Hello everybody,
First post so be gentle. I currently use a compact digital multitrack in my modest home studio. I have nice mics and pres and a nice little vocal booth and I reckon I do the best I can with my set-up. The thing is the multitrack is uncompressed 16bit wav format. I'm looking at some of the 24bit multitracks out there and wondering if I trade up will I notice the difference between 16 and 24bit?
Thanks muchly.

You'll have more dynamic range with 24bit. Each bit in the digital word is worth 6 decibels. However with much of today's popular music you probably wouldn't notice the difference because dynamics aren't stressed much and the CD is only coded to 16bit.

With that said though and for the purposes of starting at the highest reasonable quality level I suggest working at 24 bit when tracking because if you mix in a sequencer you will be altering the gain structure with virtually each thing you do and that will require extra calculations from your sequencer which will result in larger digital word-lengths or rather more bits,(usually sequencers calculate at least as high as 32bit even though you might have only tracked at 24). That being the case, I say take advantage of the available bits in a 24bit set up and then after you mix and if you don't intend to master then dither it down to 16 and be done with it.

Liz

--------------------
MIRRORMIX STUDIO
blog

I guess it all depends on your hearing.
If you have been exposed to constant loud noise then the extra detail 24bit offers, could be missing.

Your age often determines the amount of high end loss - ie the older you get, the less you can hear.
Which is one reason why a 25yr old is a better age for an engineer than a 40 yr old.

Obviously experience and a trained ear helps counter act this !

as for the technical side :-

24bit is much clearer, akin to wiping the mist of a window, revealing far more detail than 16bit.

The converters used in your recorder will have an impact, which is why we prefer to use dedicated and expensive ! ) devices maked be compaines like Apogee etc.
These units use a very stable system clock - a detail often lost in cheaper boxes. The satiability of the clock has immense implications when converting digital data to an analogue voltage.

A good clock will add further detail, increasing the stereo image and depth of sound.


Finally your monitors have a huge effect.
Most decent speakers will convert all that extra noise into something worth listening to. but again quality costs.


In the most part, moving to 24bit is a good thing.
It is unlikely that you will not notice the difference, but as to how large that difference is, depends on the rest of your setup.

Quote tigali:

---

Hello everybody,
First post so be gentle.

---

OK!

Quote tigali:

---

I currently use a compact digital multitrack in my modest home studio. I have nice mics and pres and a nice little vocal booth and I reckon I do the best I can with my set-up. The thing is the multitrack is uncompressed 16bit wav format. I'm looking at some of the 24bit multitracks out there......

---

Which 24 bit multitracks are you looking at? Do you mean something like an "all in one," such as an AW1600-with mic preamps and everything else, or do you mean something like the Alesis HD, Radar units, Tascam, etc.?

Quote tigali:

---

and wondering if I trade up will I notice the difference between 16 and 24bit

---

There is nothing intrinsic to 24 bit recording that offers better sound. No extra detail, etc. But 24 bit hardware might sound better because the converters are simply better.

OTH, unless everything else in the multitrack box is well built-the analogue signal paths, the power supplies, the quality of the materials used, and so on; your older gear may sound much better than the new gear.

24 bit does offer an extended dynamic range which is very useful if you know how to exploit it. It can make your analogue gear sound noticeably better, because it can be allowed to operate at its most advantageous levels.

Processing and mixing have some implications which have already been mentioned.

Your questions can only be answered by considering specific gear, and not by genres: ex., 24 bit v. 16 bit.

Quote Scope:

---

I guess it all depends on your hearing.
If you have been exposed to constant loud noise then the extra detail 24bit offers, could be missing.

Your age often determines the amount of high end loss - ie the older you get, the less you can hear.
Which is one reason why a 25yr old is a better age for an engineer than a 40 yr old.

Obviously experience and a trained ear helps counter act this !

as for the technical side :-

24bit is much clearer, akin to wiping the mist of a window, revealing far more detail than 16bit.

The converters used in your recorder will have an impact, which is why we prefer to use dedicated and expensive ! ) devices maked be compaines like Apogee etc.
These units use a very stable system clock - a detail often lost in cheaper boxes. The satiability of the clock has immense implications when converting digital data to an analogue voltage.

A good clock will add further detail, increasing the stereo image and depth of sound.

---

--------------------
Nik
Godin, Axon, Tonelab, Repeater & the skin of my teeth!

Quote Shingles:

---

---

Could you elaborate ?

