Wouldn't converting twice harm the signal?

[Sugarman]

I am thinking about buying a pair of outboard EQ's for mastering my ITB Pro Tools HD mixes. Perhaps I will buy some old german broadcast modules from Telefunken or Neumann and have them racked.
Anyway.
I am thinking if the audio will be "damaged" from going out of the 192 converters, through the analog EQ and back to digital again when recorded to a new stereo track in Pro Tools..? Would the sound quality be better if I just stayed ITB and instead used a really good software EQ like Waves SSL?
I know all mastering studios use outboard, so I am probably wrong here! But could someone please explain to me why it wouldn't be bad for the audio to be converted twice.
Thanks

[beanface]

Originally Posted by Sugarman

I am thinking about buying a pair of outboard EQ's for mastering my ITB Pro Tools HD mixes. Perhaps I will buy some old german broadcast modules from Telefunken or Neumann and have them racked.
Anyway.
I am thinking if the audio will be "damaged" from going out of the 192 converters, through the analog EQ and back to digital again when recorded to a new stereo track in Pro Tools..? Would the sound quality be better if I just stayed ITB and instead used a really good software EQ like Waves SSL?
I know all mastering studios use outboard, so I am probably wrong here! But could someone please explain to me why it wouldn't be bad for the audio to be converted twice.
Thanks

If you stuck on 2 extra channels of good ad/da, it would probably be worthwhile depending on what your working on. I personally wouldn't do it through digi conversion though, but you could use it for your monitor dac still. Any extra pass of conversion is gonna take something away from the sound, its just best to make it as little as possible. What about something from benchmark, mytek or lavry. Secondhand ads might be worth checking, maybe audiogon.com too, I've seen some really good stuff on there for fairly cheap, but most of it is just high end hi-fi. Worth a look though.

[MrVelvet]

Originally Posted by beanface

If you stuck on 2 extra channels of good ad/da, it would probably be worthwhile depending on what your working on. I personally wouldn't do it through digi conversion though, but you could use it for your monitor dac still. Any extra pass of conversion is gonna take something away from the sound, its just best to make it as little as possible. What about something from benchmark, mytek or lavry. Secondhand ads might be worth checking, maybe audiogon.com too, I've seen some really good stuff on there for fairly cheap, but most of it is just high end hi-fi. Worth a look though.

I actually rated the 192's very high in a blind test... I think it's pretty common here to bash them, just check for yourself what you loose. Simply send a signal out and record it back. Sure, you'll loose a tiny little something with extra conversation, but don't worry too much about it. People always make an issue out of it as if it sounded like going through a sample rate converter at 32 khz during that extra pass. At this time, every run through outboard benefitted enough for me to warrant the extra "mile". It's such a small bit that one pass of already digitized material will loose while making the roundtrip. Not much of a tradeoff when you consider how you affect the music with the outboard EQs.

[beanface]

Originally Posted by MrVelvet

I actually rated the 192's very high in a blind test... I think it's pretty common here to bash them, just check for yourself what you loose. Simply send a signal out and record it back. Sure, you'll loose a tiny little something with extra conversation, but don't worry too much about it. People always make an issue out of it as if it sounded like going through a sample rate converter at 32 khz during that extra pass. At this time, every run through outboard benefitted enough for me to warrant the extra "mile". It's such a small bit that one pass of already digitized material will loose while making the roundtrip. Not much of a tradeoff when you consider how you affect the music with the outboard EQs.

No, I'm not bashing them, I just wouldn't personally use them for master processing, thats all - when compared to better converters like prism, lavry etc, they just don't sound as engaging or as detailed.

[Sugarman]

thanks for your replies.
I actually do have a Benchmark ADC1 and DAC1 (stereo AD/DA) that I don't use anylonger because I got the 192 and it was easier to just use that alone.
I am just about to sell my DAC1.
Would you recommend that I keep the ADC1 and use it for a stereo input? Or should I keep both so I also use the DAC1 for the stereo output?
Would it really make a difference you think?? (I could use the money from selling it)

[bob katz]

Originally Posted by Sugarman

I am thinking about buying a pair of outboard EQ's for mastering my ITB Pro Tools HD mixes. Perhaps I will buy some old german broadcast modules from Telefunken or Neumann and have them racked.
Anyway.
I am thinking if the audio will be "damaged" from going out of the 192 converters, through the analog EQ and back to digital again when recorded to a new stereo track in Pro Tools..? Would the sound quality be better if I just stayed ITB and instead used a really good software EQ like Waves SSL?
I know all mastering studios use outboard, so I am probably wrong here! But could someone please explain to me why it wouldn't be bad for the audio to be converted twice.
Thanks

It's a matter of balancing the losses against the improvements! There is no perfect D-A-D chain, but some come close enough to "transparent" so that the improvements in the analog chain (EQ and/or analog compression) outweigh the slight transparency losses. It used to be that adding a D-A-D chain added some harshness due to monotonicity issues and other losses, but at high sample rates these days, the remnant effect is a kind of "softening" of the sound but usually no harshness. And that very softening effect of adding a high-quality D-A-D chain can become part of the very sound quality that we are seeking during the mastering, so it becomes part of the processing. Other losses of course include a slight loss of soundstage width, but then we seem to gain it back through the phase shifts and other euphonic anomalies provided free of charge in typical analog processing chains!

