quote: Harold Hallikainen

---

Finally, there's interoperability. Stuff I design has to work with a bunch of stuff I did not design.

---

And each bunch is a unique combination, meaning that an update intended to fix a problem with one combination can cause one in another. I've only been working with digital servers and projectors for a few months, but the horror stories I've heard during that time have convinced me that firmware and software updates to these machines (and any other mission critical machine driven by a built-in computer, for that matter) should only be applied if they are absolutely necessary to solve a problem, or reduce or avoid a clear and present threat. If an update has been published to correct a bug that doesn't affect you or to add functionality you don't need or want, why risk it?
 

---

Posted by Steve Guttag (Member # 268) on 07-12-2014, 02:53 AM:
 
There you have it...updates come for several reasons and Leo hit upon the questions I ask when each update is issued. I'm a stickler for wanting release notes before I apply updates. I want to know what is being updated, what it is trying to fix and what "features" it is trying to add.

In the server world...most server companies had a "good" software version before the DCI compliance BS hit. Many of you will have noticed that my "favorite" Dolby software was 4.3.5.9. It was their most stable software, by far. Any customer that was not compelled to update for VPF reasons were not updated by me beyond that version. The same goes for the other server companies...e.g GDC 7.83 was pretty darn reliable and I still have people on that one!

In 2010, the series 2 projectors came out and, from what I can tell, the projector companies were so under the gun to have them in early 2010 since series 1 projectors were not to be manufactured after 2009 that they didn't get into server company's hands until very late. We were getting software updates to servers to get series 2 projectors to work on a DAILY basis and for a while, some software versions only worked with certain brands of projectors...talking to Harold's point about having to accommodate things not under his control.

So then there is a settling in period as each projector company rapidly issues updates to get their new product stable and the server companies having to do likewise until series 2 and servers all work with each other.

So then you get an industry requirement of "DCI compliance" and that throws a monkey wrench in EVERYONE's software. Think about it...it adds complexity for the sake of making the show HARDER to work successfully. It does not offer new "benefits" so the challenge is to add in this security without diminishing reliability. There are going to be unintended consequences to most any software change and there were PLENTY of them. Sometimes you get a sense of the struggles they went though on the "build" number. Dolby's DCI software debuted at build 48 of System 4.4.0. GDC took 108 tries at it before 8.01 worked reliably.

The DLP projector companies get another wildcard...TI can change their software on their own time table. And there will be unintended consequences to that. 3.2.340 was their most stable...and then they went to an unstable 4.0 and didn't stabilize again until 4.2. However, influences by the industry for things like high frame rate and 4K can force them to do their updates into version 4 software...but it has ramifications in each projector.

Which brings about another reason why software changes in the servers...the industry moves (for better or worse)...somebody says the next thing is 3D...so everyone has to deal with that...then they say 4K...so everyone has to deal with that...then they say HFR 3D...so everyone has to deal with that...and it goes on and on. Note, not all of the "fads" pan out but no projector/server company is going to be left out by not accommodating what may be the next big thing. Hey the industry moved to IMBs (or IMSes)...well that has a whole NEW set of software demands as well as hardware accommodations...and the cycle continues.

When you layer the various independent systems that have to react to a change including the change caused by another device to the overall change imposed by the industry, it all adds up geometrically.

Then you get individual components trying to fix their own bugs because software people are NEVER finished working on those so layer that in. Finally you have people adding "features" to software to either make it more enticing for resale or to satisfy customer demand.

Typically, when you buy a product, you are seeing a "snapshot" of where the software was in development when that product was produced on that day. It does not mean it is best or most stable.

I don't know anyone that is responsible for maintaining this equipment that would not welcome the day when software updates were no longer needed. We have that, for the most part, for the series 1 stuff. It is now frozen in time and that is sort of nice. Then again, lamp companies come out with new models (longer life) so you need a means to allow for those. I would welcome an "annual" software update for stale equipment (or maybe every other year or at least a means to update the lamp list!).

But when one updates a piece of equipment's software it is kind of like taking medication...you may be doing it for valid reasons but you always are concerned about the unintended consequences. And there are ALWAYS those.
 

---

Posted by Marco Giustini (Member # 4544) on 07-12-2014, 04:30 AM:
 
Back in the 90's when Internet had just born, it was not easy to get 'updates'. computer components came with their own floppy disk and a driver on it. That was it. The driver was only one and was supposed to work for the whole life of the component.
Upgrading Windows 95 was a pain in the neck, it was not automatic as today's Windows.

As already pointed out, this has an advantage - you don't need to worry about updates - and a disadvantage - your software never improves, never gets new features.