( please ? )

At best, a 16 bit system will provide a dynamic range of around 96dB. That is fine for a final delivery format when the dynamic range of the source material has been controlled, but it is a little difficult to use with potentially unpresictable and dynamic source material when recording raw. It is workable, but only with considerable care that takes attention away from performance and production.

In the good old days of analogue, any half decent desk would manage a dynamic range of about 120dB, and the way the metering was set up, a considerable headroom margin was left (typically 20dB or so) to avoid internal clipping on transients. So engineers really didn't have to worry about transient peaks -- they wouldn't clip the desk and the tape would ignore or squash them in a benign way -- and could concentrate on more important aspects of the recording session.

16 bit recording was only ever intended as a stop gap until the technology improved, and it quickly did with 20 bit and then 24. Although few budget 24 converters actually better 21 bit performance currently, there are true 24 bit converters out there, and some appear to manage even better (Neumann's digital mic converters are 28 bit and Stagetec's truematch converter is 30+).

So, all of that is a long winded way of saying that working with 16 bits is harder and technically inferior to working in analogue, while 20 bits is the same and 24 bits is better.

The downside of working at 24 bit instead of 16 is 33% more data to move around and store.... but then hard disk capacity and bu data rates just aren't an issue anymore, so there is no really valid argument for not recording, archiving and post-producing your material at 24 bits (or more).

For consumer release, 16 bits is still a totally appropriate format, since very, very few home music systems could cope with a greater dynamic range anyway (even if they wanted to)....

hugh

--------------------
Technical Editor, Sound On Sound

One very significant advantage of working at 24bit when recording is that many lower cost mixers will exhibit higher
distortion as they reach the last few dB's of headroom on each channel. By running the channels a little lower than 0VU when you PFL on the mixer the advantages in less distortion offset the relatively low signal at the A/D(input on recorder).Some transients will actually be higher than the VU/Peak meter will indicate.

I often track at -18dBFS and this leaves plenty of analogue and digital headroom. You may have a slightly lower S/N ratio in the analogue domain (i.e. the mixer) but if you are using something like a Mackie or better then this will be pretty much unnoticeable.

It surprises me when I see LED VU meters on live desk channels running at +15VU (1 LED before max) which occasionally ends up briefly clipping an A/D and probably has a sharp rise in analogue opamp distortion as the device reaches is operational limits.

--------------------
Mastering online mastering

I'd just like to reinforce Hugh's comment about 24bit recording being so much less stressful because it can be set up to give a lot more headroom.

That is the biggest difference I noticed when moving to a 24 bit system.

J

When I went to 24bit, I sold a single CPU G4 & bought a dual processor version.
I did not notice any significant increase in plugin capacity. ( ie virtually none ! )

24 bit audio places a lot of extra load on a host system, but then you are looking at dedicated units, so I guess this is not really an issue.

The extra headroom is a definite bonus, especially when using compressors/ limiters to max your input signal at the converter.

every bit helps !

Quote SafeandSound123:

---

By running the channels a little lower than 0VU when you PFL on the mixer the advantages in less distortion offset the relatively low signal at the A/D(input on recorder).Some transients will actually be higher than the VU/Peak meter will indicate.

---

0 VU is normally aligned to be about +4dbu out of the mixer, which is probably around 18 - 20db below the desks maximum output anyway.
Further the VU is NOT a peak reading device, so there will of course be higher peaks, but probably not 18db higher peaks! Now there are some meters used on cheap gear that are (despite the labelling) NOT VU as they have completely the wrong ballistics (possibly and scale markings), but 0VU should be WAY below clipping for any sane analogue desk.

Quote:

---

I often track at -18dBFS and this leaves plenty of analogue and digital headroom.

---

-18dbFS = 0VU = +4dbu is one of the two standards for this, but if you have aligned as above, 0VU will be -18dbFS, and digital full scale will be +22dbu (which is reasonable given balanced connections and a sane desk design, and is way past the top of the VU scale).

Quote:

---

It surprises me when I see LED VU meters on live desk channels running at +15VU (1 LED before max) which occasionally ends up briefly clipping an A/D and probably has a sharp rise in analogue opamp distortion as the device reaches is operational limits.

---

If it is an LED meter, it probably is not a VU (in most cases), and is most likely much closer to peak reading then a true VU (it may even be peak reading), which probably explains why the user is getting away with that.

I would also note that some desks (and especially some preamps) distort when driven hard in musically useful ways (Not that I would class a Mackie in that category), so hammering the channel can be a useful trick.

Ref the 16 Vs 24 bit thing, there is NO extra resolution, just a lower noise floor. Both systems are perfectly linear (If implemented correctly (Not always a given)), the ONLY advantage of 24 bit an increase in available dynamic range, nothing else (That is reason enough to use it in production).

