I have noticed in N-Track when chopping sound is present, the graphic enhancements for each track occasionally disappear and reappear as the anti aliasing, shadows, lines and all the things that make up the visual aspects of the track seems to be reloaded after stopping and then restarting the playback of tracks. I believe this to be the cause of intermittent chop during recording and playback as the track textures are dumped then reloaded after stopping or starting a recording track. Why would theae textures get dumped in the first place? Memory Leak? I just bought a brand new labtop Asus M50vm a few days ago running Vista 32 bit and it too does the same thing. In fact because it uses less CPU power and I have to run higher buffer settings in line6 tone port and turn off all graphic enhancements for N-track or I get chop. I still get the odd chop but nothing like it was on default line6 toneport buffer settings and N-track graphic defaults
Right clicking on a track just recorded to bring up the option to delete so a new track can be recorded, it happens, texture dump and then graphic reload. I usually wait the 1 or 2 seconds it takes to reload these textures before I hit the record button. These textures should not disappear at all, as it takes CPU cycles to reload the textures and this can cause chop in playback and recording. I firmly believe once this issue has been addressed N-Track will run much better as stolen CPU cycles are limited if not stopped all together.
It appears when a track is deleted it dumps all the textures from all the tracks instead of just the highlighted track textures being deleted. I think this plays havoc on CPU cycles and physical memory infrastructure. Meaning that when this space in memory, reserved space for textures is dumped it gets replaced. The system then creates new space to reload textures in ram, shifting wave file data, but if the ram cache is not cleared, (dumped textures) when the file dumps, you’ll run out of memory and the swap file will take up the headroom lost by lost data that wasn’t cleared in the ram cache. This is known as a memory leak if I’m not mistaken. Also this presents other factors which can effect the playback of audio as the ram is being refreshed, it can move the Wave file data to a different section of the ram, and if this ram is being used, then the swap file will be used in it’s place. In fact, bits and pieces of the wave files can be in memory and in the swap file which will create chop. This is senn as my hard rive light goes on and off. That’s what I think is going on.
If I was to make any suggestions regarding this matter, Texture management needs an overhaul, space reserved in memory should be locked for textures for individual tracks (not global) as not to allow wave files already in memory the chance to be shifted or moved to the swap file as a result of the reallocation data process, (reloading of textures), that takes place as the result of data dumping.
Please pass this information along to those who may be interested in it.
As a side not Albeton lite ran flawless on Line6 buffer defaults on both systems. I hate Albeton by the way.
PACO
I just wanted to reiterate what I was saying in understandable terms.
The way it should work.
N-Track should be set in concrete, when you create a track, it’s beside it, when you remove it, only the wave file and associated textures gets punted. N-Track stays intact, set in concrete.
The way it is now
Right now N-track is not in concrete, when you remove a track, N-track breaks up into little pieces and then tries to put itself back together but those pieces don’t fit anymore so space is created to make it fit. Once the space is totally filled the bucket is used to make more space but it’s across the room wasting precious time. Johnny goes home sad.
LOL
PACO
are you talking about tv GUI - ? The graphics?
Yes…The wave image that is presented across the screen for each track.
PACO
Sorry…I don’t have clue what you mean.
The dumping of textures is causing fragmentation in memory leading to poor performance of N-track. I’m going to make a video showing it in action.
PACO
cool - I’ll make popcorn.
AH! Gotcha.
Maybe it’s time for a pool party.
As promised here’s the video, it’s 27 meg and 5 minutes long.
A few comments first. Test setup:
Asus m50vm labtop 2.56ghz/ 4 gig ram
Vista 32bit sp1 fully updated.
Running TonePort Line 6 UX2 gearbox
Ntrack v6
System monitor.
All default programs disabled, system setup for performance.
During the test system memory used was 30% of appox 3 gig available and stayed that way.
Here’s the link to the video.
http://files.filefront.com/Ntrackrar/;11873988;/fileinfo.html
I’m of the opinion, that when I remove a track “only” the track and it’s wave form textures should be removed. However this is not the case, all the textures from all tracks are getting dumped. I’m not certain when all the wave textures get removed that any data from the other waves files are effected. Nevertheless there seems to be some hard drive activity shortly after the textures are reloaded, indicating the possibilty of swap file use. I firmly believe that fragmentation occurs as result in memory due to the reloading of textures and wave file stored in memory might very well be effected and possibly sent to the swap file.
Furthermore, to combat this I sometimes I have to hit the play button, starting and stopping the song several times until the skipping stops. This indicates to me that it takes some time for all the wave files to be in memory. The skipping usually begins after deleting a track. This is obviously a problem, spans WINXP 32, Vista 32bit.
Since my 2 system vary in horse power, I have to run a much higher ASIO buffer on the slower machine to aid in combating the skipping issue.
I hope this information is helpful.
PACO
OH That’s what you’re talking about. Interesting theory. Great movie.
Will you let Flavio know about it, maybe show him the video, see what he thinks about it?
