tracks with no sound ? r they the same as tracks with ?

Spider

New member
hi all , i know what your all thinking , huh???????? but all i was wondering was if i record say 30secs of nothing (by this i mean have everything hooked up , just mute the mixer) is it the same (cpu usage, that sort of thing) as if i recorded sound on the track?? I heard some where it was the same , mainly heard about how muting in between where the vocals are in a vocal track made no difference at all. So this means the size of the wav file would be the same to right?? well i hope im not to far of the mark there ... But the reason i ask is i did a few little adjustments (heehee) to my pc and i wanted to put it under some pressure to see how it would hold up. So in increments of 8 tracks(thats all my card allows simultaneously) i recorded 32 tracks into cubase .... these tracks were fed from the direct outs on my mixer which were all muted resulting in a physical audio file being created , just no sound in it. I was able to add 7 plugins (comp, couple of reverbs , waves eq's) and the cpu usage in the performance section was still reading only half.... Am i getting excited over nothing here or am i right in saying these no sound tracks will give a basic idea of trackage ?? (trackage.!!!!! dragon , add that to the bbs dictionary) I really dont need 32 tracks , so i would replace the 16 tracks with plugins if this is true.. Well thanks all ,
spider
 
I wouldn't be sure

The experiment is more meaningful if you have an ordinary .wav file (one with a signal, just as easy to come by as one with silence) copied to multiple files each with a different name, then load each one into your multitracker. No telling how the lack of signal might affect the CPU usage for any of the plug-ins.
 
A silent audio track will be the same size on the drive as any other track, all other things being equal (length, sample rate, bits, etc.). If you're sending a muted signal from your mixer to the sound card, you probably won't end up with true silence. The sound card is going to generate some amount of internal noise. As far as "naked" tracks go, they don't suck up much CPU time at all.

EQ plug-ins usually have an even CPU load, regardless of the audio content. Compressors, on the other hand, have to work harder depending on the level of the signal. Both will be affected by the plug-in settings you use, as well. It's really tough to do accurate plug-in benchmarks. You'll find out what you can support through experimentation. But it sounds like ya got a prety good setup.
 
so pg , doesnt it matter if i have a 3min song saved as a wav in comparison to a 3min blank file?? Thats pretty trippy... Cos i would think its like a pic ,the bigger and more content it has the larger the saved file will be..
But ill definetly try it with some real audio...

Ive had probs with it for ages , well not major probs , just wasnt right ...u know? Then ive been trying to tweak it up a bit with settings ..it seems to of done something.. Though when i say i had 32 tracks and 7 plugins on , i really didnt feel confident that the system was going to hold up.. Lets say i wouldnt trust it with 32 on someones work, but i think 24 or something should be quite stable..
I found this book at the library the other day , although it seems to be all about 'how to setup your first pc for racording' it really does give out some great info .. its called making music with digital audio.. Thing is it seems a little old , about three years i think so there using really crappy pcs and win95... but thats kinda good cos u see how they configure the test pc they use in the book and then what there doing with it , it gives you some comfort thinking that yours is like twenty times as fast if not more and the drives a billion times bigger , well not that big...
Anyone know of any other good books on setting up your pc for recording?? well cheers all, ill let u all know how the test goes , try it tonight maybe...cheers
spider
 
It's the same as with a picture. “The bigger and the more contents, the larger the file size.” Number of pixels compares to number of samples and the bit depth works in the same way in both formats. The file size of an uncompressed picture is independent of whether you have just one colour in the picture or use the entire palette. It’s how many colours you can use that matters.

Picture file size = Number of pixels X bit depth
The number of pictures is of course the height X width

Sound file size = Number of samples X bit depth
The number of samples is of course sample rate X time

Pretty similar huh?:)

Correct me if I’m wrong though

/Ola
 
sounds pretty right to me ola... when u say 'can' , does this mean the file sizes , so long as they were all recorded at 16 bit then they should be the same size? is that right.. I ended up trying full songs , just changed the names like the doc said , i was using 6 stereo tracks (is this the same as 12mono tracks?) and had 13 plugins open , mainly reverbs and compressors..It was using about a third cpuaccording to the meter..it seemed rock solid , just ran out of time to add anymore tracks.. i had the performance window open while turning the plugins on and off .. its amazing to see how much more cpu usage some plugins use. I found , in cubase anyway, the steinberg (resident so to speak) plugins seem the most solid , is this cos they were able to design them specifically for steinbergs programs? think i really just need to try recording a large sesion and see where she breaks , if at all i hope , heehee ..... well cheers all....
 