Regards, Dan.

--------------------
Audiophiles use phono leads because they are unbalanced people!

Just a quick hi-jack question - I am intending to record some vinyl onto my PC so that I have it available digitally for listening and maybe using in Ableton / Traktor etc.
I was thinking 16-bit 44.1kHz would be o.k. simply because it is all modern electronic / dance music with very predictable peak levels and dynamics (i.e. I can run it close to 0dBFS without worrying too much).
Does this seem sensible?
The only processing I intend to do would be take out any obvious clicks by hand drawing them out in Soundforge and then a peak normalize to -0.3dBFS.
The other half of me thinks I should just go for 24-bit / 96kHz so I have an 'archive' copy at 'higher' quality - I have enough space for it. But re-sampling and bit-depth converting if I burn to CD etc. seems like a hassle.

Quote Scope:

---

When I went to 24bit, I sold a single CPU G4 & bought a dual processor version.
I did not notice any significant increase in plugin capacity. ( ie virtually none ! )

---

You would only see an increase in capacity if the software was written to share the plugin load across both chips. My betting is that it was not (It can be very hard to do for some processing cases, not all audio processing graphs parallelize well).

Quote:

---

24 bit audio places a lot of extra load on a host system, but then you are looking at dedicated units, so I guess this is not really an issue.

---

In all probability whatever the source bit depth, it is converted to either a fixed or floating point format that is independent of the initial format before being fed to the plugins, so there is actually negligible difference there. Most plugin systems are effectively O(1) in input word length and O(n) in sample rate.

The plugins themselves may obviously have any time behaviour in terms of sample rate, but as the data format is normally fixed at that point, will be O(1) in bit depth as far as the external system is concerned.

Quote:

---

The extra headroom is a definite bonus, especially when using compressors/ limiters to max your input signal at the converter.

---

Ouch, that is very poor strategy IMHO as not only does it force you to make processing decisions that you then cannot undo, but the make up gain raises the room/mic/preamp noise floor during quiet parts to no benefit.
For most real world rooms, with most real world mics/preamps, the self noise of the room/mic/preamp will (even allowing for say 20db of headroom) be greater then the self noise of the AD by quite some margin. Squeezing the audio before recording it was maybe necessary with the early 14 or 16 bit converters (where leaving 20db of headroom would raise the AD noise to objectionable levels), but a modern converter is not generally the weak link in the chain in that way.

Regards, Dan.

Edited by dmills (24/05/08 11:32 PM)

Quote Hugh Robjohns:

---

At best, a 16 bit system will provide a dynamic range of around 96dB. That is fine for a final delivery format when the dynamic range of the source material has been controlled, but it is a little difficult to use with potentially unpresictable and dynamic source material when recording raw.

---

I'm not sure I've ever recorded in a room that has a dynamic range in excess 65 dB.

I'm now a little confused.

If I want to record stereo (with two mics of course), should I keep the recording level on the meters down to about -18dB?

I normally record with about -6dB with only the very loudest sounds reaching -3dB. (This is classical music - say piano and violin - in a concert hall or church). I'm good at predicting peak levels as I was a musician myself in this area.

Quote Scope:

---

Quote Shingles:

---

---

Could you elaborate ?

( please ? )

---

Quote dmills:

---

Ref the 16 Vs 24 bit thing, there is NO extra resolution, just a lower noise floor. Both systems are perfectly linear (If implemented correctly (Not always a given)), the ONLY advantage of 24 bit an increase in available dynamic range, nothing else (That is reason enough to use it in production).

---

--------------------
Nik
Godin, Axon, Tonelab, Repeater & the skin of my teeth!

Thanks for all the replies. Brilliantly informative stuff. I appreciate the weakest link argument (that it is very dependent an montiors, mics,etc) but as far as I can gather, it's better to have it than not but not to expect a revelation in sound. The machines I'm looking at are compact mutitracks like the Tascam 2488 mk2 and the Korg D3200. I have often toyed with the idea of completely changing the studio to run something like Logic on my iBook or going with a digital desk and something like the HD24 but I just really enjoy the dynamic of using the compact multitrack. Workflow is everything to me and I can't see me ever being happy p*ssing about with plugins and a mouse. Anyway, thanks again people.

