How much of a difference will an audio card make?

[Smenelian]

I currently have onboard audio which has gotten me by so far without any trouble. That is until about the start of this year, when my CPU has decided that my audio projects are too much for it. I have a P4 3.2 GHz with 1 GB of RAM and normally live with about 46 ms of latency, which I know isn't considered good. How much would I benefit from an audio card?

[Prasa_U.]

I went from onboard audio to a M-Audio 24/96. I noticed some improvement but generally my upgrade in terms of computer specs caused a much bigger change. But it was worth it for the reduced latency. I didn't realize how much that affected how I worked.

[SnappleMan]

It'll work faster, sound better and cleaner, and give you professional inputs for recording and a spiffy internal mixer and all that spectacular stuff. It should be the first step of building your home based studio thing, if you're serious about music, at least.

Now get a good sound interface thing and start getting rejected from OCR!

[Villainelle]

What Prasa said.

You can use the ASIO4ALL driver with just about anything, including onboard sound chips, and get very reasonable latency. A nice audio card will be slightly better, but it's not a huge overall performance difference. What would make a big difference is a DSP card, which does some of the processing of plugins and takes that load off your CPU.

If you're recording any live sources, having a good sound card is important. It will affect your latency and quality while recording, and has a lot more inputs/outputs than onboard sound. If you do a lot of MIDI recording and your latency with ASIO4ALL isn't acceptable, then yeah, a nice audio card can cut it down a bit.

If you just want to be able to use more VSTs and effects, the biggest difference you could make would be upgrading your CPU, or adding a DSP card. After that, RAM. After that, sound card, fast hard drives, optimize your OS, etc.

[zircon]

DSP cards are cool, but expensive. If you're going to spend $700 on a Powercore MK II you might as well just save up for a Core 2 Duo processor... upgrading to one of those from a P4 like you have will let you run at least 4-6x as many plugins.

[SnappleMan]

Asio4all LOLOLOLOLOLOLOL

The Great Debate.

[Justus]

I used to think getting a good pro audio card didn't make much difference other than latency, efficiency, and having advanced routing options too.

Then I did some back-to-back testing between my generic onboard soundcard and my M-audio FW1814, and was actually rather shocked by the results!

I plugged the same synth into both cards one after the other, and made sure to set the recording levels to where it was hitting both at the same volume. The recording sounded significantly better when done using the M-audio! And that's recording line-level, so the quality of the mic pre doesn't have anything to do with it!

And when I say "significant", I mean it. It's difficult to tell if you don't literally try both back to back, but the pro card recorded the same source with a far more natural sound and more robust dynamics. For lack of a better term, it sounded more "analog", more real. There was also a noticable drop in background noise when listening at loud volumes. It's not just recording either. The same difference in quality can be observed just in playing sound back!

Here's a WAV I compiled real quick to demonstrate: I played my Ys 3 CD in my CD player, and recorded results from the same passage on both my onboard sound and my FW1814. The first half is the onboard, and the second is the 1814.

http://www.justusjohnston.com/temp/compare.wav

You can hear that during the second half, the strings sound warmer, the drums are fuller and punchier, the bass is better defined, and....well, the whole thing is just better.

What's really funny is that M-audio makes about the cheapest professional audio interfaces money can buy, and they're also...well, I hesitate to use the word "crappiest", because they really truly do give good bang for your buck if you're looking to "go pro" on a budget, and I definitely recommend them as an entry level professional interface....but I'll put it this way: when I did the same back-to-back test comparing this M-audio FW1814 to an RME Hammerfall, the difference was just as striking as the one between the M-audio and the onboard!

