Help with digital

[Ethan Winer]

Shailat,

> I will do what I can with getting audio tracks for you to listen with a detailed describtion of what I did ... <

Before you go to that much trouble, just tell me how you are keeping all variables the same except the one part that changes between analog and digital. Maybe I'll be able to see something wrong with the test.

--Ethan

 

[chessrock]

Ethan,

It's very nice to have you on this board, and we're lucky to have someone like yourself share your knowlege with us.

But from an objective standpoint, do you have difficulty admitting that some things can't be explained with 00's and 01's? It's just that this whole argument is beginning to look like a "Science versus Religion" arument between an atheistic existentialist and a priest.

"There is no scientific proof of a God" versus "Some things are beyond the realm of scientific proof and logic."

Most of what we hear is an illusion -- our brains interperate what our ears are perceiving to create what we know as sound. Perhaps mathematics says the summing of all the 01's and 00's is going to equal a given sum, so there shouldn't be any difference or loss in fidelity . . bla bla bla bla bla.

Maybe digital summing would sound perfect to a robot or a piece of test gear. But perhaps there's something about the subtle innacuracies of analog summing that our brain finds easier to make sense of -- and thus makes it easier for it to create a truer 3-D landscape for us to perceive. We certainly don't hear perfect, and most of our brains certainly aren't perfect.

 

[Ethan Winer]

Chess,

> But from an objective standpoint, do you have difficulty admitting that some things can't be explained with 00's and 01's? ... "Some things are beyond the realm of scientific proof and logic." <

This is an excellent topic and surely related to the discussion at hand. But I suspect you won't like my answer.

I reject the notion that some things are unknowable. Just because science hasn't yet found an answer, or found the right thing to test, or the best way to test something, does not mean an answer can't be found. To think otherwise is to believe in superstition. We might as well go back to burning witches and reading Tarot cards. Oh wait, some people still do believe in Tarot cards.

More to the point, summing a bunch of audio signals - whether through resistors and op-amps or by adding their equivalent numbers - is extremely basic stuff that has been understood completely for many years. All that's really needed to settle this particular issue is a scientific double-blind test. This can't be difficult to do!

> But perhaps there's something about the subtle innacuracies of analog summing that our brain finds easier to make sense of -- and thus makes it easier for it to create a truer 3-D landscape for us to perceive. <

Nah. An audible sense of space and depth has nothing to do with simple distortion, which is really what the differences between analog and digital boil down to. Depth is related to the time and phase relationship between the left and right channels. And depth in a mono recording is related to ambience and room tone captured by the mikes. Again, this is very basic stuff that is well understood.

Further, the theory put forth here is that digital summing is somehow "thin." To me, thin is an EQ issue. And that is very easy to test and measure.

--Ethan

 

[Shailat]

Shailat,

> I will do what I can with getting audio tracks for you to listen with a detailed describtion of what I did ... <

Before you go to that much trouble, just tell me how you are keeping all variables the same except the one part that changes between analog and digital. Maybe I'll be able to see something wrong with the test.

--Ethan

I was thinking of running a mix that was recorded on the computer to a DAT and then use the same outputs to ADAT to DAT
and then try - mixer straight to DAT. Then possibly both to computer to burn a wav.

The adats are 16 bit and the converters are no better (and actully worse) then the ones on most audio cards.

If you have a better idea then tell me.


Sjoko,
Do you mean that even though the fader seems to not effect the output to HD it actully does? is that what you are saying?.

 

[sjoko2]

Shailat - lets put it in the easiest way: "therein lies the summing problem" This is not just a problem with Pro Tools, most DAW software has a same or similar problem. When tracking with a DAW, always consider it as NOT having a level control option at all, and you'll be fine.

I first discovered this problem years ago when I was wondering why one session I had recorded using a Neve Legend console as a front end, going through good A/D converters straight into Pro Tools, sounded great (and I mean.... GREAT), and a subsequent session I did using the same gear, same artist, sounded considerably less.
The only difference was that in the first session I had not touched the faders.
A subsequent test proved this to be the definite problem / answer.
Please note that, yes, the summing problem does exist. Don't overestimate it though. If, as I had said to many times, you use accurate clock and good converters, you have won the biggest part of the battle.