I do suggest something better than a Digi 192, though. It's not quite "damage" (the days of really bad converters are gone) but it is a significantly lossy path compared to a mastering-grade converter set.

BK

[beanface]

Originally Posted by Sugarman

thanks for your replies.
I actually do have a Benchmark ADC1 and DAC1 (stereo AD/DA) that I don't use anylonger because I got the 192 and it was easier to just use that alone.
I am just about to sell my DAC1.
Would you recommend that I keep the ADC1 and use it for a stereo input? Or should I keep both so I also use the DAC1 for the stereo output?
Would it really make a difference you think?? (I could use the money from selling it)

I'd keep both of them and you've got yourself a nice little loop for your eqs.

[Sugarman]

thanks bob and beanface.
some great info!!

Are my Benchmark ADC1 and DAC1 what you would call "mastering grade"..? Are my Benchmark converters a lot better than my 192?
what about just running drum bus, vocals etc through outboard and back to digi when mixing? In that case, are the converters less important? I know no mixing engineers that are using any other converters besides their standard 192 or Apogee when using outboard.

[dcollins]

Originally Posted by bob katz

Other losses of course include a slight loss of soundstage width, but then we seem to gain it back through the phase shifts and other euphonic anomalies provided free of charge in typical analog processing chains!

That doesn't really make sense to me. And if it was the case, running through your analog chain 20 times it would make it as wide as the great outdoors........... Does that happen?

I do suggest something better than a Digi 192, though. It's not quite "damage" (the days of really bad converters are gone) but it is a significantly lossy path compared to a mastering-grade converter set.

Speaking of multiple generations of A/D/A, here is 60, yes 60!, passes through a converter that I would say BK is not exactly going to rate as "mastering grade" or give an A++ grade to.

ZMIX Converter Tests

The signal is clearly rendered totally unrecognizable..............


DC

[Sugarman]

Originally Posted by SafeandSound

Doing another AD/DA ALWAYS degrades the audio (at least in theory), it is whether equipment in between can subjectively improve the music in a meaningful way which determines whether it's a worthy trade off.

It's the argument everyone forget about in the big OTB vs ITB discussions.
There MUST be a degrade in sound, and that should be a good reason for using more plugins (of good quality) and keep it in the box, no?

[bob katz]

Originally Posted by Sugarman

thanks bob and beanface.
some great info!!

Are my Benchmark ADC1 and DAC1 what you would call "mastering grade"..? Are my Benchmark converters a lot better than my 192?

I believe they are. Significantly more transparent to my ears and definitely mastering-grade. Why not do a shootout. Match up your levels carefully with a 1 kHz -20 dBFS test tone. The big problem is maintaining clocking integrity so I suggest you capture two musical files back into your DAW and then play them back in sync and switch between them with a common monitoring DAC.

With the Benchmark, the ADC should be your master clock, and with Pro Tools, the system should be in internal sync unless you have a demonstrably-superior external master clock.

BK

[philip]

Originally Posted by bob katz

I believe they are. Significantly more transparent to my ears and definitely mastering-grade. Why not do a shootout. Match up your levels carefully with a 1 kHz -20 dBFS test tone. The big problem is maintaining clocking integrity so I suggest you capture two musical files back into your DAW and then play them back in sync and switch between them with a common monitoring DAC.

With the Benchmark, the ADC should be your master clock, and with Pro Tools, the system should be in internal sync unless you have a demonstrably-superior external master clock.

BK

And when your at it. You can do a proper testing easy.

Feed your ADC with any analog signal (CD player will do, music or noise), route the signal to both DACS, route the DACs to monitorcontroller, route the analog signal to monitorcontroller. make sure you match the signal within 0.05 dB.

Now you can listen to the dacs output (eg A/B) but the real advantage is that you can compare the DACs to a REFERENCE, the analog signal. Compare DAC1 with reference - how much degradation do you hear? Compare reference to DAC2, how much degradation do you hear? The DAC who degrades the signal least is possibly the most transparent (the only thing you want from a mastergrade DAC)

After you done this test it's much easier for you to hear differences between the DACS.