Why is there such a difference? The quality of the electronics for one. Even if the ultimate goal is a digital recording, there's still plenty of internal analog signal path involved. This will also have a significant effect on the device's self-noise. When recording or playing back audio on either my onboard or FW1814, the onboard has audibly louder noise when not routing audio, and by about 10 decibels at that, according to sound forge's meters (this means it literally makes twice as much noise for the same amount of healthy signal). Admittedly, part of that is because the M-audio is an external firewire device located on a grounded, noise-isolated rack, and plugged into a power conditioner, whereas the onboard is a series of connections and DSP chips inside a computer surrounded by fans and attached to other computer bits with power flowing through everything, but still. The other reason there's a difference is the quality of the converters, the interface's ability to change analog to digital and vice-versa. Better converters mean that the device generates a more accurate wordclock for more accurate analogesque performance and superior dynamics.

So the big question: What does this mean for you? If you're a normal guy who just likes playing games/watching movies/listening to music, then not much unless you're a perfectionist. Even though the playback quality on a pro interface is audibly better, most people (including myself, I'll admit it), grow accustomed to minor quality differences over time, so the experience itself is not significantly enhanced.

If you're into recording and mixing on the other hand, it makes a WORLD of difference. First, consider monitoring a mix. Even the most minor differences in listening back will have a profound effect on mixing. You'll use different levels of compression, different EQ settings, different reverbs, etc. The noise level is also a big deal. Trace amounts of on-track noise that you'd hear and fix using a good clean interface would be masked by the self-noise of the cheaper one. If you use lots of outboard gear, then the quality of the converters come into play in a big way! It may seem like we're being nit-picky about the quality of signal that comes in, but consider this: a single sound in a typical mix goes through a good half-dozen or more processes before the final master, each of which further digitizes the sound (a snare drum, for instance, which might have a meter change, EQ, compressor, reverb, sub-grouped with the other drums for another meter change, EQ, compressor, then yet another meter change at the master fader, with yet another EQ, compressor, and maybe even a peak limiter). Each of those processes will exacerbate any flaws in the original recording. The effect is further compounded when you consider that there may be several dozen such tracks in a mix!

On the other hand, if you only produce electronic music using software on a computer, then you can throw out everything I just said except for about monitoring. But then there's always the efficiency (how fast it plays with the rest of your machine), and latency. At 46 ms, you must not be doing any real time MIDI sequencing or you'd probably find that quite unacceptable unless you're monitoring direct from an outboard synth's outputs. It's also no good for recording live performers with a monitor mix.

[Villainelle]

Justus...welcome to the forum; let me point out a few things.

1) Yes, a good sound card is very important if you're recording anything that originates from OUTSIDE of your computer, like guitar, voice, a hardware synth, whatever...but is basically irrelevant for VSTs (except for how much latency you can tolerate), for example.

2) Part of what makes a good recording is how ACCURATE it is to the actual sound recorded. While a "warm" quality is usually desireable for a final mix, if it's introduced during the recording process, it's called sound coloration, and it's not supposed to happen. Even if it sounds good, coloration is an alteration of the clean source sound, something that should be done with great care either deliberately during the recording with e.g. an effects box or hardware compressor etc., or added as an effect once the audio is in your DAW, so that you can control it or avoid it altogether if you prefer.

3) Regarding your example WAV...it doesn't sound "warmer" to me; what it does sound like is that the second recording is clearer and has more detail - in fact, very much like it was recorded at a higher sample rate than the first, and then both recordings were sampled down to 44.1khz in your DAW or WAV editor. I noticed you didn't mention the sampling rate, so I suggest you run the test again and make sure both sound cards record at the same rate, volume, etc. Then take both WAVs into your editor, lay one on top of the other, and invert one. They should cancel out.

4) If all sound cards recorded or rendered audio differently with their own special sound coloration, we wouldn't be able to do the above phase inversion test across computers - your card would add a little something whenever you rendered, mine would add something different. Eventually you'd end up with something like what happens to a JPEG or an MP3 that gets compressed over and over - complete noise.

Your specific audio interface can record at up to 192khz - I have an M-Audio Audiophile 192 that can as well. I could do your WAV test again and get you similar results - recording at 192khz will sound amazingly pristine in comparison to murky old 44.1.

Hope this has been informative.

