Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[bruno putzeys]

dcollins wrote on Wed, 18 July 2007 07:12

Why we are arguing this in 2007 is beyond me..............

Maybe we should make it compulsory for anyone trying to enter an AES convention to show proof of having attended one of JV/SL's digital audio basics tutorials at least once.

[Jim Williams]

I was a member of AES for many years, but no longer.

If digital converters did do an honest 24 bit resolution, I would be happy with a 140 db dynamic range, unfortunately the 24 bit resolution is nowhere close to that. Most of those last bits are just dither, not audio. 24 bit audio is still just a dream, not a reality.

[bblackwood]

I'm curious, is there any analog audio gear that ha a 127dB dynamic range?

[bruno putzeys]

A power amplifier easily does 130dB (includes some class D designs by yours truly). Some microphones pull off 130dB as well, as do some mic pres. Dan's converters are also in this ball park, as are Salzbrenner's although the latter uses gain ranging to get there. TI has DAC chips that promise 129dB when wired in mono. All depends on what kind of product you're making.

[bblackwood]

Bruno Putzeys wrote on Wed, 18 July 2007 13:49

A power amplifier easily does 130dB (includes some class D designs by yours truly). Some microphones pull off 130dB as well, as do some mic pres. Dan's converters are also in this ball park, as are Salzbrenner's although the latter uses gain ranging to get there. TI has DAC chips that promise 129dB when wired in mono. All depends on what kind of product you're making.

Well, I meant in context of this discussion (summing, etc). Dan's Gold ADC is supposed to hit 127dB dynamic range - I'm curious if analog audio production gear is quieter than that. The argument that digital summing is limited by the converters seems strange considering most analog gear I'm familiar with is lucky to hit 100dB dynamic range...

[Jim Williams]

The BurrBrown PCM1792 dac in a dual sine/cosine configuration reaches -132 db. That is within 8 db of perfect 24 bit resolution.

The ADC's are not there yet.

Most "pro" audio gear is centered more on specific distortions than maximum dynamic range. Many of these currently made devices are still using 30+ year old analog stages. This also includes many popular converters.

Then again, with a 1 db dynamic range on most current CD releases, what's the point anyway?

When one masters a CD without any compression or limiting, one will start to appreciate the 90 db dynamic range this format is capable of and has never been fully used for.

[bblackwood]

Jim Williams wrote on Thu, 19 July 2007 09:43

The BurrBrown PCM1792 dac in a dual sine/cosine configuration reaches -132 db. That is within 8 db of perfect 24 bit resolution. The ADC's are not there yet. Most "pro" audio gear is centered more on specific distortions than maximum dynamic range. Many of these currently made devices are still using 30+ year old analog stages. This also includes many popular converters.

Oh, yah- I don't know if we'll ever achieve full 24 bit signals - thermal noise is tough to beat.

But my question remains unanswered - if the ADC is limiting digital summing, then what analog gear has a higher s/n ratio than 120dB? Much less a chain of analog gear...

I'm just unaware of any analog recording gear that beats that dynamic range, which confuses me as to your earlier point (the ADC being the weak link in digital summing).

[Jim Williams]

The thermal noise is tough to beat. It's not too much of a problem for me, I concentrate on the linearity of the system.

With the analog distortions below -125 db and the digital sources increasing in distortion at lower bit levels, it's the old decreasing distortion vs amplitude of analog against the increasing distortion vs amplitude of digital at low levels.

Another very important facet that was passed over in my previous post is the lack of depth in the digital mix core that the analog system seems to retain.
This is the main driving force for those to forfeit digital summing and buy an outboard box for this function.

I demonstrate this easily as I can post up a mix ITB and then feed that via a BurrBrown PCM1792 into my analog console's 2 track return. Then I'll post up the same mix on the analog console feed from a HDR and make the switch back and forth.

It's like a book opening and closing, analog = wide, digital = narrow. The front to back depth completely disappears in the digital mix where the surround elements just pop out in the analog mix.

I've yet to read or hear a suitable explanation for this.

[bruno putzeys]

Having heard a few too many boxes where the signal came out with a bigger and "more 3D" stereo image compared to the signal that went in, I no longer buy the idea that perceived depth (etc) is something that can only get lost. The thing that needs explaining, then is not why the digital chain loses depth, but why the analogue chain adds it. Somehow this avenue looks more promising to me.