Personally, I am getter better results now using digital than I've ever done in 30 years of all analogue recording. That includes recording acoustic instruments, classical instruments and killer kick-ass drums.
Sadly, some of the strongest proponents of analogue are people who have not (as yet) come to grasp with the peculiarities of digital. Sadly and funny, some of the loudest proponents of the "analogue is better" fraternity I have found to be incapable of presenting a well calibrated and balanced console / tape combo....... It has amazed me over the last years how many sessions I have had in different studios around the globe where my first task just had to be calibration??

Ethan - I read you are using SAW. Good choice. Pitty what happened recently, I hope someone will keep the platform running and developing.

 

[Lt. Bob]

digital versus analog

ya know--it kinda depends on if you're using 16-bit 44.1 sampling or a higher medium. If you're using 16-bit/44.1 then sometimes analog can be better. Note that I said SOMETIMES. It's a fact that 16-bit cuts off reverb tails, giving them an unnatural quality. Also with the brick wall filters, there are documented problems with phase problems and other artifacts. Also there is a substantial amount of study that seems to show that sounds above 20k have an effect on how we hear sounds in the audio band even though the ultrasonics are not themselves audible. So the 20,000 hz limit on 44.1 can in some cases affect the way we hear sounds in the lower range. Also 16-bit has less dynamic range than good analog. That being said, it's clear that the higher resolution formats (i.e. 24/96k) are more accurate than even the best analog. It's only 16-bit/44.1 that has the limited low level resolution.

 

[Shailat]

Sjoko,
The faders automaticly jump to unity gain when recording if left untouched but how do you monitor during recording if you want to achieve a balance between the different parts of the kit?.

That was my main problem. How can I make decisions with out balancing the kit during monitoring/recording?

 

[kristian]

have a card that has 8 or more D/A and run a mix that in the board. send it to your headphone AUX and voila!

atleast thats what logic would tell me, i don't have protool$

 

[Shailat]

No.
Since I was going direct to computer, I wanted to hear whats going direct to PT with out any other gear to intervine.

I wanted to hear the output from the main out.

 

[sjoko2]

Sjoko,
The faders automaticly jump to unity gain when recording if left untouched but how do you monitor during recording if you want to achieve a balance between the different parts of the kit?.

That was my main problem. How can I make decisions with out balancing the kit during monitoring/recording?

Very simple, to set up a cue mix in pro tools create a stereo aux master and use the aux send faders on each channel to balance your cue mix. This way you can create a whole bunch of independent cue mixes.
You can also make discrete cues by choosing pre-fader stereo sends.

That's the pro tools 'internal' way. There are other options dependent on the type of gear you use. Another way I sometimes use is to create a cue with my converters, as I have MIDI control over the D/A sends.

 

[Shailat]

Very simple, to set up a cue mix in pro tools create a stereo aux master and use the aux send faders on each channel to balance your cue mix. This way you can create a whole bunch of independent cue mixes.
You can also make discrete cues by choosing pre-fader stereo sends.


No no no.....this is not my first day with a PT rig....
I know how to set up a monitor mix.....what I wanted to acheive
is the most cleanest purist signal path and hear the kit with out bussing it to other places as I heard the problem, I didn't want to make any other busings to an aux not knowing if that would effect the signal even more....
I checked the main out and a monitor out of one of the cards and both were different and that put me even more on hold.
The internal way was what botherd me.
a



b


That's the pro tools 'internal' way. There are other options dependent on the type of gear you use. Another way I sometimes use is to create a cue with my converters, as I have MIDI control over the D/A sends.