[Smenelian]

Aside from a midi keyboard that I will get when I have the space for one, I do everything with software. I've done the occasional playing with the typing keyboard, but the latency kills the already unsatisfying experience. So I guess that means I'll need to be looking into a new CPU first. What I really need is a completely new computer. The one I have now is starting to become old and isn't going to be worth upgrading with all the new technology that it doesn't support. Such is the life of those who like to keep up with the times.

I tried out Asio4all and it nearly tripled my CPU usage, though I could decrease my latency a bit. Windows blue screened me about 3 times with it installed too, and after uninstalling the blue screens stopped.

I'm not hearing a significant difference in that sample, Justus. It may be my headphones (Sony MDR 7506), or maybe I just don't know what I'm listening for.

[Justus]

I'm not hearing a significant difference in that sample, Justus. It may be my headphones (Sony MDR 7506), or maybe I just don't know what I'm listening for.

Those are pretty sweet headphones. We've got a couple pairs of those at one of the places I work. It's been awhile since I've used headphones for serious listening though, so I couldn't say for sure if that would have an effect. I will say that it's not necessarily a "significant" difference, but is certainly a noticeable one.

However, being that you do in fact work with all software (a keyboard is part of the MIDI signal path, not the audio path, so it doesn't count), it shouldn't be that big a deal anyways. You are in the circumstance of using software and sequencing by step, so a pro audio interface doesn't really bring much to the table until you're using outboard gear, microphones, and/or realtime sequencing. You basically get the benefit of slightly cleaner monitoring, and if you can't tell the difference between the two halves of that comparison, then that obviously isn't a big issue anyways at this point of the game for you. Once you get that keyboard though, you may want something that gives you <20ms latency. I trust you've already tried playing around with the audio buffer size, right?

So what changed about your setup to make the CPU start crapping on you all of a sudden anyways? Also, what kind of software are we talking anyways? A 3.2 P4 and 1 GB of RAM should be quite sufficient to play around with Fruity Loops, but may get you in trouble if you're trying to do a bunch of Kontakt stuff in Sonar 4.

Justus...welcome to the forum

Thanks. I figured I should give it a try since one of my remixes is getting posted, even if it is just some BS I came up with on my violin one evening. I hope this community is chill.

1) Yes, a good sound card is very important if you're recording anything that originates from OUTSIDE of your computer, like guitar, voice, a hardware synth, whatever...but is basically irrelevant for VSTs (except for how much latency you can tolerate)

I think I can agree with that, in fact..

On the other hand, if you only produce electronic music using software on a computer, then you can throw out everything I just said except for about monitoring.

Yup.

Part of what makes a good recording is how ACCURATE it is to the actual sound recorded. While a "warm" quality is usually desireable for a final mix, if it's introduced during the recording process, it's called sound coloration, and it's not supposed to happen.

Sorry if it was confusing. You'll have to forgive me for using such a subjective buzzword. In this situation I meant the word to mean "present and intimate", rather than any specific harmonic characteristic. Both my M-audio and my onboard add no (or at least identical) coloration in the harmonic balance, as my spectrum analyzer responds to both identically. I don't think the vast majority of interfaces (pro or not) do, though I do remember hearing awhile back about some Creative soundcard with a tube in the output signal path. We're not comparing Akai samplers after all.

Yet I would have to argue with having no use for coloration during recording. At a place I work (the same place), we use API pres to track vocals, which are good clean pres, but we'll often switch to an Avalon 737, which is a very colorful tube pre, and for certain vocals, it's like putting a silk tie on. Even running a synthesizer through a colorful DI can do wonders for having it stand out (if that's what's needed), and can breath some great life into an otherwise sterile sound. It depends on what you're recording. Too many people go the opposite extreme too. I'd say "it's not supposed to happen" is a more accurate statement when talking about monitoring than recording, though I'll certainly admit that one should avoid unmeasured coloration for sure regardless.

very much like it was recorded at a higher sample rate than the first, and then both recordings were sampled down to 44.1khz in your DAW or WAV editor.

No, I didn't do that.

