TranceAddict Forums (www.tranceaddict.com/forums)
- Production Studio
-- When does bit loss occur?
Pages (3): [1] 2 3 »

When does bit loss occur?

I bought a volume control unit a while back because I did some reading on bit loss since I didn't have my soundcard's level set to 0.0 db, rather I had it -5 db or whatever to control volume. I had no other way to control my monitor's sound output. With the cound control unit, I can keep my soundcard at 0.0 db, avoiding bit loss.

However my question is, does the whole keeping at 0.0 db to avoid bit loss apply to hardware AND the software? In other words, what if I have my master output in my sequencer at say -4.5 db while my soundcard's level is still 0.0 db, am I ok? Should I try and keep my sequencer's output at 0.0 db always?

Re: When does bit loss occur?

i don't think there is any way to attenuate volume, in the digital domain, without bit loss. therefore, in order to have no loss from volume changes, you would have to have every fader and plugin not affecting gain (obviously impossible).

i wouldn't worry about it dude, unless you're talking about -40db or something

It's virtually impossible to get any degradation using floating point. The "bit loss" term generally applies to 16- or 24-bit fixed-point digital transmission protocols like S/PDIF or ADAT.

I'm not sure I understand the problem exactly. Can't you turn the sequencer's main volume down so you're not overloading your monitors, then when you're ready to export the track, just put the volume back up so it peaks at -0.1 dB or so? (to answer your question - in the end you want your track to be around -0.1, -0.0. Whether you do that or you send it to a mastering studio, in the end it should be up there. It's not going to make much difference at the end of the day, if you have a completely clean mix which doesn't clip and peaks at -0.1 or one which peaks at -3.0. A mastering studio will be able to use either. Compared to many other production techniques, this one is not a major concern. Worry about a lot of other things before you worry about this.)

I do most of my mixing at extremely low levels, just by having an extra processor on the master channel to drop the level (and also to give me the option to listen/ mix in mono). When it comes time to export, I just remove that processor.

You got it the wrong way round, the soundcard volume is in the analogue domain, so you won't get bit loss there. You get bit loss from the level controls in your DAW software. But even at 16 bit, its not a big deal, don't obsess.

quote:

Originally posted by kitphillips You got it the wrong way round, the soundcard volume is in the analogue domain, so you won't get bit loss there. You get bit loss from the level controls in your DAW software. But even at 16 bit, its not a big deal, don't obsess.

Some soundcards have digital master volume. And no, you can't get bitloss in a DAW with a 32-bit floating point engine.

when you can't actually hear a difference, nobody gives a crap.

Thanks for all the info everyone its very insightful. One question though since you mention the 32-bit floating point engine. Does this mean I should be writing my tracks in 32bit float? Since you mentioned this, I looked and I see now in Cubase that is an option under project setup. I've always had this set to 24bit. Does it matter than my soundcard, the Audiophile 2496, is only capable of 24bit, 96khz?

I guess I'm a bit uncertain then what I should be writing my tracks in. I currently write them in 24bit, 48khz. I realize everything gets broken down eventually to 16bit, 44khz so what would be ideal for my situation? 32bit float, 44khz, 32bit float, 48khz, etc?

I appreciate all the help.

quote:

Originally posted by Pjotr G when you can't actually hear a difference, nobody gives a crap.

Actually the whole reason I came on here and asked about this is because I can hear a difference and I'm trying to determine what caused this difference because many factors were involved in me getting the 2 conflicting outcomes, confusing the hell out of me. I will see what answers come for my response above this post and then explain the details of my problem (if anyone is interested of course). Again, I'm thankful for all your help.

I used to always run at 32bit/96khz, which I always thought sounded better but, then I never liked the way it sounded when I dithered down to 16bit/44khz. So nowadays I usually run at 32bit/44khz and then render down,and it actually sounds better to my ears but, maybe I was doing something wrong when I ran at the higher rates. The only time I record at the higher rates anymore is if I am recording a vocal, so I can capture all nuances.

before i used the digital fader volume on my echo audiofire, then i bought a passive volume control by sm pro audio and set master out to +4 and the fader -4 so the output is 0dB (dont know if the sm pro audio is balanced with jacks), i recieved alot of dynamics at low volums. also i stopped rendering at 48kHz 24bit and use instead normal 44.1 16bit as i had problems when converting to mp3.

quote:

Originally posted by palm before i used the digital fader volume on my echo audiofire, then i bought a passive volume control by sm pro audio and set master out to +4 and the fader -4 so the output is 0dB (dont know if the sm pro audio is balanced with jacks), i recieved alot of dynamics at low volums. also i stopped rendering at 48kHz 24bit and use instead normal 44.1 16bit as i had problems when converting to mp3.