Paco
Flavio will see it I’m sure - I have no more an ‘in’ with Flavio than anyone else. ?He checks this discussion group almost daily. I’m sure he’ll explain how the graphics are interfaced. You may be on to something or not. I have noticed those track visual changes for a while but never associated it with overall performance.
I’ll see if I can upload the movie file to my server so it will stream.
I came across this issue in my attempts to get n-track to behave better on a brand new high end labtop.
I bought a new labtop, the reason for this is because my main system makes too much noise, fans and what have you. It interferes with vocals. I actually placed cardboard around my PC and then placed a few blankets around it in an attempt to isolate the PC sound it gives off. I had a couple of choices, build a sound booth to isolate myself from the PC or buy a quiet labtop. If you have ever seen what I can do as a carpenter, this choice was obvious. I’m no carpenter, a fence I built was proof of that, as it had to be torn down and replaced all 1000ft of it, but on the other hand I was an expert with a rake being a superintendent/rakerman of a asphalt paving company for 17 years.
I was about to return the labtop because I couldn’t get n-track to behave with any degree of stabilty. It was my wife, who at the last minute suggested I give it one more try. This is when I discoverd the issue at hand, the issue of this thread. As soon as I disabled the wave form features I notice right away the jerky playback was lessoned, further approaches with ASIO buffering seemed to help and finally I was able to get a song with many tracks to playback without any hiccups but this was with all wave form options disabled and high buffering in my ASIO driver more so than my main PC.
It was clear to me that this reloading of textures was playing a part in N-Tracks performance. Ntrack performs better with more horse power and the amount of ram appears not be an issue. Hence the ram fragmentation theory.
My main PC runs at 4ghz with 2 gigs of ram, the labtop 2.56ghz with 3 gig of ram but it requires higher buffering, no illusions here. More horse power is better for Ntrack at least in my case.
I hope this information will help make N-track better.
PACO
I’m sure Flavio will have some comment on this issue.
PC sounds are always a pain for doing vocals. I considered the same on my last system crash but went with another desktop.
In the mean time you might consider a “Quiet Case”. A lot cheaper than a laptop, and I hear they’re great with a quiet fan package.
Already got the labtop, and I’m lov’in it. It’s really nice. Thank you for the advice though, well appreciated.
PACO
Paco, just a couple of questons because I am not too good at reading the specs for a laptop:
It looks like it is a setperate video card (not on motherboard) - it is?
My laptop has turned out to be a big disappointment for audio. it was expalined to me that the IRQs on laptops are not assigned very well for audio and they cannot be reassigned as on a desktop (I thought Windows fixed all that, but apparently not.)
Do you know what the chipset is for the usb and firewire?
I have gone through a couple of laptops and finally moved the desktop in to another room, got 50 foot video cables and wireless keyboard and mouse. My video card is 1 gig pcie and I have some of the same video issues you are discussing. I will close a window (mixer, or the sound card or something) while I have ntrack up and the section that previously had an image leaves a white area instead of repainting. I was thinking maybe I had a bad card, but . . .
Bax
First off, none of the labtops I’ve looked at have any support asio except an add card from the creative, audigy 2 but it’s hardware is limited to software rendering only all done in software that bites, I use the UX2 Line6 Asio driver and this works good, by passes the onboard and uses it’s “toneMonitoring” which can be montiored thur the headphone jack provided or the SPDF to a set of speaker or 2 lines out to a montior, so you can by pass the onboard audio on the labtop no worries.
One of the advantages of the Line6 Tone port is the live monitoring, where as m-audio the signal must leave the card dry and it and then ads effect afterwards. Tone port wil process the effects in realtime to the line outs and SPDF channels, including the headphone jacks. A great plus. Ntrack see-s tone port ASIO drivers and theres is no tcack bleeding creatives ASIO causes.
Here; the link to my labtop
http://www.memoryexpress.com/Products/PID-MX21515(ME).aspxthe
1 gig 9600gt onboard. My main system has 2x 8800gt@sli
The problems I mentioned span XP on my main system in my sig and my new labtop and Vista. Buy the way I got the Gold bundle testerday. Line6 rocks…
I hope I answered your question?
Pick my brsin more if need be, I’ll be glad to answer more questions if you have any.
PACO
I noticed some redraw issues after reading this thread. I still have a pause in playback or recording if I have the plugin latency box checked (even if i don’t have any plugins on a project). In addition if I have the latency box checked there is a cpu spike upon starting playback…right to 100% (confirmed by the task manager)…The latest build has trouble starting. Since I have no midi input devices or cards it seems to hang for a while and then displays “midi device input error”…sometimes it won’t load at all. I checked and unchecked the midi preferences that Flavio suggested to no avail.
When I stop playback, the program hangs for about 5-10 seconds…frozen solid and then returns to a normal, stopped state. Does this with all my projects regardless of track count or plugins. This does suggest some type of memory leak and page file shenanigans. I am one of those people Diogenes refers to…a person who just wants a glorified pc based tape deck. I would LOVE to be able to shut off all the bells and whistles and simply play and record…mix with plugins and be done
Ray
Here’s a link to check out fellers.
DPC Latency checker
Some useful info on the page also.