 

[Ethan Winer]

Shailat,

> I was thinking of running a mix that was recorded on the computer to a DAT and then use the same outputs to ADAT to DAT
and then try - mixer straight to DAT. Then possibly both to computer to burn a wav. <

I guess that makes sense, assuming you have a sound card that can send all of the tracks separately from the computer to the ADAT. I also assume the tracks have any processing already in place, so the only difference is that the computer mix goes to a DAT as a two-track feed, and the ADAT mix goes through the mixer only and not to any other outboard gear.

The key is to have both scenarios the same except the part where the tracks get combined. So you'll also want to set all the track levels to exactly zero before exporting them to the ADAT.

--Ethan

 

[Ethan Winer]

Sjoko,

> Personally, I am getter better results now using digital than I've ever done in 30 years of all analogue recording. <

Yes, that's been my experience too.

> I read you are using SAW. Good choice. Pitty what happened recently, I hope someone will keep the platform running and developing. <

Yet another reason to never buy software that's copy protected. If SAW was protected, with the next computer I buy I'd never be able to install it again or even load any of my past projects.

I recently switched from SAW to Sonar, though I still use SAW for some types of work. But buggy as Sonar is, it has some really nice features that have changed the way I work for the better.

--Ethan

 

[Ethan Winer]

Re: digital versus analog

Bob,

> It's a fact that 16-bit cuts off reverb tails, giving them an unnatural quality. <

I've never experienced that with normal program material recorded at normal levels. Yes, when a reverb tail dies down to below -80 dB. it might sound affected. But in most cases room noise or a synth's electronic noise are louder than that making the lack of bits insignificant. If you have to raise the volume to 40 dB. above normal to hear a problem with reverb tails as they fade out, I'd say that's not really a problem.

> Also with the brick wall filters, there are documented problems with phase problems and other artifacts. <

I've been trying to find such documentation for years. I read all the time in Electronic Musician and other such magazines how phase shift is so terrible. But I've never experienced a problem with phase shift, or seen any scientific tests that show phase shift per se is bad. Remember, phase shift occurs in nature all the time. If you're standing on front of your loudspeakers and move your head an inch to the left or right, you've just introduced a healthy amount of phase shift at high frequencies.

> Also there is a substantial amount of study that seems to show that sounds above 20k have an effect on how we hear sounds in the audio band even though the ultrasonics are not themselves audible. <

Again, the only places I've seen that are the pages of hobbyist audio magazines like EM. It is possible for very loud ultrasonic sounds to cause a nonlinearity in your ears, such that you hear sum and difference tones. But aside from putting your ear right next to a very loud glockenspiel, this is not likely to be an issue.

--Ethan

 

[Shailat]

OK

I said I would try to do some tests and so I did.

Before I explain what I heard, let me say this-
I have no interest in bashing heads oon the matter as I know what I hear and no matter if 20000 people would say to me otherwise, I know what I am hearing.
So.......take it as you like.
It wasnt a sientific lab conditioned test. I did what I could with the little time I had.

I finished recording a session of drums to a computer that consisted of the Digi 001 system. I then added to it a synth bass and a midi piano for the test.
Then I did the following things -

1. I sent it to the mixer using the direct outputs from the 001
to the board and mixed it straight to DAT. This way I had the same computer tracks going into the board as the computer tracks but summed in the mixer.

2. I mixed it again inside the computer making sure the mix was identical in every possible way. Dumped it in to the same DAT

I A/B the two mix's to make sure they sound identical and checked it with two people to be sure they could not tell the difference between mix A and mix B.

For the record...these people are excellent Engineers.
One is a friend who has been one of the top engineers in my country and is a leading world expert in modifing vintage pres/compressors/and modifing mixers.
The second person aside from me was a assistant engineer with a wonderfull set of ears.

I then "Bounced" the mix in the computer to a single stereo wav file.

So now I had 3 mix's
1. Computer mix
2. Board - mix
3. Bounced mix - (in tweakhead mode)

Made sure they were all outputing the same level and A/B/C'ed them one after another jumping from one to the next.

Conclusions in short hand were -
The bounced mix had suffered from the bounce. Some of the proportions were lost as the harmonic content had been changed.
In short.....it sucked

The computer mix had also changed the harmonic content. It wasnt as deep and it made for example the Bass lose some of its resolution.

