam1991

Posts: 2893
Joined: Sat Jun 11, 2011 2:31 pm
Location:

HTPC Specs: Show details

#30

Post by adam1991 » Wed Apr 12, 2017 10:41 am

last Black Friday I picked up a Tivo for $200, lifetime service. Glad I did.

I also have a Fire TV box on order; I will Kodi that thing and then pretty much not care if I miss a recording.

The end, she is nigh.

dkrom

Posts: 255
Joined: Mon Nov 17, 2014 11:39 pm
Location:

HTPC Specs: Show details

#31

Post by dkrom » Sun May 14, 2017 9:12 pm

technodevotee wrote:Continued my ramblings at http://www.thegreenbutton.tv/forums/vie ... =9&t=10803

On a side note, I see that http://www.babgvant.com has finally disappeared.

The repository at Sourceforge https://sourceforge.net/p/babgvant/code/526/tree/ is still available at the moment but basically, DVRMS Toolbox users are completely on their own now.
I just found that babgvant.com is back up and running!

marvin-miller

Posts: 221
Joined: Tue Sep 18, 2012 9:50 pm
Location:

HTPC Specs: Show details

#32

Post by marvin-miller » Mon May 15, 2017 4:08 am

technodevotee wrote:Nah, may have been a bit hasty there.

Some more testing shows that .97 IS better at detection than 1.1 in programmes recorded on certain channels.

Back to the drawing board then.

Maybe a mix and match approach?
Yes, as mentioned, .97 was good and 1.x was bad :wtf: :yawn:
The benefit of Schoolhouse was that you could alter the detection settings and re-run the detection, live, in real-time, and see whether you adjusted them correctly. This gave the end user the ability to tailor the detection on a per-channel basis.

To date, the functionality has never been eclipsed. As I mentioned earlier, I took a modern H.264 recording, converted it over to MPG2 DVR-MS and ran it through Schoolhouse. With a minor adjustement it found every single commercial on a recording that Comskip routinely puked on.

So it still really is that good. The problem was that it would not work with H.264 and .WTV files. That was it. Unforunately, Jere deicided to take the next version in a 'new direction'. Many people, at the time, tried to warn him about this but he did it anyway. All the problems in 1.x killed off Show Analyzer entirely. Jere himself mentioned several times that he had programmed himself into a corner. The writing was on the wall.

So the creator of Show Analyzer himself knew that 1.x sucked - to the extent that it ended his work on it. Not sure why anyone would then think that 1.x was better in some fashion - it's not. It never was, and clearly, since the author abandoned it, it's never going to be as good as .97

Now if someone has the key for .97 - that would be really handy....

Post Reply