Back to Music front pageBack to CDM community forum indexBack to Motion front page
Create Digital Noise Forum Index  »  Hardware  »  Laptopping?
 This forum is locked: you cannot post, reply to, or edit topics.   This topic is locked: you cannot edit posts or make replies.
Page 1 of 3    Goto page 1, 2, 3  Next

Laptopping?

Author Message
midiguru
Posted: Wed Jun 21, 2006 6:24 pm Reply with quote
Joined: 21 Jun 2006 Posts: 28 Location: Northern California
In Keyboard's special Laptops Live issue a few months ago, to which Peter Kirn and I both contributed, editor Ernie Rideout opens the proceedings by stating, "...it's possible to have a laptop music system that totally rocks."

Nice hooplah -- to which I would now add, "Uhh, sometimes. Tried it, Ernie?"

I just bought a new Hewlett Packrat Pavilion dv8000, which has a 2.16GHz Centrino dual-core CPU and a gig of RAM. I installed a bunch of software, attached an M-Audio Fast Track Pro, and heard clicks and pops in the audio output when I tried playing softsynths. Switched to an M-Audio Firewire 410 -- same deal.

Neither of these interfaces produces clicks or pops when attached to my normal desktop machine, a custom-built 3GHz Pentium 4.

Over the course of a week, HP's online chat tech support provided many suggestions for how to tweak the system, none of which worked. They have now agreed not to buy the computer back, which is what I'd prefer, but at least to ship it back to the factory, check it out, and reinstall the factory Windows configuration. Which might conceivably help, but probably not.

I should note that the clicks occur with a reasonably large buffer (1,024 samples), no other software running, and no network connections hooked up. The computer is, in theory, just a musical instrument with a big display, and nothing else.

According to M-Audio, the most likely culprit is that all of the USB ports and the Firewire port are on "virtual IRQs," which means the system can juggle them around and thereby cause problems in the audio output buffer.

I've asked M-Audio if they can supply a list of laptops that don't use virtual IRQs. We'll see what they say.

In the meantime, I'm out more than $2,000 for a computer that I can't use to make any sort of music, and I'm just a teensy bit cheesed about it. So for the present, I'll be looking at all claims of, "Hey, you can make music with a laptop!" with an extremely jaundiced eye.

If anyone has horror stories, buying tips, or troubleshooting ideas to share, this is the thread for them.

--Jim Aikin
View user's profile Send private message
PeterKirn
Posted: Wed Jun 21, 2006 8:39 pm Reply with quote
Site Admin Joined: 03 Feb 2006 Posts: 822 Location: New York, NY
Hi, Jim, nice to see you!

Darn. My troubleshooting tip was going to be never buy HP. (I'm not joking.)

This sounds truly, truly bizarre. Laptop music works just fine; I've used FW 410 and Fast Track Pros with much more modest systems, like 1.5G PowerBook G4s and my trusty (okay, not so trusty) 1.5G Pentium M PC. Your dual-core machine should run circles around them. Something has to be truly screwy with this particular laptop.

I mean, I wouldn't say the "you can make music with a laptop" idea is suspect, because the issue here sounds like some sort of I/O bottleneck. Not to ask a dumb question, but I presume these problems occur with the ASIO drivers. (Tried WDM, out of curiosity? It should be a step backward, but just to see.)

I still see people playing around with old PowerBook G4s. My first Ableton Live machine was a PowerBook G4/400 (running Live 1.x). It worked, using a FireWire interface, though with limited processor load, of course.

I'd keep yelling at HP, too. Considered playing your press card to try to get them to take it back? If it does turn out to be the machine, you might consider which credit card you bought it on. Sometimes American Express will force a company to reverse charges on a non-functional product when the manufacturer refuses.

It is conceivable (despite my anti-HP rant here) that they'll be able to fix it, or that something is off on your Windows install. It's even possible there's actually a hardware problem inside the machine. As the recent MacBook debacle proves, these brand names often ship from the factory in condition no DIY builder would ever allow.

Speaking of which, though it doesn't help you at all now, Maximum PC did a nice feature a few months back, now available in their DIY special issue, on building your own laptop. It requires steady hands and nerves (to say the least), but the payoff is pretty nice.

