PWT buffer


#1

Yeah, someone better get in or this thread will go off in never never land. :-B



Sorry, I took Sunday off to play with my tapes. Lots of old memories there. I had forgotten just how bad some of the recordings of the bands were, which is a shame since the music’s really good. Oh well, there are a few gems that can be saved and lots and lots of interviews. It’ll be a fun project.



I am not 100% sure on the buffer. It is certainly an intelligent buffer, nothing like a straight FIFO of a fixed length. That much I remember. It is pretty consistent that when it’s empty, say at the beginning of a new CD placed in the instrument, it fills up as quickly as it can and there you get about 20 seconds almost immediately. And it pretty much keeps that going for most tracks.



Part of its variability depends on the disc itself and how easy it is to read that disc. Also, it depends on the optical ROM mechanism as well. If the two are in synch then the buffer time is smaller and if they are out of synch larger. We determined long ago that 1 second is all one needs for sonic purposes. Anything longer buys nothing sonically. The rest is up to the machine and its needs.



Also, I had asked for another rule set to be installed. If the device goes from nothing to playing, that first track is supposed to load as much into the buffer as possible regardless of synch. I asked this because when I do shows and seminars, demonstrations of the PWT, I wanted to be able to perform my showman trick to maximum effect. I usually ask someone in the audience for a disc, and someone inevitably brings one. I play perhaps half the track, maybe more, then I open the drawer and hand the disc back to the owner as the PWT continues to play for another 20 to 30 seconds. Or, sometimes I simply play the track, pull it out and show the folks it’s in memory, then put it back in. The PWT is designed to find its place and keep going if you can do that within 10 or so seconds. That was programmed in for no other reason that to help me sell PWT’s when I show it off.



The rest of its logic rules is a bit of a mystery to me. The fellow that wrote the program is somewhat of a quiet genius in Romania. Getting the rule set out of him is near impossible. He just tells me “it works”. Indeed, it does.


#2

This discussion was created from comments split from: How much of a memory player is the PWT, really?.


#3

So Paul, even in your case you’ve seen instances where, when the drawer is opened, the music might cut out after only a few seconds?



BTW, I have seen the case where the same CD reinserted will pick up right where it left off. That’s pretty clever.


#4

Yes, certainly, it’s different once the album is playing for a while. Generally I do get about 20 seconds but not always.



It is a clever trick and one that the programmer spend an additional two weeks of programming time on just to make my little demo work. It has zero value otherwise.


#5
Gordon said: his discussion was created from comments split from

Excellent thinking, Mr. G.

Paul McGowan said: 1 second is all one needs for sonic purposes.

Very cool, but makes sense. As long as the bucket remains full enough for the water to flow out the bottom hose, it does not matter how big the bucket is.

#6

Can anyone comment on whether the drive in the PWT has itself a proprietary FW installed, or is it the stock FW that comes with the Asus drive? Note I’m not referring the PWT FW.


#7

Yes, it has proprietary firmware in the drive itself that Asus writes for us. The primary purpose is to make sure the drive can do low speeds.


#8

Thanks Paul.


#9

I own a PWT and i’m just wondering what happens when I need to change the drive. Do I order one from PS Audio or will an off the shelf DVD writer work without the special Asus firmware?



Also, is there a way to tell if my PWT has a SATA or PATA drive without opening up the unit? I know the PWT is quite easy to open up but I’d rather not do it until the drive breaks.


#10

Paul said:

"The rest of its logic rules is a bit of a mystery to me. The fellow that wrote the program is somewhat of a quiet genius in Romania. Getting the rule set out of him is near impossible. He just tells me “it works”. Indeed, it does."



This explains why an email conversation I had with Dave Pannamen 2 years ago to try to sort out one annoying feature of the PWT memory buffer was only partially successful in achieving a solution. The ‘Pause’ function does not produce a pause on play in all instances: if the play is not resumed within a short time (about 5 minutes) then play resumes at the beginning of the track rather than where it was stopped. This is very inconvenient if there is a need to interrupt one’s listening during a long track (e.g. to answer the phone). I made a series of tests; to quote my mail:



1) Original firmware

"Using a stop watch (iPhone app) I looked at the effect of gap length when pausing a ‘typical’ track:

(a) control by remote

For a 3 minute track with gaps between pause and play of 0, 5, 10, 15, 30, 60, 120, and 300 sec - all restarted correctly.

As I practically never pause short tracks (typical jazz & pop) this was news to me!

For a 25 minute track (a far more typical use of pause for me) all but the 300 sec gap gave correct restart of play. Behaviour for the latter case was as described in earlier email(s); restart, play for a few seconds before stopping, with further depression of play causing the track to start playing from the beginning. Further investigation showed that a gap of 180 sec produced correct restarting of track but a gap of 240 sec did not, giving the same behaviour as the 300 sec gap.

(b) control by touchscreen

I repeated the above tests with the 25 minute track - all played perfectly!

I wondered if the problems encountered via the remote might be due to ‘spent’ batteries so I put some new ones in the remote, but this did not change the behaviour described above.

I repeated some of the tests with the remote control, but using a long track on a CD with pre-emphasis. This produced the same behaviour as that described above for the 25 minute track.

One other thing I did notice was that almost always at 128 sec into a pause there was a mechanical noise from the drive (parking the laser head?). Probably not significant as the 180 sec gap gave no problems."



2) After Dave Paanamen supplied modified firmware

"As before I paused in a 25min track for a range of times.

5 minute and 7 minute pauses resulted in the music continuing to play correctly whereas longer pauses of 8 and 10 minutes caused the incorrect behaviour described previously.

Definitely better than before, where failure occurred for 4 or more minutes.

In the real world this behaviour is probably adequate (to answer the phone, make a coffee etc. etc) so I guess most folk could live with it, but ideally a pause should be a pause whatever the length. I assume that this would be difficult to achieve with a ‘memory’ player."



David