I prefer hardware too! the Yamaha units are pretty good - Aw1600 or AW2400 and both can do 24bit. If you can find a second hand AW4416 in good nick they are brilliant and have everything in the box

Cheers

What kind of price can you expect for a second hand one? I was very impressed with the spec of the Roland VS2480 cd but it seems to be discontinued despite still being advertised on the Roland website. I just love working with hardware and I'd be love to see a high quality 24bit 24 track recorder with some decent EQ. I don't need a million features I just want the best quality recording of what I'm playing. Having said that, I do miss the vari-speed function on my old Tascam 4Track portastudio. Fun with tape and varispeed. Now that would be a seller. A nice reliable,affordable 1inch 4track retro chic tape recorder.Why doesn't someone lash something like that together?

Quote tigali:

---

Thanks for all the replies. Brilliantly informative stuff. I appreciate the weakest link argument (that it is very dependent an montiors, mics,etc) but as far as I can gather, it's better to have it than not but not to expect a revelation in sound.

---

Quote:

---

The machines I'm looking at are compact mutitracks like the Tascam 2488 mk2 and the Korg D3200. I have often toyed with the idea of completely changing the studio to run something like Logic on my iBook or going with a digital desk and something like the HD24 but I just really enjoy the dynamic of using the compact multitrack. Workflow is everything to me and I can't see me ever being happy p*ssing about with plugins and a mouse. Anyway, thanks again people.

---

--------------------
Nik
Godin, Axon, Tonelab, Repeater & the skin of my teeth!

Quote Ariosto:

---

I'm now a little confused.

If I want to record stereo (with two mics of course), should I keep the recording level on the meters down to about -18dB?

---

-18db relative to what exactly?

If the meter has VU dynamics but is calibrated so that 0dbFS is at the top of the scale (which does exist on gear and is stupid), then yes, -18 or so is a good place to be (with 24 bit).
If the meter is actually VU and is aligned such that 0VU = -18dbFS, then somewhere around 0 on the VU is a good place to be.
If it is a PPM aligned as above then PPM6 is a good place to be (PPM is closer to peak reading so 14db above what it sees as peaks is probably OK)....
For an IEC style DPM, I would aim to leave about 10db, just because it avoids loosing that one otherwise perfect take.
For a meter with no documented calibration, who the hell knows (Horribly common)?

HTH.

Regards, Dan.

--------------------
Audiophiles use phono leads because they are unbalanced people!

Yes corrections appreciated and all important. Slap on the wrist for me.

Confusion stems from Mackie manual suggesting VU at every opportunity and to confuse matters "0VU" as Mackie would have is ii is actually 0dBu in there design. I prefer to be safe than distorting which is pretty much irreversible after the event. I compared the newer DIP8 opamp based desks with discrete circuitry of the past in higher end equipment which had very high headrooms of +30dBu or more.It's a fair comment IMO for the sake of cleaner audio on equipment with a little less headroom.

The meter on my Mackie mixer is peak reading and not VU
but when I see +15dBu on a Soundcraft desk channel (not exactly known for transparency, esp the budget desks)This does not make me confident what is getting to the recorder is clean esp as opamps distortion tends to be higher in the final few dB of headroom.

The ballistics of a BBC PPM meter is incomparable with a digital peak meter, it will not register brief digital overs. And assuming a peak of PPM6, (8 dB above -18dBFS) being -10dBFS (or on a transient even higher), 14dB above that is surely crispy time.

If you have to use a meter which has no documentation or is abiguous it is safer to peak a little lower for safety than to cause irreversible damage to a recording.



--------------------
Mastering online mastering

Quote SunShineState:

---

I prefer hardware too! the Yamaha units are pretty good - Aw1600 or AW2400 and both can do 24bit. If you can find a second hand AW4416 in good nick they are brilliant and have everything in the box

Cheers

---

They can do 24 bit, but the mic preamps-at least in the 1600-are not remotely capable of exploiting 24 bit dynamic range.

Not necessarily a reason not to record at 24 bit.

Quote SafeandSound123:

---

Yes corrections appreciated and all important. Slap on the wrist for me.

Confusion stems from Mackie manual suggesting VU at every opportunity and to confuse matters "0VU" as Mackie would have is ii is actually 0dBu in there design.

---

The point is not to debate various manufacturers' gain structure and level definitions, rather it's achieving optimal levels-dynamic range and low noise, not to mention just sounding good-throughout whatever device you're using.

Set up the gain structure on your Mackie exactly as Mackie recommends. If this results in peak levels of -18, -16, or -12 dBFS on your recording device, so be it. That's what 24 bit is all about, and good 16 bit gear will sound fine with this arrangement as well.

I think the debate about conventions between manufacturers gain staging and good sound are inseperable.

--------------------
Mastering online mastering