The most natural was the board mix where the snare was like a snare should sound !

Then I normalized all three mix's and gave it a listen and that was a DISASTAIR !!!!. All three mix's suffered greatly !.

* A side note - The monitor outs of the 001 are bad news.
I used the Main outs and difference was night and day between the two.

This was all done using only 8 tracks of a acoustic drum kit and 2 stereo midi piano tracks with a mono midi bass.
I have NO doubt in my mind that the more acoustic intstruments I would load on to the mix to be summed, would create a bigger difference between a board mix and a computer mix.

I asked the other engineer to please come and tell you all the results as I am sure some of you are doubting me as I am biased,
Yet he can not write and express himself in english very well.
He offered that anybody who would like to hear his opinion could email him and he will try to reply in Broken english

I did this for myself and dont want to preach or knock any path choosen by others. I did record all three after the normalizing to disk but I wont convert them to mp3 for understandable reasons.
If anybody wants to hear it let them send me a CD and I'll burn it for them.

I hope next week to do the same test on the TDM PT to see if it will make a difference.
Of course I must add that I used a excellent mixer and not the typical mackie homerecording mixer. That is a factor.


My for the time being conclusion is to mix on the board and leave computer mixing to others.

 

[Ethan Winer]

Re: OK

Shailat,

> I A/B the two mix's to make sure they sound identical and checked it with two people to be sure they could not tell the difference between mix A and mix B. <

So you're saying that the mixes were in fact identical, right?

> I then "Bounced" the mix in the computer to a single stereo wav file. <

You mean you recorded a stereo Wave file from the output of the stereo DAT? If so, did you use a SPDIF or analog connection?

> The bounced mix had suffered from the bounce. <

This is the part I'm interested in. Especially if it was just a straight two-track transfer.

> * A side note - The monitor outs of the 001 are bad news.
I used the Main outs and difference was night and day between the two. <

I don't have Digi hardware so I can't speak to its quality. When I record and play through my Delta 66 it sounds the same in every way.

Thanks for taking the time to do that test, and report here the results.

--Ethan

 

[Shailat]

Re: Re: OK

Shailat,

> I A/B the two mix's to make sure they sound identical and checked it with two people to be sure they could not tell the difference between mix A and mix B. <

So you're saying that the mixes were in fact identical, right?

Yes.

> I then "Bounced" the mix in the computer to a single stereo wav file. <

You mean you recorded a stereo Wave file from the output of the stereo DAT? If so, did you use a SPDIF or analog connection?

The computer mix (not the bounced) and the board mix were put on the dat via analog. The computers main outs L&R and the board as well analog to DAT then I transfered them back to the computer using SPDIF so I could A/B them with ease....
Not extremly lab conditions but it gave me a close enough feel.

> The bounced mix had suffered from the bounce. <

This is the part I'm interested in. Especially if it was just a straight two-track transfer.

This part was amazingly nasty.... I took the 11 tracks and bounced them to a single wav file using the tweakhead option.
The processing here was a huge dissapointment for me.
I use in the TDM PT the same option but I have read that it might be better to use the second level of processing then the tweakhead -first option.

> * A side note - The monitor outs of the 001 are bad news.
I used the Main outs and difference was night and day between the two.
<

I don't have Digi hardware so I can't speak to its quality. When I record and play through my Delta 66 it sounds the same in every way.

I suspect the fact that the monitor outs have the option of a volume knob making it go through that would alter the sound....

Thanks for taking the time to do that test, and report here the results.

--Ethan

No problem. I commend you in your civil way of disscussion.
You keep your cool and talk with respect to your opposers.

I can make a cd of the 3 files after normalizing if you want, and I can try to send it to you some time in the near future.

It was fun and informative. If I find some time I'm going to do it better with a TDM PT and maybe another DAW. I did hear the problems with a Paris but that was far from being lab conditions.

Ethan if you wish to ask the other engineer questions he would be glad to discuss it with you as well.

