20bit vs 24 bit, how big is the difference?

Dobro, one thing you've got to take into consideration. This is evolving technology. If, about 3 years ago, you would have told me: "within a couple of yours you are going to be recording and mixing digitally by preference", I would have asked you if you'd gone totally berserk. More and more things are being discovered and developed, as a result the possible quality of digital sound has / is improved dramatically.

Clocks were one item pretty much ignored in the audio industry. Not in the video and post production world, where synchronization is a much higher priority. You can ascertain the low priority f clocks by looking at the small number of manufacturers. Aardvark, NVision, Lucid, as well as some others whom are way off the price scale. If you look at the specs for the clocks, you will see that their main commercial use is video and post production, where the clocks function to keep entire multi - room operations with sound and image "in sync". This is why these clocks are expensive, and feature packed, providing for just about every standard and clock speed, as well as having a capability to deliver a variety of clocks. This is why they are expensive, average around $2000. This is also why they are still not as good as possible for audio, the massive amount of features in the clock necessitate some compromises. Lucid developed a stripped-down version, one without all the video stuff and excessive choices, especially for DAW users. This is the first time a good clock has been available at a ‘low’ price, just over 500 bucks. Due to its lack of video features, its jitter rate is very low, thereby providing for really amazing audio quality.

Regarding another part of your question, like I have stated many times; Manufacturing companies design recording products AT A PRICE, aimed at a certain market, aimed at a group of people who can afford to spend between x and y. And thank goodness they do, otherwise you’d still have to go to a large studio and pay a fortune to record anything. Something somewhere has to go in order to achieve the manufacture of a product people can affort. If that wasn’t so…. All there would be in the digital realm would be consoles made by Neve, Euphonics, Studer, Stagetec etc., all costing hundreds of thousands. The fun thing is ……. It IS now completely possible to get the same quality audio produced on those expensive pieces of gear on something costing but a fraction. In order to achieve that, you have to pay close attention on those elements which make that happen, and that is the reason why I’m always harping on about what makes digital sound good. To get there, focus on the basics of you chain. Get that right, and you’ll hear what you are doing, and you’ll enjoy the results. Then – perhaps start thinking about pre-amps, microphones etc.

Next, quote: “. But if you're *not* stacking up tracks, then it wouldn't make so much difference, right?” Of cause it would! All I did is outline a way to evaluate converters, the easiest manner in which to make quality differences audible. Quality = quality, no way around that.

Quote: “Finally, the 24/96 thing. The box studios they're selling these days that claim to track at 24/96 actually *do* track at that, right? And computer soundcards also, right?” Yes they do. Read the things above – make it at a cost and…… Apart from that, there are multiple items yet unsolved with 24/96, including processing power requirements, and storage capacity requirements. Look at the specs and manuals from, for instance the Mackie and Tascam 24/96 recorders. Welcome to the hype!! They are called 24/96. Start recording at 96. What do you get? 12 track maximum. So really – they should all be called 24/48’s or 12/96 recorders!!
Now also think of this. The reason the track count is halved, is that the units would need over 2 x the processing power in order to cope with real 24/96 recording. Please note OVER 2 x. Same with disk space. Now look further. How many effect processors, plug-ins, etc, etc, have got a 24/96 capability? Imagine how much computer power you need to run them, imagine how much processing power you need to start editing? Why has Digidesign not launched a 24/96 Pro Tools version?
The time is near – but it ain’t here yet!

You should be happy with what you’ve got. Why not? Its just that if and when anyone wants to improve their recorded sound quality, it helps a lot if they understand where to look to make the most effective improvements.
 
Cheers, Sjoko, thanks - that was really useful. I'd give this one another five stars if I could. :) It's nice to have priorities and directions clear.
 
Thanks for the explanation, sjoko. I have been recording at 44.1kHz (stuck at 20-bit dithered to 16-bit resolution for now) but I will definitely give 48k a try. Is there anything in particular I should know about SRC? Will, say, Sound Forge's SRC suffice or is there something significantly better? SRC should be the very last step applied to the stereo master, right? Also, I have read that whether or not you can actually record any higher than 16-bit, you should still work with 32-bit floating point files because it allows more final resolution after tracks are mixed, processed, etc.