I have the feeling that when manufacturers knew that updating a piece of software was damn hard, they would put lots of efforts and testing trying to make sure the code was ok before releasing it to the field.
Today you can upgrade in minutes, sometimes automatically by the internet. Software houses just don't bother to check the code thoroughly: if a bug is found in the code, they can easily fix it.

I like the idea of the software being developed and improved. But as Steve said I would appreciate if the software was not rushed on the marked, tested properly and released every now and then.
I remember when the 745 was out, I was not finishing to download the new software and a new one was released!

To add to Brad's comment about new bugs being introduced every time, you're right, upgrading doesn't necessarily mean 'fixed bugs'. However, if it worked as you suggested, that would mean the very first version was the best one and all the subsequent ones would be just worse and worse

slightly OT, I had a customer the other day asking why a CPL would play subtitles on his 4.6.1-4 and not on his 4.3.5-13! (space in the subs file)
 

---

Posted by John Roddy (Member # 7348) on 07-12-2014, 05:46 AM:
 
Even when the effort is put into fully running a large set of proper tests on new software before it's released, it's still possible for even showstopper-severe types of bugs to pop up out of nowhere. I've watched that happen with GPU drivers on the PC market for years. They spend months developing it and testing everything, but when it's finally released, hardware starts exploding (no really, that actually happened. Three cheers for nVidia!). There are just too many variables out there to account for them all.
 

---

Posted by Carsten Kurz (Member # 5396) on 07-12-2014, 07:00 AM:
 
And ISDCF is struggling hard to find ways to make cinemas update their software in order to transition to SMPTE DCPs...

For most cinemas, software updates are not free, they would need to be forced to upgrade both server and projector.

'The Hobbit HFR' was not such a bad case - even as I consider HFR useless effort, it allowed at least a part of the installations to grow to a defined state. HFR would be advertized and cinema operators had a clear incentive to update their systems. Test footage was made available and only when successfully played the HFR feature could be booked.

Unfortunately, it was still rushed on the IMB side (Doremis runtime switchable HFR firmware is a real joke, Dolby wasn't even ready), and it only targeted a small portion of the installed equipment, most of it quite new anyway.

They would need something similiar to get the SMPTE transition going. But which studio will take that risk to force cinemas into software updates in order to play a certain feature? Maybe the best idea would be to issue an important trailer long ahead of the main feature, then weigh the responses. Sooner or later this transition to SMPTE DCPs needs to be done.


- Carsten
 

---

Posted by Marco Giustini (Member # 4544) on 07-12-2014, 07:13 AM:
 
John,

nVidia have to deal with thousands of hardware configurations and millions of hardware/software combinations.

A DC server deal with a minute fraction of those combinations and the server itself runs on a very defined HW setup where everything is known.
 

---

Posted by Steve Guttag (Member # 268) on 07-12-2014, 10:46 AM:
 
Marco...the DCinema server companies probably have also comparatively smaller software engineering teams too...it all scales.

I guess they will need to do something like they did with Cyan soundtracks...start with a trailer that most people will want (a tent pole movie) and offer it only in SMPTE...make sure it is flagged as only playable on SMPTE compliant servers...and then you find out who is ready before it becomes a disaster. You can make the first feature a non-tentpole...like they did with cyan but you would want to pick a trailer that would get everyone's attention. Even better, get the heads of the five families...er studios to each agree to issue one trailer that way.

On the other side of the coin...as an exhibitor...it is definitely pissing them off for HAVING to constantly spend money (in labor or licensing or hardware) to update a WORKING SYSTEM to accommodate the whims of the industry.

I already have some customers that are getting sticker-shock over what this stuff costs to repair/replace. We are seeing "End-Of-Support" notices on things from the series 1 era (both projectors and servers) so even getting stuff fixed is becoming iffy. You may not even get 10-years on some of this equipment if it was bought towards the end of the series 1 era.
 

---

Posted by Marco Giustini (Member # 4544) on 07-12-2014, 01:37 PM:
 
well you're supposed to have your equipment serviced every now and then and SW upgrades can be done then. I don't see it as a big deal as long as updates do not come out the first Tuesday of every month!

Sure Steve, but some SW released were obviously lacking basic tests. I can think of 4.6.0 - the cat 862 was flashing white when selecting the show - or some of the Doremi SM which were downgrading the firmware at the same time. I'm sure you'd agree with me that those are not rare glitches, they should have been picked up during a proper beta testing stage.
 

---

Posted by Steve Guttag (Member # 268) on 07-12-2014, 01:59 PM:
 
I think it was clear with Dolby 4.5 and 4.6 that their "head" was in getting the CAT745/DSS220 going and the other stuff mainly came along for the ride to remain compatible in a mixed booth.