You can do the same test with the ADC, one by one with the new DAC reference or both ADC and DAC in one go (witch is preferable in you case you're looking for the least colored AD/DA chain in this thread)

/philip


And btw, higher sample rates then 44,1Khz is never needed except for measuring.
(or if you use plugins with great aliasing artifacts)

[24-96 Mastering]

Originally Posted by beanface

No, I'm not bashing them, I just wouldn't personally use them for master processing, thats all - when compared to better converters like prism, lavry etc, they just don't sound as engaging or as detailed.

That's kinda what Mr Velvet meant by bashing, i.e. folks saying they aren't as good as others, whereas he disagrees. I'm not saying he's necessarily right, but it's not like his opinion isn't valid. I know which converters he did the shootout against

[Sugarman]

Thanks for your advice guys!
okay, I want to hook up my DAC1 and ADC1 for mixing/mastering purposes OTB and back into Pro Tools.
So how should I do this?
How do I avoid that the 192 is doing the converting job when sending sound out to the DAC1 and back into the ADC1 and into Pro Tools? I don't want to use the DAC1 and ADC1 for recording. For that I am fine with my 192. So I don't want to make my Benchmark the "master clock".
And....!! At the same time I want to use the DAC1 as my soundcard for my Mac Pro S/PDIF output (iTunes, web etc). This should be setup in a way so I can still use my Danfield Monitor2 controller as the "controller"...
I guess my DAC1 and ADC1 should be hooked up to my patchbay somehow...?

[Audiop]

Originally Posted by Sugarman

It's the argument everyone forget about in the big OTB vs ITB discussions.
There MUST be a degrade in sound, and that should be a good reason for using more plugins (of good quality) and keep it in the box, no?

No, there must not be a "degrade in sound". There will always be some small amount of noise, distortion added and some frequency response alteration but that does not mean it will be audible. These errors can be magnitudes below JND.


/Peter

[dcollins]

Originally Posted by Audiop

No, there must not be a "degrade in sound". There will always be some small amount of noise, distortion added and some frequency response alteration but that does not mean it will be audible. These errors can be magnitudes below JND.

Everyone should hear those Zmix tests. I think they are quite revealing.


DC

[Audiop]

Originally Posted by dcollins

Everyone should hear those Zmix tests. I think they are quite revealing.


DC

Yes, I listened to the piano clip and even though the 60 gen file is clearly audibly different to the original it really does a good job of putting things into perspective.

I have been thinking on a similar test with Lynx Aurora 8 for some time and will try to get that done. Shooting for a 8 gen loop.


/Peter

[bob katz]

Originally Posted by dcollins

Everyone should hear those Zmix tests. I think they are quite revealing.


DC

I'm not familiar with the Zmix tests, can you please provide a link or description?

BK

[First Aid]

post #9 i think Bob.

[bob katz]

Originally Posted by dcollins

That doesn't really make sense to me. And if it was the case, running through your analog chain 20 times it would make it as wide as the great outdoors........... Does that happen?



Speaking of multiple generations of A/D/A, here is 60, yes 60!, passes through a converter that I would say BK is not exactly going to rate as "mastering grade" or give an A++ grade to.

ZMIX Converter Tests

The signal is clearly rendered totally unrecognizable..............


DC

I see. I wish he had done the 60 generations without the normalization in between, as it mixes up the variables. The way I see a multi-generation test is to train your ears to learn the artifacts. The fun part comes whether you can then use your trained ears to recognize the first generation. It's EXTREMELY difficult, with today's grade of converters. On the one hand I claim I can hear it, but when it's blind, it suddenly becomes EXTREMELY difficult. But the only one I have to prove it to is myself since it's my money and my clients I'm serving.

Someday I may be able to afford a pair of Lavry Golds (Is the Weiss, ADA or Prism in that league?) and then I'll have something to shoot out against my HEDD.

BK

[First Aid]

haven't tried the weiss or Lavry but in a shootout between the orpheus, aurora16 and the ADA8XR the ADA stomped them both.

lynx and orpheus were very good (remarkably similar) but every single ref track we played shone in terms of neutrality and imaging on the ADA and was the only clear choice as far as we were concerned (no lavry available to demo at the time). Dual path architechture and scene memorys (for diff samp/wordlengths etc) are very handy too!

[24-96 Mastering]

Originally Posted by bob katz

I see. I wish he had done the 60 generations without the normalization in between, as it mixes up the variables. The way I see a multi-generation test is to train your ears to learn the artifacts. The fun part comes whether you can then use your trained ears to recognize the first generation. It's EXTREMELY difficult, with today's grade of converters. On the one hand I claim I can hear it, but when it's blind, it suddenly becomes EXTREMELY difficult. But the only one I have to prove it to is myself since it's my money and my clients I'm serving.