Interesting you say this because that is what I noticed too and part of the reason for my questions. When I render out the project at 24bit 48khz, I noticed the resulting MP3 when I converted it sounded way shittier than if I rendered out of Cubase at 16bit, 44khz and turned it into an MP3. Sorry if this sounds dumb but why is this?

i have no idea but i realy dont see the reason why to render at 48kHz and 24bit anyway. yes it sounds slightly better spesialy in the bass but that wont help anyone if the format is unsupported anyway. as far as i understand 24bit gives u more headrom and better signal-noisa ratio and therefore also better dynamics but i might be out there. im not a theroetic i only trust my ears and they are rarely wrong (except with eqing ). My advice record everything in 44.1 16bit to avoid what happened to my previous release on joof, the file got screwed up on beatport and the other sites beacuse i provided 24bit. same happeneded with whirlloop. u could render a 24bit version for yourself and send 16bit to labels. also max your soundcard output and use an analog volumkontrol for monitoring, this is what i got: http://www.google.com/products?q=sm...roducts&show=dd

quote:

Originally posted by palm i have no idea but i realy dont see the reason why to render at 48kHz and 24bit anyway. yes it sounds slightly better spesialy in the bass but that wont help anyone if the format is unsupported anyway. as far as i understand 24bit gives u more headrom and better signal-noisa ratio and therefore also better dynamics but i might be out there. im not a theroetic i only trust my ears and they are rarely wrong (except with eqing ). My advice record everything in 44.1 16bit to avoid what happened to my previous release on joof, the file got screwed up on beatport and the other sites beacuse i provided 24bit. same happeneded with whirlloop. u could render a 24bit version for yourself and send 16bit to labels. also max your soundcard output and use an analog volumkontrol for monitoring, this is what i got: http://www.google.com/products?q=sm...roducts&show=dd

This is what I was afraid of. I sent 24bit versions to be mastered and I've been 2nd guessing myself now thinking, should I have just sent 16bit, 44khz? Then part of me is like the mastering people know what they are doing so I shouldn't be worried. I'm still trying to understand why the by-product of a 24bit render sounds worse the the straight 16bit render straight out of the sequencer.

quote:

Originally posted by Magnus This is what I was afraid of. I sent 24bit versions to be mastered and I've been 2nd guessing myself now thinking, should I have just sent 16bit, 44khz? Then part of me is like the mastering people know what they are doing so I shouldn't be worried. I'm still trying to understand why the by-product of a 24bit render sounds worse the the straight 16bit render straight out of the sequencer.

I went to a mastering engineer last year with a track.

I rendered one wav file as 96khz 32bit float, and one 44.1khz 16bit just in case, and he told me that was a good idea, he couldnt put his finger on what exactly, just that the 96khz file sounded better, and he used that throught the mastering process, which he rendered down to 44.1khz 16bit when he was done.

So u shouldnt be afraid to send higher rates to an engineer.

yeah if it goes thru an engineer. my track was released as i delivered it, no work was done to it as they meant it sounded good enough. then when they uploaded it to beatport, which requires 16bit 44.1 wav, caus the mp3 decoding lies there, everything went wrong. somewhere in the chain this should be have discovered but it didnt before it was too late. this happened eventho i clearly stated that the file was in 24bit 48kHz. so just be carefull. it was fixed a few weeks later tho but there where some angry people around caus they have bought a wrecked mp3. everything 44.1/16bit for me now

Thanks again guys I'm going to render back out my tracks and re-send them.

Just a thought but......

The difference in sound qaulity between an mp3 encoded from 16b 44k and 24b 96k (or higher) could be due to the programming of the common mp3 encoder?