4.7.0 was the the one that they seemed to really try to get the bugs out...that was in beta for at least 3 months with as many systems as they could find, from what I can tell.

4.8.1 is now out and that addresses about a million CAT745 issues...it also opens up the Barco presets to 40 from a mere 14. I hope it is a good one...we are starting to "beta" test it now...even though it is production release for Dolby...it remains beta to us until proven otherwise.
 

---

Posted by Marco Giustini (Member # 4544) on 07-12-2014, 02:55 PM:
 
sure but isn't that what we are talking about? 4.6 was released for Atmos support but, still, a white screen when SELECTING the content is not a small bug.
I mean: if you test that software you WILL HAVE to select some content to play!

We can talk about that old release (4.2.x?) which was never clearing the "checking licenses" message from screen, not a rare issue.

But don't get me wrong, we both agree that Dolby have been doing a good job on the software, that is not the point. Also, I happen to mention Dolby examples but I'm sure I (or anybody else) can come up with similar examples for any other manufacturer!
 

---

Posted by Steve Guttag (Member # 268) on 07-12-2014, 03:05 PM:
 
As I recall, 4.2 finished on 4.2.1 (3) and it was reasonably stable for what it was. It is such a blur now but the whole series 2 fiasco was a nightmare in 2010 and it took some time before that settled.
 

---

Posted by Marco Giustini (Member # 4544) on 07-12-2014, 03:58 PM:
 
I can't remember which version it was.

Fiasco or not, if you test that version for more than 45 seconds you WILL realise that it's seriously bugged.
 

---

Posted by Monte L Fullmer (Member # 2797) on 07-13-2014, 12:09 AM:
 
Thx for all of the comments to my question - very informative and a wealth of information to learn from being rather new in this digital world.

With all the mentioned complications that can follow updates, it prob makes one wonder on, "why did we have to go this route in the first place, where film and related items never had to go through all of this outside of technology to present the medium better."

..but we know all of those answers why the change had to come into place, whether we liked it or not.

- Monte
 

---

Posted by Marco Giustini (Member # 4544) on 07-13-2014, 06:01 AM:
 
when Cyan came out an engineer had to install a new light source on your projectors, didn't they?

Consider that as the 35mm 'software upgrade'. On Digital you could have supported Cyan soundtracks with a SW patch!
 

---

Posted by Steve Guttag (Member # 268) on 07-13-2014, 09:59 AM:
 
Ahem....

I told everyone then...that the whole Cyan thing was a big waste of everyone's time and money and didn't do a damn thing! It was a pitiful attempt to keep film around just a wee bit longer by removing one annoying process from the processing stage that helped the exhibitor not one bit. As a result, we will forever have to handle cyan/non-cyan in the archival world (different crossmods) all for the last 10 years of film's mainstream use (and only for 35mm). It was stupid.

If nothing else, it was a primer for the studios forcing exhibitors to change the way they do things, spend their money on a whim.
 

---

Posted by Marco Giustini (Member # 4544) on 07-16-2014, 06:48 PM:
 
I see now why you said "tons of fixes for the 745" on the 4.8 release... WHY Dolby are not sending a notification email to notify when a new SW is released???

It looks promising!
 

---

Posted by Monte L Fullmer (Member # 2797) on 07-16-2014, 11:49 PM:
 
Probably didn't want to look bad due to 4.7 having issues after 4.6 ...
 

---

Posted by Marco Giustini (Member # 4544) on 07-17-2014, 06:16 AM:
 
they never send notifications, you have to go and discover new versions by yourself!
I saw the issue corrected but I don't know what kind of issue to be honest. Another thing to be discovered
 

---

Posted by Steve Guttag (Member # 268) on 07-17-2014, 06:34 AM:
 
4.7 was not a bad release, in my opinion. The fixes in 4.8.1 to the CAT745 go back to the CAT745's debut. 4.5 and 4.6 was not Dolby's finest moment, for sure. 4.7, for us, was reasonably stable. 4.7 address numerous problems too. 4.8.0 was buggy, in my opinion. It never got out of our beta test site for us. 4.8.1 has been glitch free for us, so far but we've only been playing with it for a few weeks.

I agree Marco...it would be nice if Dolby and Christie and well...all of them got on the same bandwagon that Barco is already doing and announce software updates as they become available. It is not that I advocate always updating...far from that. But one should know that a new release is out, be able to look over the release notes to see if there is a fix for a problem you are having.