I noticed you didn't mention the sampling rate

44.1khz, 16 bits per word on both. No resampling, no dithering. The WAV file is representative of the raw work of the original wordclock in both instances, and the comparison is purely that of one A/D converter and internal analog routing vs. another under identical load.

so I suggest you run the test again and make sure both sound cards record at the same rate, volume, etc. Then take both WAVs into your editor, lay one on top of the other, and invert one. They should cancel out.

I'm hitting record by hand here, so it would be time-consuming to line up both recordings so that they are timed sample-perfect with each other. Even if I did, I'm sure a good portion of it would cancel, but it would not be a perfect 100% cancellation anyways since the actual content does in fact have differences. As for making sure they record at the same sample rate, I mean...how could I not? That would really defeat the purpose of the experiment. The whole point was to leave every variable identical and simply use different interfaces.

4) If all sound cards recorded or rendered audio differently with their own special sound coloration, we wouldn't be able to do the above phase inversion test across computers - your card would add a little something whenever you rendered, mine would add something different. Eventually you'd end up with something like what happens to a JPEG or an MP3 that gets compressed over and over - complete noise.

I'm not sure what you're getting at here. If you recompress audio to MP3 continuously it does indeed become less coherent. I have no idea what that has to do with comparing phase inversion tests of an uncompressed WAV file on multiple computers, or why we'd even want to do that in the first place.

Your specific audio interface can record at up to 192khz - I have an M-Audio Audiophile 192 that can as well. I could do your WAV test again and get you similar results - recording at 192khz will sound amazingly pristine in comparison to murky old 44.1.

And if I actually had recorded at 192 and dithered down, it would have been even MORE pristine (with the addition of a negligable amount of dithering noise of course).

[Villainelle]

the comparison is purely that of one A/D converter and internal analog routing vs. another under identical load.

Is the CD drive using an analog connection to the audio card? Obviously any analog element introduced into the signal chain is going to introduce imperfections, however pleasant they may be. A purely digital signal chain would not.

The analog circuitry of a given sound card has nothing to do with digital information. Which again, is why the sound of a software synth will be exactly the same with either sound chip, as it shouldn't pass through any analog routing. But if a CD was ripped from a drive with an analog connection to the sound card, there would certainly be some coloration. Perhaps this explains why your recordings differ?

I hope this doesn't come across as antagonistic. I'm just curious about how you set up this test. If all conditions are identical, a purely digital signal chain will give identical copies of a sound. So something is amiss here.

The MP3/JPEG artifact thing is an analogy. Sorry it was unclear, I was illustrating what happens when inaccuracy is introduced while working with source material. Here's a different example.

Take any sound.

Play that sound through any signal chain that contains some analog component.

Record the sound after it exits the chain.

Do this repeatedly.

The analog inaccuracies introduced will gradually accumulate until the sound becomes noise.

Even if it only happens once, it's no longer the same as the source and can't cancel out with it.

If one person does this while another copies the sound through a wholly digital signal chain, the resultant files will not cancel out upon phase inversion.

If both pass through purely digital signal chains, they will remain identical and cancel out, regardless of the sound chips used.

I hope this is clearer now.

And just some food for thought--I'm not sure what other sites you read or participate in, but I lurk around some like KVR, NorthernSounds, Gearslutz etc. These are frequented by audiophiles and professionals of all sorts. Topics like "DAW X sounds better than DAW Y" pop up all the time, and have been conclusively proven wrong. But I don't see many claims about sound cards sounding better than one another. If there was a real difference in the digital sound processing amongst sound cards, you can be sure these audiophiles and professionals would be all over it, even if it was negligible. Everyone would want the best, hobbyists would do tests, people would recommend Sound Card X's warmer sound over Sound Card Y's colder sound, etc. Product marketing would follow suit. But none of this is the case. There are endless battles over every type of analog component, but digital components are...digital. All that matters is whether they have the range of features you need.

Congrats on your mix being accepted, btw.