...there are still a lot of problems associated with 24 / 96. Without going into the technical detail, just look at the hard-disk recorders, Mackie, Tascam, Radar, etc. Everyone advertising 24 tracks at 24bits, 96 kHz, but the fact is, if you want to USE the systems at 96 ---- you only have a maximum of 12 tracks available, as non of the systems have the processing power required to run at 24 / 96. There are a millon reasons why doing a project (properly) start to finish at that speed is still a dream. On top of that - ever calculated the disk space required to record at that rate?
Click to expand...

Even if the HDR's don't quite cut it at 24/96, if you have a DAW with enough processing power and hard drive space, what's the holdup?

I run a SRC192 in my studio, which converts anything into anything (up to 192kHz!) instantaniously, without quality loss (other than that naturally incurred by reducing things to a lower rate / speed). It even allows me to work on a project with material in different bits / speeds simultaniously.
Click to expand...

That is one cool piece of gear...I don't even want to know what that cost. ;)
 
Conversion should be the absolute last thing you do, keep all processing at higher clock speeds and bitrates wherever possible. The SoundForge dithering process should do the job, but of cause, there is always better, at a cost. Listen to your conversion critically, do it at a "best" or "audiophile" setting if your software allows for it.

Quote: "Even if the HDR's don't quite cut it at 24/96, if you have a DAW with enough processing power and hard drive space, what's the holdup?"
Oscillation, latency, clock accuracy, clock distributon (if a generic clock cannot cope with 44.1 accurately - imagine what it does at 96?), dithering down processing, lack of plug-ins, processors etc. And that is without going into all the techno crap.
Saying that, try it. It should sound amazing if your system can cope with it. If it doesn't, 24 / 48 is extremely cool as well.
 
So, a fast computer with a good clock and sync generator ought to be able to handle 24/96. As for the other stuff, you have to deal with that anyway no matter what bit/sample rate you record at.
 
quote: "So, a fast computer with a good clock and sync generator ought to be able to handle 24/96."
Dobro - the clock IS the sync generator. Ought to, perhaps. But just apply some logic. If systems costing thousands loose half their tracks if you choose to record at 24/96 .............
In a year or 2, yes.
Another good point to consider is that the 6 songs I did record and mix at 24/96, admittedly a lot of tracks per song, occupied the largest part of three 72GB drives.
 
So what about good dedicated converters, say the Lucid AD9624, are their clocks good enough? Also, if you're going from the Lucid to a soundcard via SPDIF, do you lose a lot of the benefit external converters if you don't sync via word clock?
 
Dolomite - Like always - good enough is relative.
The 2496 can generate clock. is it better then most generic clocks? very likely, by a fraction. Is it a very good clock? No, its not ment to be, and it can't be. If it was, it would have 2 independent power supplies, just for starters. And it would cost more than a normal 2496 plus a GEN6 combined.

The second part of your question is a very good one. Do you lose etc.
The answer is no. Again, I will refrain from going into technical stuff, because regarding converters and clocks, the jargon is pure gobbledegook. Look at clocks and converters as totally independent, but dependent on each other at the same time.
A good converter will improve sound quality.
A good clock will improve sound quality.
A good converter working with a bad generic clock will still make a major improvement.
A good clock working with a bad generic converter will still make a major improvement.
Both combined will make it as good as it can get.
 
Hmmm, interesting stuff!

I was under the impression that a good dedicated converter such as the Lucid or Apogee would have vastly superior clock circuitry to "stock" converters and would be only marginally improved by a dedicated clock. So, here comes another relative and highly subjective question: say you have a "prosumer" level sound card that has word clock I/O such as a Delta 1010 that, in all likelyhood, has the same (or quite similar) Crystal converters as many dedicated converters like the Lucid. You can probably guess where I'm going with this, but will you see a greater improvement by bypassing the Delta converters with the AD9624 or syncing the existing converters with the Gen6 (models just used for example)?

Also, do any of the prosumer sound cards come close to the conversion quality of the Lucid products? I hear very good things about the LynxOne and Aardvark has always claimed to incorporate some of their clock technologies into their DirectPro cards....
 
Dolemite - there's nothing like experience, of course, but I reckon you could answer the last part of your question by looking at the prices involved. You can get four channels on the LynxOne for $460. You get two channels of AD conversion only on the Lucid 9624 for $900. I bet they're pretty different. Either that, or the Lucid's overpriced, good as it is.