BTW could you elaborate on the converters used when replaying from the HDR into the analogue board? Are they the same as the ones you're using to replay the ITB mix?

[Jim Williams]

I use a combination of AKM, Crystal and BurrBrown dacs. If I feed a stereo drum overhead pair from the computer to the PCM1792, it has no depth. If I feed it into the console from the HDR routed into the same PCM 1792 it has depth.

I don't believe the analog console adds anything, the AP does not show any added THD or noise from this set up. Crosstalk is at -90 db at 10k hz.

I can only surmise that the digital mix core is losing something.
All I need to do is put on a Mutt Lange Shania Twain recording to demonstrate that.

There are countless examples of folks experiencing the same phenomenom, I have no answers other that to use what works and let the AES code writers explain it.

So far, none have.

[bruno putzeys]

I see my question wasn't specific enough. I'll try again:
Lets look at the signal chain step by step. You have a pair of mics feeding two mic pres etc. From what point onward do the signal paths diverge? Do you use one converter going into the HDR and into the computer, or are you using the HDR's internal converters for one signal path and the DAW's internal converters for the other? Is the digital data that is on the computer exactly the same as the digital data that's on the HDR?

[Jim Williams]

The difference happens when the stereo tracks are fed into the digital mixer. If I route the stereo tracks directly out of the PC into the BB PCM1792 dac, the image is intact.

This effect only happens in the digital mix core, regardless of brand.

I only notice it as I use analog consoles to mix, I always have since the early 70's

This effect never happened with analog tape either, with it's limited crosstalk at high frequencies, older consoles with -60 db crosstalk. It's not so much the width left to right, but the total collapse of the depth and surround of the stereo tracks.

I cannot measure this with any test gear I have, but the audible difference is not subtle.

Since this is a new phenomenon, I believe it's source is the digital mix as I mentioned before, it never happened with any source mixed in the analog domain. It also carries over to the final product, whether it's vinyl with it's -25db crosstalk at 10k hz or the CD medium.

[maxdimario]

Bruno Putzeys wrote on Fri, 20 July 2007 16:52

Having heard a few too many boxes where the signal came out with a bigger and "more 3D" stereo image compared to the signal that went in, I no longer buy the idea that perceived depth (etc) is something that can only get lost. The thing that needs explaining, then is not why the digital chain loses depth, but why the analogue chain adds it. Somehow this avenue looks more promising to me. BTW could you elaborate on the converters used when replaying from the HDR into the analogue board? Are they the same as the ones you're using to replay the ITB mix?

are you saying that a passive resistor network ADDS something?

wouldn't it be MUCH MORE LIKELY that a digital system adds (or subtracts) something?

[Jon Hodgson]

maxdimario wrote on Tue, 24 July 2007 18:19

Bruno Putzeys wrote on Fri, 20 July 2007 16:52 Having heard a few too many boxes where the signal came out with a bigger and "more 3D" stereo image compared to the signal that went in, I no longer buy the idea that perceived depth (etc) is something that can only get lost. The thing that needs explaining, then is not why the digital chain loses depth, but why the analogue chain adds it. Somehow this avenue looks more promising to me. BTW could you elaborate on the converters used when replaying from the HDR into the analogue board? Are they the same as the ones you're using to replay the ITB mix? are you saying that a passive resistor network ADDS something? wouldn't it be MUCH MORE LIKELY that a digital system adds (or subtracts) something?

He's not talking about passive resistor networks.

[Schallfeldnebel]

If you want to know that, mix two stereo or four mono signals from different sources together, all with the same clock, pan it as you like, use any gain setting, it does not really matter. By routing the same sources into 4 other channels, using exactly the same levels and pan, but switched in reverse phase, the netto result should be 00000. If this works, digital mixing is clean. If you would do this trick with analog mixing, you will keep always a residue of noise and sound.

I am mixing digital since 1995 (Weiss), and I would not like to go back to any analog desk, although I do respect that others may find digital mixing not satisfactory, I doubt if their conclusions are based on the correct facts. Mixbusses with transformers give a nice signal soup of delicate distortion, just like the 1:20 step up transformers in old desks, which were a need to avoid noise in those days, but nowadays honoured as the only way to good sound.

Schallfeldwebel