[zircon]

Actually there is an advantage to upgrading your interface, even if you do work entirely "in the box"; that advantage would be superior ASIO drivers. I used to use an EMU 0404 PCI sound card. With it's onboard EMU ASIO drivers, I was able to get more out of my CPU (read: more plugins at the same latencies without crackling) as compared with asio4all and non-ASIO devices on my system. Then, when I upgraded to a Presonus Firebox, I got even better results. In the end I think it was a very worthwhile investment.

[Justus]

Is the CD drive using an analog connection to the audio card?

OK, I think I see where you're coming from. I didn't see it because I don't think you understand what I'm testing. You're talking about digitally transferring audio, and not any kind of actual recording. Even hooking a lightpipe from my CD player to my interface and hitting record would just be transferring literally identical bits without any transformation, and proves nothing. The phase test would still be a bitch unless I used a sync to record both signals at the same mark. For that matter, my onboard sound doesn't have any digital audio inputs anyways. Even if it did, comparing results would be like trying to compare how accurately two different hard drives could copy a file.

That wasn't the point of my test. The point, as I've said, is to compare the A/D and D/A converters and the internal analog circuitry of the interfaces, and hence how well they treat microphone signals and analog outboard gear (an important part of most people's music work). As I've said a few times now, it won't have any effect on software synths except for the monitoring, since the monitoring is the only part of the analog signal path.

I don't know about recording it 192 DIGITALLY and dithering it down. I think that'd only have an effect on things recorded analog. In fact, wouldn't the wordclock mismatch cause the audio to play back 4.3 times too fast until I resampled it down? And even then, it would sound identical, and could indeed be perfectly phased. I suppose I could upsample it to 192 with interpolation and then dither it back down, but I think that would accomplish little other than the introduction of dithering noise. Of course this no longer relates to the OP, so the discussion is purely academic at this point.

I hope this doesn't come across as antagonistic

Likewise. Too many people on the internet seem likely to get up in arms about stupid BS. I know you're not being antagonistic anyways. You just don't realize that I know what I'm talking about yet. Whether or not you're one of these internet guys that will post ANYTHING to avoid admitting that someone else knows their stuff remains to be seen of course.

And just some food for thought--I'm not sure what other sites you read or participate in, but I lurk around some like KVR, NorthernSounds, Gearslutz etc.

Not many anymore, I will admit. I do most of my shop talk in the meat space with other industry professionals. I don't really care if two dozen people think their pirated version of Tracktion sounds better than their pirated version of Sonar and that anyone who uses Cubase is an idiot who couldn't possibly have a different approach to writing music and different needs.

Congrats on your mix being accepted, btw.

Thanks. I submitted it back in November, and it was added to the "to be posted" list in January. I didn't realize the turnaround time was like this. I'm really curious to see the judges' comments. I honestly didn't think it would be accepted. I thought for sure they'd say it was too short, could've done more with the arrangement, and could've been better performed. I guess maybe the combination of it being an uncommon game and an uncommon arrangement for the site was enough.

Actually there is an advantage to upgrading your interface, even if you do work entirely "in the box"; that advantage would be superior ASIO drivers. I used to use an EMU 0404 PCI sound card. With it's onboard EMU ASIO drivers, I was able to get more out of my CPU (read: more plugins at the same latencies without crackling) as compared with asio4all and non-ASIO devices on my system. Then, when I upgraded to a Presonus Firebox, I got even better results. In the end I think it was a very worthwhile investment.

Yup, this is true. It's nice that ASIO4all exists, but ultimately, using it with a consumer-level interface is similar to playing a Direct X 9 video game with a video card that only has DX8 support. It works, but not as efficiently. In the pro audio world, you usually really DO get what you pay for, and there's no simple way to completely outsmart the sound equipment industry.

Zircon, out of curiosity, have you ever seen this? I'm dreadfully curious about it. I've thought about buying their Solo driver to see if it makes a difference (it's supposed to work with my FW1814). And they have a driver that's specifically for the Firebox, too. I get OK latency as is, about 4ms, but I usually have to bump it up once I start loading on dozens of audio tracks, and there are a couple of things I find obnoxious about M-audio's drivers anyways. Besides, it would fascinating to try chaining a second interface for more ports. I'd love to buy a cheaper RME firewire interface and use it as master wordclock while still being able to use all 18 M-audio inputs for extra I/O.