What's a Gen6? A clock?
 
sjoko2 said:
Again, I will refrain from going into technical stuff, because regarding converters and clocks, the jargon is pure gobbledegook.
Click to expand...
Hey sjoko do you know of any good web pages that do explain all (some) of this "gobbledegook"? :D

-tkr
 
Dobro, I don't doubt that the Lucid is superior is ways that I probably can't fathom, but I think your logic is a little flawed. Although money is definitely a limiting factor in my life and in the gear that I purchase, price is not some magical quality factor that turns shit into gold. ;)

There are loads of relatively expensive pieces of gear that sound like the drippy, runny poo that they are. I think its been established that the Lucid products are a bargain in the realm of dedicated converters, equalling or surpassing Apogee units which are multiples of their price. Is it not equally possible that the LynxONE is a bargain among sound cards? I have heard things to that effect and I merely wondered if the LynxONE might "come close" to Lucid quality and I wished to hear Mr. sjoko's thoughts on the matter. I welcome your thoughts as well but be prepared for a bit of a sermon with your "duh, look at the price" sort of answer. ;)

Its fairly well-known that a whole load of various pieces of gear that need to go from A to D and back again use the very same Crystal or AKM chip to do that conversion. These chips run a couple dollars each in quantitly. What is in certain ways the most critical link in an $3000 Apogee converter is a $5.00 or so part that you can also get in a $200 sound card. One would think that a $3000 converter would use at least, say a $50 converter chip. Nope. As I understand it, the real difference is in the supporting circuitry of the Analog I/O, clock generator, power supply, etc. The design is the key element here, and while the design does add to the overall cost, along with the parts that implement that design, a good design is probably the secret to the "cost defying factor" that make the Lucid, and from what I've heard, the LynxONE, sound so good for such a reasonable price.

Furthermore, given that the Lucid can compete with the Apogee, I imagine that they could charge Apogee-level prices without improving the quality of their product. In this case there would be zero quality increase for the increased price. So once again, cost does not equal quality. As a final thought, with converter chips running a few dollars each, you can imagine that its not terribly expensive to implement 2 channels of D/A in addition to 2 channels of A/D, so thinking that the LynxONE's lower cost per channel is an indication of its quality is again, flawed logic, especially since many converters can handle 2-in and 2-out with one single chip.

I apologize if I come off a little harsh but I hate when people give me this, "you gotta spend the big bucks to get quality" sort of argument, when, in the capitalist world of ours, that has been disproven time and time again. ;)

Dolemite - Stopping Misinformation One Post at a Time

Tekker,

I would check out Bob Katz's site, www.digido.com, if you haven't already. It does help wade through the details of digital conversion, jitter, dithering, etc., and the terminology should be quite applicable to converters.
 
Dolemite said:

I was under the impression that a good dedicated converter such as the Lucid or Apogee would have vastly superior clock circuitry to "stock" converters and would be only marginally improved by a dedicated clock. So, here comes another relative and highly subjective question: say you have a "prosumer" level sound card that has word clock I/O such as a Delta 1010 that, in all likelihood, has the same (or quite similar) Crystal converters as many dedicated converters like the Lucid. You can probably guess where I'm going with this, but will you see a greater improvement by bypassing the Delta converters with the AD9624 or syncing the existing converters with the Gen6 (models just used for example)?

Also, do any of the prosumer sound cards come close to the conversion quality of the Lucid products? I hear very good things about the LynxOne and Aardvark has always claimed to incorporate some of their clock technologies into their DirectPro cards....
Click to expand...

Couple of answers.
The second part of your question:
No, not a single soundcard comes close to the conversion quality of good stand external converters. Apart from costs, it is physically not possible to build a good converter in a card.

The first part:
The clock - when I said you would have a marginal improvement in clock accuracy - I still meant you would get a definite improvement over a generic soundcard clock.

In your example, I would say you stand to benefit more from the purchase of a good converter, in which case you get high quality conversion as well as a clock improvement.
An accurate clock can make further, substantial improvements, but can never correct the effects of bad conversions.
However, there is another element, which comes into play, which is the amount of equipment in a recording system requiring clock. The more pieces of gear, the worse the effect of bad clocking, the greater the improvement a good clock will make to the audio.
In other words, whereas one will give you consistent down-line denigration, the other will keep quality consistent 'as-is'.
 
Shit there is more :)

Dolomite, you are correct. Even the runny poo - just listen to a Manley Gold microphone....