Keep me posted on this; I hope you have better luck.
View user's profile Send private message Visit poster's website
masterslave
Posted: Wed Jun 21, 2006 8:54 pm Reply with quote
Joined: 18 Jun 2006 Posts: 37
ok now this may not fix your problem, but it's workaround for me. i've got an hp laptop and was running some drum tracks off acid while playing to it. then the battery died and i replaced it with hp travel battery. started getting skips and stutters from time to time and then noticed that it was due to the battery, if the laptop was on dc power everything's fine... weird but true!
View user's profile Send private message
midiguru
Posted: Wed Jun 21, 2006 9:11 pm Reply with quote
Joined: 21 Jun 2006 Posts: 28 Location: Northern California
PeterKirn wrote:
Darn. My troubleshooting tip was going to be never buy HP. (I'm not joking.)


Yeah, they had what looked like the best deal going on the Centrino duo T2600 2.16GHz processor.

PeterKirn wrote:
Something has to be truly screwy with this particular laptop.


What's almost as screwy is that in repeated interaction with HP's online instant messaging tech support, they can't figure it out. I keep saying, "Background processes interrupting the CPU." They don't get it. I could be wrong, of course, but it's a reasonable supposition, and should be tested.

PeterKirn wrote:
Not to ask a dumb question, but I presume these problems occur with the ASIO drivers. (Tried WDM, out of curiosity? It should be a step backward, but just to see.)


WDM not got. MME/DX got. When I switch to the latter, I can push the processor much harder -- like, up to 60% -- and only hear a very occasional click. But of course the latency is unacceptable for live synth work. With the ASIO driver and a 1,024 sample buffer, I have to keep the CPU usage under 10% in order to avoid clicks.

PeterKirn wrote:
I'd keep yelling at HP, too. Considered playing your press card to try to get them to take it back? If it does turn out to be the machine, you might consider which credit card you bought it on. Sometimes American Express will force a company to reverse charges on a non-functional product when the manufacturer refuses.


To be honest, I think the music technology industry is too small to impress them. They're going to take it back and reinitialize it. Maybe test it and find a problem, who knows? As far as an actual refund goes, the credit card charge went through in May.

After the reinitialize, I can sell it on eBay for a loss and buy ... well, I can buy something else that may or may not work.

--JA
View user's profile Send private message
PeterKirn
Posted: Wed Jun 21, 2006 9:38 pm Reply with quote
Site Admin Joined: 03 Feb 2006 Posts: 822 Location: New York, NY
Boy, that's bizarre. You should be getting exactly the opposite with the ASIO driver and MME/DX. Counterintuitive, but did you try *reducing* the sample buffer size?

Which host are you using? In a perfect world, you and I would know the answer, but I often find the people who do tech support all day get a larger sample size of problems and solutions.

I'd also be suspicious of some sort of bizarre power management setting. The Core Duo machines have had power management issues on PC (and some have argued, more subtly, on Mac with CPU throttling when connected to the AC adapter).
View user's profile Send private message Visit poster's website
midiguru
Posted: Wed Jun 21, 2006 11:10 pm Reply with quote
Joined: 21 Jun 2006 Posts: 28 Location: Northern California
PeterKirn wrote:
Boy, that's bizarre. You should be getting exactly the opposite with the ASIO driver and MME/DX. Counterintuitive, but did you try *reducing* the sample buffer size?

Which host are you using? In a perfect world, you and I would know the answer, but I often find the people who do tech support all day get a larger sample size of problems and solutions.


At 256 samples, the clicking is lots worse. I didn't try throttling it down to 128.

The host is Live 5.2. I haven't tried anything else ... maybe I should, before I box the machine up and send it home. I've set up a test song in Live that runs Absynth and Sytrus very quietly, so as to tax the CPU, along with FM7 playing one-operator sine waves at higher volume, so that the clicks and pops will be clearly audible. If I freeze the Absynth and Sytrus tracks, so that the CPU is only being used 10%, there are no clicks.

Power management is another amusing issue. The computer was causing significant buzzes and grinding noises through my speakers when I first started using it. I suspected a grounding problem in the computer chassis. An HP tech asked, "Does it make the noise while running on batteries?" Turned out it didn't, so he recommended using a ground lift adapter plug. That solved the noise floor problem, but I can't prove that lifting the ground didn't cause other power management issues that might affect the CPU.

Then there's the fact that the Windows Classic skin gets lost every time I power down. It goes back to XP style. I may have a lemon here.

--JA
View user's profile Send private message
PeterKirn
Posted: Thu Jun 22, 2006 9:32 am Reply with quote
Site Admin Joined: 03 Feb 2006 Posts: 822 Location: New York, NY
Maybe a lemon? Good grief. Hard to know where to start.