[analoq]

Is the CD drive using an analog connection to the audio card? Obviously any analog element introduced into the signal chain is going to introduce imperfections, however pleasant they may be. A purely digital signal chain would not.

he said CD player not CD drive. it's clear he's just using an analog line-level source (external CD player in this case) to do his test.

fake edit: nevermind, he clarified while i was writing this.

And if I actually had recorded at 192 and dithered down, it would have been even MORE pristine (with the addition of a negligable amount of dithering noise of course).

doubtful. dithering only applies to bit-rate, not sample-rate. and resampling algorithms aren't going to do much for you.

downsampling from 192 to 48 would be neutral, but going from 192 to 44.1 would give you a marginally worse sound than if you had just recorded at 44.1 in the first place since it's not evenly divisable.

downsampling from 176.4 to 44.1 might be better than just recording from 44.1 if you had a crappy ADC. a good ADC would give you the most accurate samples possible at any rate.

[zircon]

Zircon, out of curiosity, have you ever seen this? I'm dreadfully curious about it. I've thought about buying their Solo driver to see if it makes a difference (it's supposed to work with my FW1814). And they have a driver that's specifically for the Firebox, too. I get OK latency as is, about 4ms, but I usually have to bump it up once I start loading on dozens of audio tracks, and there are a couple of things I find obnoxious about M-audio's drivers anyways. Besides, it would fascinating to try chaining a second interface for more ports. I'd love to buy a cheaper RME firewire interface and use it as master wordclock while still being able to use all 18 M-audio inputs for extra I/O.

Nope, I haven't seen or used that... sorry!

[Justus]

doubtful. dithering only applies to bit-rate, not sample-rate. and resampling algorithms aren't going to do much for you.

Yup, you're right! I've gotten into the nasty habit of referring to any process of interpolating digital resolution downwards as "dithering", when it is in fact just the addition of strategic noise before re-quantization in order to avoid quantization error.

downsampling from 192 to 48 would be neutral, but going from 192 to 44.1 would give you a marginally worse sound than if you had just recorded at 44.1 in the first place since it's not evenly divisable.

And that is a debate I'd really rather not get into. We don't need to fill this up with pages-long arguments about oversampling and mixing benefits. You can find enough of that at gearslutz or homerecording.com

[analoq]

We don't need to fill this up with pages-long arguments about oversampling and mixing benefits.

i tend not to make such contributions, but given the amount of discussion in this thread that is not really pertinent or helpful to the OP there is no harm. par for the course around these parts.

besides, mixing benefits aren't what i'm talking about. that's a different debate.

You can find enough of that at gearslutz or homerecording.com

i think you are like me in that i don't look to internet forums for this sort of knowledge.

i'll stick with the trade publications and trade shows, thanks.

[SnappleMan]

So when will you people cut out the technical nonsense and start making some *music?

*By music I mean music that doesn't suck.

[Villainelle]

Likewise. Too many people on the internet seem likely to get up in arms about stupid BS. I know you're not being antagonistic anyways. You just don't realize that I know what I'm talking about yet. Whether or not you're one of these internet guys that will post ANYTHING to avoid admitting that someone else knows their stuff remains to be seen of course.

Yes, you seem to know what you're talking about and have more knowledge about it than I do. I don't have a problem admitting that. The reason I replied was that your first post in this thread came off as misleading, qualified with meaningless buzzwords--no offense, you admitted that yourself. Now that you've clarified it I understand the conditions of your test and so should anyone else, which is great.

It does seem self-evident however, and I'm not sure why it needed to be said in the first place--the original post in this thread is talking about CPU load and latency, not recording quality. So an audio card records an analog line-level source better than an onboard sound chip? In other words, as stated previously and more simply...you'll get a better live recording of anything with an audio card as opposed to onboard sound? ...duh? Did we need all this to come to that conclusion again? The Ys CD example just seems misleading in this context.