If they're recorded at the same bit depth and sample rate and are equal in lenght, then they're the same. However, not all 16 bit files are the same size even if they are the same lenght.

6 stereo = 12 mono - I would be surprised if it wasn't but I'm not 100% sure. Sorry. Maybe some plugins require less power on one stereo than on two mono though. Just a gut feeling though.

As for the limit, I think that disk speed is the bottle neck in almost all DAWs. As soon as you start deleting tracks and recording new etc., your disk gets fragmented and that's a real performance killer.
 
forget about bits for a second ...

Digital audio exists as 1s and 0s regardless if there is an actual signal or not.

If you have a 30sec .wav file of the London Symphony Orchestra ... it will be the same size as a 30sec .wav of absolute silence (assuming all that the samnple rate, bit depth, etc etc etc are the same). The size of the .wav is not governed by the content (content meaning the complexity of the audio signal ... not the # of samples). Also, even if it is a null signal ... your CPU still has to process that information the same as if it were the William Tell Overture.

An attempt at an analogy could be ... if you record silence to analog tape ... does the tape still roll during those sections? Yep ... it does ... so the capstan and rollers and head still need to work just as hard to translate whatever is on tape through the output to your monitors ... even if it's nothing but hiss. :)

Please tell me I didn't make that more complicated! But I get the feeling I did! ... :D
 
As for the limit, I think that disk speed is the bottle neck in almost all DAWs

This is the only place where I disagree with ya, Ola. I was able to support at least 48 tracks (16/44.1) with no effects on a single 7200RPM drive. A seperate "dedicated" audio drive formatted with the largest possible block size should perform even better in my situation. Plug-ins have always been my bottle-neck in projects, without exception.
 
hey cheers for that everyone , and no it wasnt made more complicated , understand it a lot better now.. Well i heard somewhere that cpu speed governed the amount of plugins u could run and harddrives size and speed governed the amount of tracks , does this sound right??
Hey pg , im glad u mentioned block size ... I kinda understand what this is about but not really much about how to change it so it makes an improvement .. A book i was reading the other day mentioned this , but they were talking about using a 3 gig drive with 95 on an old pentium mmx.... Tell me if this is right regarding blocksize..
When we format the drive is divided into blocks or clusters with FAT (file allocation table)telling the cpu what is where ... So lets say a 20gig harddrive is formatted in fat32 .. Is this right , as a default it will make the blocks roughly about 4mb.. This appears to be great for saving space on the drive, cos from what ive heard if your blocks were 10mb each and u only wanted to save a file worth 1 meg , it would still have to take up the full block size of 10meg.. But the impression i got from this book was that increasing the block size (so long as u could afford the space) helps digital audio quite a bit.. Any ideas on this and how would i change it if it would help?? im using fat32 at the moment , so i would prsume there quite small at the moment ... please someone correct me if i was way wrong with all that .... cheers all
spider
 
You've got most of it right, except the actual sizes. I think the default block size is 4k. It's best for audio to use the largest block size. That does mean that more space might get wasted, since even one byte used in a block will cause the whole block to be marked as used (wasting all but that one byte). Like everything, it's a tradeoff. Audio deals with a few huge files in contrast to most applications which utilize a gazillion little files, so you're not likely to lose all that much space even with a large block size. I can't advise ya on the actual process of partitioning or formatting... I leave that to friends that work with those tools on a regular basis. I'm not a hardware guy, it's a messy area to me :).
 
pglewis - I only use a wee bit of plugins so CPU has never been a problem for me. I did however hit the limit on my old DAW - 5 or 6 tracks of 24/96! It was a PIII 5000MHz with a dedicated 7200rpm drive. I couldn't get it to do more without messing up the recorded tracks. It turned out that my disk was so fragmented, it couldn't handle more. I started defragmenting regularly (daily) and never hit a limit again. I think we can easily agree on that the limit is all up to the recording situation.

Spider - Search for "partition" and "block size" etc. and you'll get a bunch of info on that as well. I did a fair bit of benchmarking on my disks as I had one IDE and one SCSI and wanted settle the eternal question. It turned out that the only thing that made any significant difference was switching bus mastering on or off on the IDE disk (search for that as well). Bus mastering rocks I can tell you. When both disks performed at their best, there was virtually no difference and the theoretical number of tracks they would support was out of this world anyway.
 