And yes, sounds now like there's an internal electrical problem and a botched Windows install on top of everything else.

Send it home to mama. Total bummer.
View user's profile Send private message Visit poster's website
rlainhart
Posted: Thu Jun 22, 2006 10:17 am Reply with quote
Joined: 07 Apr 2006 Posts: 146 Location: Rockland County, NY
midiguru wrote:
Then there's the fact that the Windows Classic skin gets lost every time I power down. It goes back to XP style. I may have a lemon here.--JA


That sounds like it could be either corrupted NVRAM or a dead motherboard battery. Not that you shouldn't send it back, but resetting the NVRAM or replacing the battery might help.

_________________
Richard Lainhart
http://www.otownmedia.com
http://www.downloadplatform.com/richard_lainhart
http://www.vimeo.com/rlainhart
http://www.youtube.com/rlainhart
http://www.airglowmusic.com
View user's profile Send private message Send e-mail
PeterKirn
Posted: Thu Jun 22, 2006 10:25 am Reply with quote
Site Admin Joined: 03 Feb 2006 Posts: 822 Location: New York, NY
Interesting . . . could NVRAM cause power management issues (thus creating these processor symptoms)?
View user's profile Send private message Visit poster's website
rlainhart
Posted: Thu Jun 22, 2006 3:17 pm Reply with quote
Joined: 07 Apr 2006 Posts: 146 Location: Rockland County, NY
It certainly seems possible, since some power management settings would likely be stored in NVRAM. More likely would be parameters like port status, which could be responsible for the FireWire problems he mentioned.

_________________
Richard Lainhart
http://www.otownmedia.com
http://www.downloadplatform.com/richard_lainhart
http://www.vimeo.com/rlainhart
http://www.youtube.com/rlainhart
http://www.airglowmusic.com
View user's profile Send private message Send e-mail
midiguru
Posted: Thu Jun 22, 2006 8:51 pm Reply with quote
Joined: 21 Jun 2006 Posts: 28 Location: Northern California
I have no idea what NVRAM is. I would think that a dead motherboard battery would trigger an alert, but perhaps I'm being naive.

The M-Audio support people advised reinstalling Windows, with the idea that this might cause the system to reallocate IRQs in a different (random) manner. Their other suggestion was that I could try using a USB hub, which aside from the awkwardness of having an extra piece of hardware at a gig seems like a suggestion born of desperation.

Talking to HP support, I learned that if I do a vanilla install of XP, I won't have a driver for the touchpad -- stuff like that. Not something I'm eager to delve into. The other alternative, which I have now taken, was to reinitialize the whole system from their Recovery Partition disk image. This restores all of the crap that comes with the computer, but at least I have all of the drivers.

The next step is to poke around and find out if the IRQs are any different this time. And maybe do some other stuff randomly differently. Like maybe this time I don't try to get rid of the wireless connection entirely, I just leave it disabled. Or leave the game software installed while testing the audio.

Bottom line, though: If it's actually possible to use a laptop to play softsynths, somebody is going to have to prove it to me, because I ain't gonna believe it until I see it myself. And if I've wasted $2,100 buying a laptop that won't do it, I'm gonna want to know why M-Audio doesn't have a big red flag on their home page saying DON'T BUY HEWLETT PACKARD LAPTOPS! Because if they had had such a flag, I wouldn't have wasted the money. But we're not there yet. Maybe it will all be fine this time....

--JA
View user's profile Send private message
Egz
Posted: Fri Jun 23, 2006 3:55 am Reply with quote
Joined: 26 May 2006 Posts: 10 Location: CA, USA
Bummer about the new laptop being less than perfect! Where does one start?

Speakers pop/crackle with system is a ground loop (despite ground lift setting on computer, I've never heard of that on a laptop... and no desktop in a decade! ) - Hardware needs to seat, and ground, factory issue....

So, despite how ASIO/DX hardware buffers also pop/crackle - when set too low, that is NOT the case if the system does it through speakers.... could other issues & quality issues be more common... sure, but why go on now?

I'm curious if a "shared" USB bus still exists, and if that will be an issue ( you know trackpad is also USB - try killing it see if it helps?)

I hope it gets sorted out - background processes might have confused the tech support people... nvram is video, Power Management Unit is power, um, yeah.... HP will straighten this out right?