Since you like to speculate about the psychological motivations of people on message boards, let me return the favor--you seem like someone who knows a bit about sound tech and likes to flaunt it, unsolicited, on message boards; who makes provocative yet ambiguous qualitative claims without full disclosure of all relevant data, so that you can respond patronizingly to those who, in their confusion and relative ignorance, take issue with you. You do in fact remind me a lot of an ex-member of this forum, Compyfox.

Anyway, done with this thread. I hope the OP got whatever answer he was looking for.

[analoq]

So when will you people cut out the technical nonsense and start making some *music?

my technobabblery output is miniscule compared to my musical output.

*By music I mean music that doesn't suck.

oh. well in that case probably never.

my musical output is miniscule compared to my self-deprecational output.

[SnappleMan]

Dramarama.

At least he doesn't pimp Asio4All :DDDDDDDDD J/K LOOLOLOLOL

And Anal-oq, good sir, I have been without tunes from you for over 2 years. Granted, I rarely check OCR for new music, nor do I visit your website for music, but give me a break, I'm lazy.

[Yoozer]

Topics like "DAW X sounds better than DAW Y" pop up all the time, and have been conclusively proven wrong.

But that's due to the used summing engine and its algorithms, which are often at a much higher resolution.

But I don't see many claims about sound cards sounding better than one another.

Ask on Gearslutz about the difference between an Apogee Rosetta and an E-mu 0404 .

Both the way in (ADC) and the way out (DAC) call for converters of a high quality, low noise and low jitter. Once it's in the machine however, there's not much happening; but it's the conversion step that does the trick.

Plus, there's the phenomenon that an on-board soundcard's input suffers from the electronic noise of the motherboard and the other components.

You do in fact remind me a lot of an ex-member of this forum, Compyfox.

I can assure you that Justus is not Compyfox .

What's with the "unwanted" anyway? The topic wasn't resurrected from the dead, the response was on topic and in time, and it underscores the importance of having a reasonably decent audio interface; if not for the latency, then at least for the better quality of I/O (and having more of it if the project calls for it).

[SnappleMan]

This intense PWNing of Leah is making me sad.

[Smenelian]

Well, this thread has gotten quite a bit of attention...

So what changed about your setup to make the CPU start crapping on you all of a sudden anyways? Also, what kind of software are we talking anyways? A 3.2 P4 and 1 GB of RAM should be quite sufficient to play around with Fruity Loops, but may get you in trouble if you're trying to do a bunch of Kontakt stuff in Sonar 4.

Nothing has changed except I keep loading more and more things into my projects. I use FL as the host and usually have some random VST instruments thrown in, like EWQLSO Silver, Plugsound, Sytrus, and a couple others. I use DirectWave for a sampler since it works pretty well and it came with the XXL version of FL anyway.

On one of my newer songs, I've got about 35 channels, about half of which are VSTs with 1-5 effects in the mixer or on the plugin itself. My CPU usage is about 30% and there is a lot of crackling during some parts. How much more performance can I expect from something like a Core 2 Duo? Let's say one that runs at around 2.8 GHz.

[zircon]

Hmm... if your CPU usage is at 30% and you're getting crackling, that sounds like a BIG driver problem. You should be able to push to 90%+ before you even start. So I suggest upgrading your interface *first* (eg. don't use onboard audio) - perhaps to an EMU 0404 PCI card. Then you'll also want to take a look at your disk streaming settings and hard drive speed. Even a fast computer w/ good drivers might choke on plugins that rely on disk streaming (QLSO might, depending on configuration). For example, dense QLSO Gold setups in FL7 for me, even at <20% CPU usage with good drivers, can cause some crackling for me because of heavy disk streaming.

Anyway, I want to emphasize... it sounds like you are doing OK with your current CPU because your % usage is low. No need to upgrade that yet. Just get a new interface.