Ahhh... I didn't know you were pushing 24/96. I don't even consider trying to go 96k given the space requirements :(. How much did that 5000MHz! processor run ya :D?
 
It costed 7 gazillion dollars. It should of course have been 500GHz:rolleyes:. Also, now you know why processor power never was a problem...

Cheers

/Ola
 
all i can say is "I WANT ONE" gimme gimme.....
Sometimes it seems thats what u would need to hold up a large session. I couldnt imagine running 128 tracks on a pc.. Maybe one of those new 1gig Hz pcs might be in order. Anyone heard how they go on a daw.??
Well cheers for all that everyone , ill definetly have a look at blocksizes. By bus mastering , i take it your talking about DMA?? cos i have that checked , besides my soundcard just wont operate without it. So does everyone think scsi makes a big difference??
Well it all seems pretty stable at the moment ive been using it for 24 tracks just yesterday and it seems great. well cheers everyone....
hey ola u wanna post a link to the 5000MHZ pc......dieing to see it..no wonder u have a cruising time with plugins!lol...
spider
 
WOW!!!!!!
thats one mighty pc.... i thought mine was choice ... that thing rocks man!!! about the program , i think my comodore 64 came out standard with that painted on the screen .. took ages to scrape off.....
 
Nice thread, cool info. I always knew that a track of silence took up the same space as a track of music. However I still suspect that the cpu might have to work harder to digitally record a track of music vs. a track of silence. Any more thoughts on this?

In the FAT 16 & 32 file system, your default cluster sizes are:

Cluster size FAT 16 FAT 32

4K 0-256 megs 0- 8 gigs
8K 256-512 megs 8-16 gigs
16K 512-1024 megs 16+ gigs
32K 1-2 gigs

I don't know off hand what cluster sizes are used in NTFS. Easy to see why FAT32 was such a big improvement.

In Win98 SE (second edition) they made some performance tweeks in FAT32 (same cluster sizes but more speed) which means if you are using Win95 in a DAW, look to upgrade to Win98SE or WinME (and plan to reformat your drive)

Don't forget RAM, you can have a fast cpu and hard drive but if you are minimal on RAM for buffering, it's going to be a BIG SORRY....

[Edited by RWhite on 09-21-2000 at 22:58]
 
hi Rwhite, so u know if the cluster sizes can be changed , for an improvement that is? What about playing back tracks with no sound v ones with ? surely the cpu would have to process the sound in some way compared to it not having to process any sound.. My soundcard has a couple of dsp chips in it (as a lot do) .. would this (supposedly) mean during sound procesion no cpu is used , for playback at least anyway?? how important is ram anyway ? ive always thought of it as quite important , thus the reason i installed 256meg , should more help me or wont it matter ? some people i hear say it doesnt make much difference.. though i look at some apple macs and there running like 1.5gig of ram !!!!!! WOW!!!! thats pretty huge, but mind u it only costs you an extra $3000(AUS) (about$1500US)...that woulod want to make a killer difference....
hey white , your right , some good info going round here , hope its helping everyone else as much as its helping me , cheers all.....
spider
 
You can change cluster size. I've been told there is a way to do it through FDISK but I've never run accross this. The way I would do it would be via a program called Partition Magic 5 - great program by the way, lets you create, delete & resize disk partitions on the fly without reformating or losing data. I've never had it crash on me although one guy I know swears it hosed his hard disk. Interestingly it warns you not to create 64K clusters under Win95/98, to only do it under NT. But lets you do it anyways I think (haven't tried it)

The idea behind cluster size is that the bigger the clusters are the fewer of them there are, so less work for the OS. On a typical computer you want small clusters because a single 1k file will still consume a full cluster, so with 8K cluster size you lose 7K of space, 16K clusters you lose 15K, and so on.... But if you are using a seperate data drive and are restricting it to big digital sound files this issue is moot. I'm useing a Maxtor 40gig U66 data drive and I just used the default cluster size, I've no problems recording 8 X 24 bit 44.1 tracks at once (and playing back 16 through 8 channels). And yes RAM helps a lot there, probably more during recording than during playback.
 
Back
Top