July 18th: DSR530 Code CC Fixes, Features & Bugs

I just got back from vacation and am pleased to say that the machine has not locked up. For a change I don't have to crawl under the TV to unplug the 530 every day or 2.

I'm a little disappointed at the speed though. It's a really slow machine. I was hoping that the DVR LIST would appear instantaneously, instead of taking the several seconds. I also noticed that the jump buttons are slower than they were with the BC code. I mentioned this a few times before, and I'll say it again. The DSR530 has some definite hardware problems. I don't think that the firmware fixes are going to address a lot of the problems that still exist. Some of you may be waiting for dual recording. Don't hold your breath! The DSR530 is working as hard as it can just keeping up with the mono recording.

The good thing is that the lock up situation seems to have been addressed. For me, that was the biggest problem. The only thing that happened last night was the machine appeared to be locking. Everything went black on my screen and there was dead silence. About 10 seconds later it all came back up. Not sure what that was all about.

Possibly...

Kevin270 said:

...The DSR530 has some definite hardware problems. I don't think that the firmware fixes are going to address a lot of the problems that still exist. ...

Click to expand...

I guess I see the glass as half full, while you see it as half empty.
But given the fact that new code has changed the behavior in ways we thought may have been hardware failures, strongly suggest you may be in error.
I think now that there is substantially improved fundamental stability, there is clearly a time to do some optimizations and rethink the way some work arounds were done to get the box out.

I know before I left the project and had delivered the new IPG, there was noticable improvements in IPG navigation and usage due to some very simple re-thinking of the way things should be done.

While it doesn't matter to me any more, I really think you'll be surprised at how some simple changes have resulted in a much better performing box in many ways.

I hope they are able to bring it all together for the fall like they planned, but I'll be happy with a pre-Christmas release.

As for the speed, I guess they felt it was better to be stable that a speed racer that crashes.
Overall, I find the performance to be comparable with the Moto Cable PVR DCT6412 that has it's own shortcomings and is written by the Gemstar TVGuide group.

I have seen dual recording in prototype working, but it's not next on the list, Searchable & faster extended guide is.
A partial implementaion is set for Christmas with a little luck.

Kevin270 said:

While it is true that the DSR530 no longer crashes, the speed at which the machine operates has remained unchanged since May 2005. That was my point. If you have a slow processor or a slow computer, there is no software that is going to speed up the CPU.

Click to expand...

I would beg to differ. There are always optimisations that can be performed to code to increase performance on slower processors. TV Guide/Prevue Interactive worked wonders in improving the performance of the Motorola DCT2000 STB when they switched from their 'beige' IPG to the I-Guide. There's always 'fat' that can be chopped from bad code.

We know, for example, the 530 hardware is capable of dual recording, with playback of a pre-recorded stream or live buffer, all at the same time. Plus it has enough cycles left to run the IPG. Obviously the coding of the firmware/IPG is not efficient enough to handle this effectively at this time, so they limit the unit to one recording plus a recording or live buffer. Better programming will (or should) allow the unit to do what it was designed to do, all things being equal.

Now I'm getting a better idea of just how high up the totem pole ARR really was, I would like to ask one question -- how (if at all) did the choice to go for the IDE to USB connection for the hard drive impact the unit? Although USB 2.0 has theoretically enough bandwidth to handle three HD streams, the CPU utilisation with USB does tend to increase quite dramatically the more operations it has to perform (not so much with IDE from my experience). Did this hinder implementing 'true dual tuner' functionality, or was it just a time issue?

The USB<->IDE thunking was a major issue.
The engineer that spec'd it left the program quite some time ago for unknown reasons.

You are correct that there is sufficient bandwidth in USB 2.0, but there is always some latency with it.
The initial drivers for the NEC interface chip were quite problematical and underwent several updates to even become usable.
Again, the chip ( Read Hardware) was fine, but it was the driver code (Read Software) changes the solved the day.
Kevin270, are you paying attention here!

There was a very early prototype that was IDE only, but it never progressed any further.

If the 530 follows in the steps of the DCT6412 that went to SATA in their Phase III implementaion, then I'd expect better results.

But I don't see that happening until all the inital units are sold.

Dual tuning was a timing/code complexity issue, plain and simple.
The unit needed to be brought to market and a conscious decision to deliver it with a dedicated watch and record tuners was made.

As part of the upcoming searchable guide, the code foundation is being purchased from another vendor and incorporates most of the advanced features, of whcih dual recording is only part.
Things like conflict resolution and content retention are a large part of that as well.

So by having supposedly mature tested code over newly invented code and just integrating the engines into the IPG should make for a much better product as well.

The slow DVR list recall is not a hardware issues, it a s/w one in the way the data needs to be read/created.
A better form of indexing the table of contents will solves Kevin's issue.
They will get to that at some point, but as has been shown, they were focusing on some basic stability for the CC release.

I think additional optimization could be made to have performance of trick modes become comparable with the Echostar 501/Bell 5100 that does true read after write so all video is buffered through the hard drive.
That is the standard of performace I had asked for going forward.

No more clunky switching from live to LOD buffer and near instant trick mode response.

Remember, what they have achieved in ~ 1 year for a 1st unit.
Have other vendors accomplished as much with as little in such a short time?

If they are able to follow through on certain visions set out and listen to customers and prioritize activities, they will evolve the 530 into a remarkable unit.

New box - Same code - different issues

What I find curious is that now I have a new 530 apparently running the same code 00.A8 as I was with my old 530 that died, I now have new issues as well as the some of the same old ones. The DVI dropouts and delay in the optical audio compared with the normal RCA outputs is still there, but the optical audio delay only really appears to be significant on live broadcasts while the 530 is recording on the other channel. The DVI dropouts occur everywhere on HD, regular transmissions or when replaying recordings.

Howver my new box has a real problem when I hit the stop button while watching a recorded show. Unless I am watching the first recorded program on the list it defaults to another recorded program, changes channel and often locks up for several minutes. Sometimes it will sort itself out if I do absolutely nothing but if I hit delete or play it deletes the wrong program or tries to play another different recorded program. Curious. I did have cases when my old box would delete the wrong program in similar circumstances, but this is virtually 100% consistent!

I have been programing professionally at various levels since 1975 and was the CTO of a company running a software development lab in the 90's. I have seen some pretty sloppy work especially in recent years when programmers with little knowledge or understanding of machine language commands and basic DOS functions and routines attempted to fix programs by writing high level code over the top of badly designed basic functions. It never works; it just makes the problems more obtuse, more erratic and harder to duplicate. I believe that is the real problem of the 530 and until somebody bites the bullet and starts again from the basic functions and then builds a planned top level user interface utilizing sound basic routines the problems will never be solved.

Patches over patches never works for long! We can but hope for somebody to admit they screwed up and then start from scratch to get it fixed properly........

I'm with aspyrtus....

I have had similar problems with CC. I was on A8 since my 530 purchase in April, received CC in early July. I did a hard re-set (unplugged) the day after the download and have actually erased all of the programs recorded while on A8 (just to start fresh).

The 3 problems I have noticed include something similar to the one Aspyrtus mentions. While watching a recorded show the unit will just stop. Sometimes freezing. Sometimes it will drop out of the recorded show alltogether and switch to a live channel (I then have to go back into the DVR list, resume the show and it will sometimes start about 30-40 seconds before the point it stopped and sometimes at the beginning of the show). I have also hade regular audio drops on SD records. Sometimes they will recover if I "skip" back,... but that function is the second of the problems I've noticed.

Set to 10 seconds it will only skip back 2-4, so I adjusted it up to compensate. The most frustrating issue with the DVR functions under CC for me is after a FF or RW, when I hit the play button (even if I pause it first) it will start to play 30-40 seconds beyond the point I stop......frustrating.

My third issue is just wierd. The unit will just change channels! While watching live tv it will just freeze for a moment, the info bar at the bottom of the screen will pop-up and it will change channels! Sometimes 2 or 3 times. This has happened 4 or 5 times since the CC download.

I have 7 hours of recorded shows on the disc and NO timer set. ALL of these issues are new since the CC download. I had to re-set my unit once under A8 and this is MUCH worse.........

Anybody else??

R.