( Get a Macbook though... Wink
View user's profile Send private message
garyg
Posted: Fri Jun 23, 2006 6:04 am Reply with quote
Joined: 11 Apr 2006 Posts: 109
If I wanted to use a PC (desk or laptop) for anything other than 'normal' everyday usage, the first thing I'd do is wipe it and do a clean Windows install. The skin issue sounds like a corrupt registry setting somewhere (I doubt it's the cmos battery...)

My budget Acer had so much crap installed (custom power management utilities etc.) that it runs/boots notably faster since I rebuilt it. The drivers were easy to find on the Acer site, I don't remember ever having trouble finding drivers for HP kit, their site was always really good for stuff like that.

.g
View user's profile Send private message
midiguru
Posted: Fri Jun 23, 2006 12:54 pm Reply with quote
Joined: 21 Jun 2006 Posts: 28 Location: Northern California
The saga continues....

Naturally, HP doesn't ship their laptops with a copy of Windows XP. They'll sell you one for $75, but as Gary pointed out, reinstalling Windows would mean reinstalling the drivers (for things like the touchpad) separately.

So I took the coward's way out. I used HP's provided Recovery Partition to start over from the factory disk image. This means reinstalling and re-authorizing a bunch of music software, but I figured if the registry had gotten corrupted, it was worth a try.

It's interesting. I seem to have found a sweet spot (power management set to Always On, USB ports set to never turn off to save power, a few other customized settings) that works better. At least, I can push the CPU past 50% using FL Studio without introducing pops and clicks. (The reason I'm testing with FL first, before installing Live, is because it has an easy authorization. I'm out of Live unlock codes until they reset it for me.)

Along the way, I found out some stuff about the Centrino Duo that others may be curious about. Like, does the extra core help make music software run faster. The answer: Maybe a little.

Until last night, when the little light bulb went on over my head, I was puzzled about two things. First, Task Manager's Performance tab has two little animated graphs under CPU Usage History. (Okay, I feel like an idiot for admitting that I've been staring at the graphs for a week without getting it.) Second, Task Manager's CPU Usage bar graph (a momentary meter) consistently shows lower values than the CPU usage meter in FL or Live.

The reason is because Task Manager's CPU Usage bar graph shows the value for the whole computer, while the meter in the application shows only the value for the core that's actually being used. The two history graphs show the values for the two cores separately.

It appears that when you run a CPU-intensive music app, most of the activity occurs in one core. The other is sort of idling. Thus a 2.16GHz Centrino Duo is NOT going to perform as well as a 3GHz Pentium 4. It baffles me why, a year and a half after I bought this desktop machine, I still can't get a laptop that performs as well as it does, but that's the way the cookie curdles.

However, the picture is not QUITE so bleak. I grabbed another easy-install program, u-he FilterscapeVA (great synth, totally awesome) and ran it as a VSTi inside of FL Studio. I threw a slew of notes at it to push the CPU. Observing the dual history graphs, I find that when Filterscape is generating notes, the second core is being used ... to some extent. The primary core also hits a higher usage level, but some of Filterscape's threads are being offloaded to the second core. That's good news.

I'm not sure what the point of having two cores is if they're not going to share the load. I mean, maybe if you want to watch a movie while Wavelab's batch processor munches on a bunch of files. Yeah, that's a reasonable scenario. But not common.

I haven't yet uninstalled the games and other crap that come from the factory. Before doing so, I'll set a system restore point, in case the uninstallation corrupts something.

God, this stuff is SOOOO musical.

--JA
View user's profile Send private message
PeterKirn
Posted: Fri Jun 23, 2006 5:42 pm Reply with quote
Site Admin Joined: 03 Feb 2006 Posts: 822 Location: New York, NY
Richard, NVRAM isn't called NVRAM on PCs, I don't think. I was looking for a similar fix for my Toshiba laptop's increasingly nonexistent battery life in the perhaps-vain hope that it's NOT the battery. (Buying a new battery is not my favorite way of testing that, either.) But I think NVRAM (which means Non-Volatile RAM, Jim) is a Mac thing. PCs may have the same thing with a different name, though . . .
View user's profile Send private message Visit poster's website
 
This forum is locked: you cannot post, reply to, or edit topics.   This topic is locked: you cannot edit posts or make replies.
Page 1 of 3    Goto page 1, 2, 3  Next
All times are GMT - 4 Hours
The time now is Tue Feb 09, 2010 12:00 pm
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum