Ah, but they will say it 'functions as specified'; you see words like 'latencies' will not be in the spec, and provided it eventually displays that button, it is working to the spec.
It may well be unusable, and important control options may be missing, but it can still 'pass the spec'.
The fault is not so much the Programmer (who will be a low paid under-ling, not a System Architect), but more a basic failure in usability testing. Generally, the programmer is NOT the person who is best to test usability.
A programmer will test merely what they wrote, not what a customer might reasonably expect, or even pathways a confused customer might take.
Oh, I can believe the 6 seconds unfortunately. One Freeview DVD recorder I have can take several seconds between pressing the first digit for a channel before this first digit shows up on the screen. (It's a two digit channel BTW, so there _is_ a second digit to enter.)
Simon.
--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980s technology to a 21st century world
No that's the time waiting for all the filaments to warm up. It is so much longer than on WWII radios because todays TV sets are so complex, see. First we wait for a group of filaments to get hot, then the respective tubes work to produce the voltage for another chain of filaments. Which does that for the third one and so on. Similar to booting a PC, you know. A lot of warmup going on.
And your analogy is more apt that perhaps you realize, since most "Smart" TVs run Linux in a PC-like on a chip environment. (although more likely to be MIPS or ARM than x86).
I'm all about returning stuff I don't like. Of course I try my hardest to not buy stuff until I'm relatively sure it is what I want. I bought a toaster at a store where this model wasn't on display. On hooking it up I found the cord was only 20 inches or so long and came from the front of the toaster! I couldn't set it on the edge of the counter where I wanted it, so it went back. The next model was from a store where none of the display models had cords, they were cut off, I assumed because they were in the way. This model was $10 more than the first tried model. Same power cord problem, so it went back! Finally I paid $45 for one of the top of the line models at a store that had a unit on display with power cord intact so I was sure of what I was getting. 1 out of 3 isn't too bad is it?
Lots of places don't want you to return TVs, especially around Super Bowl time. That's one reason to belong to Costco, they'll take anything back although many electronics are only for a couple of months or something like that. Other things are literally forever if you have the receipt.
That is why I return things like this. It would bug me too much every time I use it... like my PC! lol I really don't have much tolerance for crappy user interfaces. Heck, I've returned $10 remote controls because they didn't work well enough for me.
I'm not too sure of the details, but my first guess would be that the quadrature amplitude modulation scheme requires a phase-locked loop to lock on the new frequency so that decoding can start. Of course, it could all be done with DSP, but it would still require a significant sample to determine the phase of the new signal.
Hopefully, someone who knows a lot more about digital TV can confirm or ridicule my hypothesis.
I don't know much about digital TV but the MPEG stream obviously takes some prebuffering to start (to fill the entire framebuffer once for the first time). Then I notice that the time to switch to a HD channel is significantly longer than to a low resolution one so these delays feel about right.
What does not feel right at all - but people have been trained to tolerate for decades now - is the boot process taking forever. If I can make DPS boot within 5 seconds to a usable state (obtaining the IP address via dhcp may take somewhat longer in background, this depends on the dhcp server etc.) then it should be possible for the others, too. Oops, wait, their code is 100+ times bloated compared to mine (which has been written to optimize my time, size has not been really a priority), among many other things.
But then they must still have those cascaded filaments I talked about earlier, not much to be done about it... :-) .
MPEG typically transmits frames(P,B) as delta's and for those frames it needs previous and/or future* frames to reconstruct the picture. Only once in a while a complete frame (I) is transmitted that can be decoded independently of the other frames.
A TV can only start displaying a reasonable picture after it has received a I frame. Therefore there is an inherent delay when switching channels. The only way to avoid that delay would be to always decode all channels regardless of what you are watching now, however that would be rather expensive...
*frames are not necessarily transmitted in the same sequence as they are displayed, therefore a frame can reference a frame that is to be displayed later.
One of the parameters one can play with is the distance between the I frames. Since I frames require many more bits to encode it is more economical transmit as few I frames as one can get away with.
The most responsive TV I have ever had was a model from 1978; after power up produced sound instantly and a picture within 3 seconds. Switching channels was instant. After that it went downhill as far as responsiveness is concerned.
I always wonder why it takes so long for a PC to boot. Given the amount of processing power, bandwidth...etc it has at its disposal it could do a huge amount of work in just a second. Looking back at the PC's I have owned the last 20 years it seems like the more powerful the PC the longer it takes to get it in a usable state.
I just checked the fairly low-end PC I use for newsgroups. It has about
75 processes running. Getting all those processes running and properly sharing resources is bound to take some startup time. Delays associated with plug-and-play hardware are probably a major factor in startup time as well.
These delays aren't just a Windows thing either---both Mac and Linux systems also have long boot times. All systems try to fool you into thinking they're ready well before they've finished loading all the drivers and starting all the background processes.
Thankfully, better-designed low-power modes have made full reboots less common.
With typical GOP (Group Of Pictures) parameters, the I-frames are transmitted at least twice a second, so I do not understand why the worst case channel zap time should be more than 500 ms.
Did you ever consider looking in the box before you left the store-- or even before you paid for the toaster?
And if we're really lucky, Costco displays intact versions of the items for sale. Now if they would only have a lady with samples in front of the wine display! ;-) (They get a sale on about 1 out of five of the sample of the other stuff I browse for lunch on a visit.)
On 2013-01-22 22:42, dp wrote:> I am setting up a new TV-set, a Philips
32pfl3807h . > > I had never seen a commercial product programmed in such a > moronic way. The programmer must have been outright illiterate. > i I discussed the settop box functionality with Viasat at length, explaining why the person specifying the functionality must be a moron on LSD, and they have decided to give me a permanent 50% discount...
If I delete some files, and change directory at the same time, it may a decide to only delete the i-nodes, keeping the files. Removing the disk and running FSCK from a desktop recovered about 1 TB...
If they run linux, like my Samsung TV, then a lot of the stuff has had plenty of review, except the frontend... Video Codec in a settop box, is likely done in hardware. BR Ulf
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.