It (the algorithm) was originally concieved to render down from 16b 44k, so maybe when converting from higher formats, the codec/algorithm itself has not been as thourougly produced.

Also, as a side note - bit depth relates to resolution in terms of dynamic range in DB. People who bang on about analog tape (etc) actually don't realise that the usable dynamic range of tape is only about 10bit, and at 6db per bit, that means only 60db of usable dynamic range (before distortion). So before you get hung up on whether you are getting enough out of your 16bit system, think about the average person's audio system (who will be listening to your track) and realise that bit depth isnot going to make such a difference.

Having said that, I know myself, there is a clearly audible difference between recordings I have made at 16bit and 24bit, so IMO your masters should be at 24bit 44k.

Anyone that tells you they can hear a difference between 44k and 48k is almost certainly lying. Anyone who can hear a difference between (and identify) a 44k and 96k is either:

A, lying,
B, A mutant
C, A very rare good sound/mastering engineer

The extra space a 96k recording will take up (double the file size) over a 44k version is really not worth it, but a 24 bit over a 16 bit is worth it (only about 20% more).

I don't think it's that surprising that some tracks sound worse when downsampled from 96 kHz to 44/48 kHz. If you have anything above the Nyquist frequency then you're going to get aliasing distortion when you downsample. If you do the entire mix in 44/48, then you just won't have those frequencies to begin with.

Try using a lowpass filter around 22/24 kHz before downsampling (in practice, probably a bit less than that because the "cutoff" is just the -3 dB frequency on the curve rather than a brick wall). It'll probably sound fine. You can definitely go down as far as 21 kHz - the most sensitive ear still can't hear sounds above that frequency (unless it belongs to a canine).

jesus christ we have been over this a million times.

ANYTHING OTHER THAN 44K IS A COMPLETE WASTE OF TIME IF YOU KNOW IT IS ONLY GOING TO END UP IN 44K ANYWAY. INFACT, HIGHER RATES MIGHT ACTUALLY SOUND WORSE AFTER DITHERING.

BIT DEPTH, THE HIGHER THE BETTER.



listen to digi and rann.
/end discussion

quote:

Originally posted by echosystm jesus christ we have been over this a million times. ANYTHING OTHER THAN 44K IS A COMPLETE WASTE OF TIME IF YOU KNOW IT IS ONLY GOING TO END UP IN 44K ANYWAY. INFACT, HIGHER RATES MIGHT ACTUALLY SOUND WORSE AFTER DITHERING. BIT DEPTH, THE HIGHER THE BETTER. listen to digi and rann. /end discussion

Nicely done.

quote:

Originally posted by Magnus Thanks for all the info everyone its very insightful. One question though since you mention the 32-bit floating point engine. Does this mean I should be writing my tracks in 32bit float? Since you mentioned this, I looked and I see now in Cubase that is an option under project setup. I've always had this set to 24bit. Does it matter than my soundcard, the Audiophile 2496, is only capable of 24bit, 96khz? I guess I'm a bit uncertain then what I should be writing my tracks in. I currently write them in 24bit, 48khz. I realize everything gets broken down eventually to 16bit, 44khz so what would be ideal for my situation? 32bit float, 44khz, 32bit float, 48khz, etc? I appreciate all the help.

You cannot choose 32-bit floating point, as it is what most daws use internally. The bit depth you choose only matters for samples in the project and exporting.
Go with 24-bit/44.1kHz or 16-bit/44.1kHz. You only need to use 24-bit if you're mastering the track externally.

quote:

Originally posted by Eldritch You cannot choose 32-bit floating point

you can actually.

quote:

Originally posted by Eldritch The bit depth you choose only matters for samples in the project and exporting.

it is of particular benefit for freezing/rendering parts.

Wow guys again thanks for all the info this has greatly helped me going forward with how I will do things.

quote:

Originally posted by echosystm you can actually.

I meant that it's not a choice. It will use 32-bit float internally regardless if you've chosen 16-bit, 24-bit or 32-bit for your project.

quote:

Originally posted by echosystm it is of particular benefit for freezing/rendering parts.

Yeah, that's what I said.

Powered by: vBulletin
Copyright © 2000-2021, Jelsoft Enterprises Ltd.