The quality of a converter does depend entirely on the chip (and no, they are not all the same), and the surrounding architecture.
I believe that the reason why there are so many bad converters around is for the reason that most converters are designed by electronics engineers, not audio engineers, and definately not audio engineers with extensive analogue experience. I think here you can find the reason why Lucid scores, they have the advantage of 25 years of experience in analogue processing.

The costs are simply a question of business philosophy. Some corporations are entirely marketing driven, with large advertising expenditure which has to be re-couped, like for instance Apogee. Lucid / Symetrix are more old fashioned, don't spend large amounts on advertising and let product quality do its job.
Apart from all that there is another element that comes into play here. Something I know, because I was / am involved in it. Lucid's CEO held the opinion (totally right) that the audio manufacturing industry was to far removed from the "ground floor" where the products are used intensively. As a result, he approached some clients, people like myself, and asked if I would mind if he sat in on sessions etc. This resulted in a series of discussions, where a bunch of 'pro's" told Lucid exactly what they wanted and needed, followed by testing of prototypes etc. This process didn't stop until all invoved were convinced they had the best possible to work with, then the products were launched.
Now there is a situation where a bunch of us, including people like Massenburg, use Lucid stuff, and if ever they feel something could be improved they shout, and Lucid reacts by making improvements.
 
Originally posted by sjoko2
Originally posted by Dolemite
Also, do any of the prosumer sound cards come close to the conversion quality of the Lucid products? I hear very good things about the LynxOne and Aardvark has always claimed to incorporate some of their clock technologies into their DirectPro cards....
Click to expand...
No, not a single soundcard comes close to the conversion quality of good stand external converters. Apart from costs, it is physically not possible to build a good converter in a card.
Click to expand...
The Aardvark Direct Pro does have an exernal I/O box apart from the card, so do you think that would make a difference? So it sorta is an external converter and they say that they incorporate some of their clock technologies into those cards.....so do you think that they might have actually come close, or are they still quite a ways off?

Just curious, because I just got the Direct Pro and I'm wondering how this thing compares to the big boys.

-tkr
 
Last edited:
I have tested a bunch od Aardvark products, both converters and clocks. It is true that the clock on their cards is marginally better than those on average found on cards. I think their converters fall in the same league as MOTU's, pretty bad.
I have some interesting graphs with clock jitter data, but that's all high end gear tests. I haven't kept ant test data from cards. Pitty.
 
Damn! This is the second time I'm typing a response to this, the first one vaporised.

My thoughts? I LOVE Spectralabs, and I couldn't live without Spectrafoo :)

The most interesting thing that struck me is that George Massenburg send us a graph of the AardsyncII, which is Aardvark's masterclock (I think it costs between 1500 and 2000, not sure), which shows exactly the same huge peak in jitter at 11k. This shows me one thing, they have a very significant design problem throughout their clock products.

My other thoughts? Read the graphs. What do they show? Some serious harmonic distorsion, jitter, a high noisefloor, resulting low real dynamic range. Translated into audio, a product with anything but smooth reproduction, and serious flaws in a number of areas.

I love the "excellent" qualification. Excellent compared to what?

Thanks for that post, it really proves a number of points I made, as it was done with Spectralab stuff, it proves it accurately and clearly!
 
Sjoko, here's the rest of the site, click on the soundcards name to get to the graphs.

The most interesting thing that struck me is that George Massenburg send us a graph of the AardsyncII, which is Aardvark's masterclock (I think it costs between 1500 and 2000, not sure), which shows exactly the same huge peak in jitter at 11k. This shows me one thing, they have a very significant design problem throughout their clock products.
Click to expand...
So would you say that Aardvark clocks really aren't all that great (or is this more of a minor detail)? Could you recommend a good clock (best bang for the buck) that's not to spendy?

My other thoughts? Read the graphs. What do they show? Some serious harmonic distorsion, jitter, a high noisefloor, resulting low real dynamic range. Translated into audio, a product with anything but smooth reproduction, and serious flaws in a number of areas.
Click to expand...
Well, I don't like the sound of that......:( So would you say this guy kinda over rated this card with the all "excellent" marks? Because (not knowing what any of that tech stuff meant) I was pretty happy to say the least, when I saw all the excellent remarks for my new soundcard. :)

-tkr
 
You must log in or register to reply here.
Back
Top