Someday I may be able to afford a pair of Lavry Golds (Is the Weiss, ADA or Prism in that league?) and then I'll have something to shoot out against my HEDD.

BK

FYI: Both the HEDD192 and Lavry gold AD are included in the samples over at A/D Converter Shootout (with samples)

Some others you may be interested in as well. I will post a list of the 6 models in the shootout as soon as I've made sure it's OK with the distributors that offered demo units.

It's just one piece of music and there's a DA involved, but I think the samples give a good indication of character. And personally, I think that most of those converters can reliably be identifed in a blind ABX under good conditions. The differences are very small, but distinctive.

[Peakly]

Originally Posted by dcollins

Speaking of multiple generations of A/D/A, here is 60, yes 60!, passes through a converter

Amazing. I can't believe it held up like that.

The signal is clearly rendered totally unrecognizable..............

Ha!

Mychal

[mischa janisch]

Originally Posted by bob katz

I see. I wish he had done the 60 generations without the normalization in between, as it mixes up the variables. The way I see a multi-generation test is to train your ears to learn the artifacts. The fun part comes whether you can then use your trained ears to recognize the first generation. It's EXTREMELY difficult, with today's grade of converters. On the one hand I claim I can hear it, but when it's blind, it suddenly becomes EXTREMELY difficult. But the only one I have to prove it to is myself since it's my money and my clients I'm serving.

Someday I may be able to afford a pair of Lavry Golds (Is the Weiss, ADA or Prism in that league?) and then I'll have something to shoot out against my HEDD.

BK

Bob, what you say is true - but have you actually taken the time to listen to this test? Yes, normalization adds to the variables, but it certainly is NOT the first thing that comes to my mind AFTER listening to these files.
The Motu 1296 is definitely a converter I would never ever consider using (not even for recording) - but I have to admit that even under these extemely bad conditions it the harm done is not this huge...

[dcollins]

Originally Posted by mischa janisch

Bob, what you say is true - but have you actually taken the time to listen to this test? Yes, normalization adds to the variables, but it certainly is NOT the first thing that comes to my mind AFTER listening to these files.
The Motu 1296 is definitely a converter I would never ever consider using (not even for recording) - but I have to admit that even under these extemely bad conditions it the harm done is not this huge...

So chances are that one trip through the converters is less damaging than we frequently hear claimed.

So far, I have not heard a single person that was not surprised how little actual coloration there was in Chuck's test..............


DC

[MrVelvet]

Originally Posted by Audiop

Yes, I listened to the piano clip and even though the 60 gen file is clearly audibly different to the original it really does a good job of putting things into perspective.

I have been thinking on a similar test with Lynx Aurora 8 for some time and will try to get that done. Shooting for a 8 gen loop.


/Peter

About perspective - what would a 60th generation tape copy sound like?

[beanface]

Originally Posted by dcollins

So chances are that one trip through the converters is less damaging than we frequently hear claimed.

So far, I have not heard a single person that was not surprised how little actual coloration there was in Chuck's test..............


DC

Yep, I was expecting the results to be a lot more tragic than that.

[mischa janisch]

Originally Posted by dcollins

So chances are that one trip through the converters is less damaging than we frequently hear claimed.

So far, I have not heard a single person that was not surprised how little actual coloration there was in Chuck's test..............


DC

I've been definitely surprised - just in case I have not been clear enough in my last post.

What surprised me even more though, is that Bob seemed to care more about the fact that normalizing was used than the fact that those files sounded SO similar. To me it's like seeing someone breaking the rules of gravity floating 6 feet above the ground for minutes and commenting: "I see. I just wish he wouldn't have worn those spring loaded Nike shoes as they mix up the variables..."

[The MPCist]

If the sound in the box sounds boring, take it out to analog, mess it up and send it back -- no worries about AD/DA conversions necessary unless you're using something like, say, Behringer or mbox...

[24-96 Mastering]

Originally Posted by mischa janisch

I've been definitely surprised - just in case I have not been clear enough in my last post.

What surprised me even more though, is that Bob seemed to care more about the fact that normalizing was used than the fact that those files sounded SO similar.

When Bob was mentioning normalization I thought that he meant it (potentially) aided the signal, as I would otherwise expect cumulative conversion to produce a loss in level, which in 60 generations might have an impact. Maybe I'm thinking too much

I really wish though they had used a drum sample with a bit more highs / lows / a less dense cymbal / room. How the signal holds up is impressive in any case, but I find it hard to 'grasp' the change in sound when the original sample makes me desperately want to reach for the EQ.