What I wish was a DCI requirement was a set of release notes to be required for each release. Sure most will send you something if you beg but you shouldn't have the ask. You took the time to release a new version...why not post what it does? At least Dolby gets that right...there are always release notes with each release.
 

---

Posted by Marco Giustini (Member # 4544) on 07-17-2014, 08:32 AM:
 
Excactly Steve,

I don't like upgrading without testing it first but if I am upgrading and a new release is supposed to fix something I have been having trouble with, I'd like to know asap!

And to be fair, and to go back on topic, I don't really like this behaviour: a X.Y.0 for 2 weeks and then the final release out with all the bugs fixed. I'd call it 'field beta testing'.
 

---

Posted by Steve Guttag (Member # 268) on 07-17-2014, 10:36 AM:
 
I agree with you on that one too. That is one reason that we beta test it ourselves for a month before pushing it out to other theatres. It should be noted...that on Dolby's software, the third digit is not necessarily a fix for an earlier version. That is, 4.8.1 is not a fix for 4.8.0. At some point, you define what a release will be and assign a number to it and it is stuck with that number and only the build number can grow...so if you see an x.y.z with a build number that is higher...then that is a bug fix for the same version. Witness 4.6.0 (31) and 4.6.0 (32).

Generally, with Dolby the second digit denotes some sort of new functionality. With 4.8, it was the "intermission" feature. So any versions that have the intermission feature is going to be 4.8 or higher. So if anything, when they increment that second digit, that is when the software is bound to be its weakest. They added something and possibly all of the ramifications are not yet known.
 

---

Posted by Marco Giustini (Member # 4544) on 07-17-2014, 02:06 PM:
 
yes, I got the picture when I installed 4.6.0-x and the 862 was displaying a white field when selecting content! 4.6.1 took just two weeks to replace that.
Note to self: never install a x.y.0-z version again.
 

---

Posted by Steve Guttag (Member # 268) on 07-17-2014, 04:19 PM:
 
The problem is not the x.y.0 versions though. Sometimes, they are very solid. 4.4.0 (48) was not particularly bad though I preferred 4.4.2 (10). 4.7.0 (42) was good too. The CAT862 firmware didn't change through all of the 4.7.x...though the DSP100 did! so you can get an idea of what they were working on there. 4.7.3 (3) is where the DSP100 will be leaving off...so hopefully it is good for those running DSS100/DSP100s.
 

---

Posted by Monte L Fullmer (Member # 2797) on 07-17-2014, 05:37 PM:
 
With all of us who are now working with digital and its related areas, a common situation has come to the surface with this.

The situation is: before when we were handling film related issues, almost the majority of the issues that we had to "be in radar range", were from human error.

Now, the issues that have surfaced in the digital realm, has almost totally reversed itself to where these issues are coming up that are well out of our hands and beyond our control.

Yet, the common denominator for both, is that we're still getting the pointing finger and being blamed for it - we have to fix it.

With film, our repairs did work. With digital, we have to see what happens.

-Monte
 

---

Posted by Martin McCaffery (Member # 37) on 07-17-2014, 05:57 PM:
 

quote: Steve Guttag

---

I agree Marco...it would be nice if Dolby and Christie and well...all of them got on the same bandwagon that Barco is already doing and announce software updates as they become available.

---

quote: Martin McCaffery

---

Sounds like a new forum category or at least thread for Film Tech.

---

I would love to see a forum dedicated to this. Way too much for just one topic. The subject of each thread in the forum would be the Name of the product and the version number. Features, benefits, fixes, bugs, etc can then be discussed in the thread.

Brad?
 

---

Posted by Marco Giustini (Member # 4544) on 08-03-2014, 05:18 PM:
 
BTW, the great software which was in beta testing for three weeks (I'm not teasing you, Steve, I'm pointing the finger at the manufacturers!) has now been replaced by another version because an important bug was discovered!
The previous version is not even in the list of old versions!
 

---

Posted by Steve Guttag (Member # 268) on 08-03-2014, 06:23 PM:
 
What version were you referring to? 4.8.1.17? It wouldn't be on the previous version when you see that 4th digit change, it is the same software but with a problem fix.

What else have I posted? Don't use software that hasn't been in the field for a month...that will find the stuff that didn't show in beta testing. The bug in 4.8.1.17 only affected some of the 6 terabyte machines and was depending on the hardware. As I recall it would affect a remote upgrade or an Install disc...not an upgrade disc. So as you can see it takes the right amount of circumstances for the problem to show itself...and the problem was just on the upgrade itself, not a system that is running it. It would hang on the update. One could run an update disc to clear the problem and not lose anything but their nerves and time.

I think they were slow in pulling off the download site, but that is my opinion.
 

---