 

[sjoko2]

Couple of observations for what its worth, because here we have entered the surreal world of digital audio..........

First of all, let me just outline "where I'm coming from" to avoid confusion.
In the first place I do record all digital by choice because, like I said before, I can now achieve better results - meaning higher quality audio. If that was not the case, I'd be working analogue, without question.

Second, and something often forgotten, there is one common element to both digital and analogue, which is that quality of equipment does matter. Easy to say "its all just data so it shouldn't make a difference", but it does, in a big way.

Shailat's test results are no surprise at all. HOWEVER, can you contribute the quality problems to a DAW? No, I'm sorry, you cannot.

Obtaining (and maintaining) high quality sound is dependent on attention to detail throughout the chain. This has always been the case for analogue, but its even MORE so in the digital domain. If an analogue signal passes through an element in the chain that might not be high quality, you can often live with the result as it can be masked by the "ear friendly" effects of analogue. With digital? No such luck. Damaged data is damage data, unwanted artifacts are there to stay, eventough putting damaged digital data through an analogue component might sometimes mask the errors.

What will have effected Shailat's tests? In no particular order:

- Conversions - Digi 001 conversions suck, I'm not hammering the product, its cool for what it is and what you pay for it - it costs less than 1 really good 2 channel converter, so you cannot expect high quality.

- Consider that the above goes for D/A as well as A/D conversions, the more conversons, the worse the signal.

- Sending the signals recorded through an 001 to an analogue board for mixdown will have increased the material's "ear friendlyness" as a reasonable quality board would result in "smoothing out" the rough sound of the 001's converters.

- The internal clock of Pro Tools is one thing only, diabolical. This has a very large effect on audio quality. Use a good clock and its almost like switching from a cassette to a CD.

- The "tweakhead" thing in Pro Tools must have been put in there as a joke, its the worse possible.

- Bouncing within Pro Tools. I have not gone into the details of why and how, but its not good, and I never do it. The same goes for every DAW I have ever used. All I can think is that it, once again, relates to clock accuracy, and this time the interrelation between a synchronisation clock and the computer's internal operating clocks (non-audio). Personally I use a different system of bouncing, which results in a large quality improvement over "internal bounces". I export the mix of the "to be bounced" material AES to a high quality sample rate converter (which recieves its clock direct from the master clock) and import it back into the DAW. This way the integrity of the files (and sound quality) is maintained 100%. Please note that this goes for bounces without conversion as well as sample rate conversions.

(Please note therefore, anyone who reads this who records and mixes on a DAW.......... do your mix and cut a disk of the mix from there...... do NOT bounce to a stereo file first)

- Digital cabling. I'm not saying its the case in these tests ('coz Shai knows his stuff), but even in pro studios I keep seeing people use the wrong value / kind of cable for digital applications.
S/PDIF has, at least for me, proven to be the most sensitive format I have yet encountered. I thought we already had the best S/PDIF cables, until someone hooked up some silvercore ones and I heard an immediate and surprisingly big difference in overall quality and extended response. Amazing ... and absolutely worth spending 100 bucks a cable on.

In conclusion, its back to the old 'you get what you pay for' routine, as well as knowledge build through experementing with what is still a relatively new and often missunderstood medium.
Had Shailat's experiments been conducted using the same DAW software, but surrounded by quality digital components, the one he found best would have been the worse, by a large margin.

 

[Ethan Winer]

Re: Re: Re: OK

Shailat,

I have to admit I'm more confused than ever. The first part of your last message said you mixed once within the computer and then once again through separate digital direct outs through an analog mixer. You said these two mixes were identical. This has been my point all along, that unless something is actually broken, there should be no meaningful difference.

Then you refer to a "bounced" mix which I don't understand:

> This part was amazingly nasty.... I took the 11 tracks and bounced them to a single wav file using the tweakhead option. <

To me, a bounce is a one-to-one transfer. You bounce a stereo mix from here to there. But how do you "bounce" 11 tracks to a single wave file without mixing them? Or do you mean all 11 tracks were in series, one after the other? Also, I have no idea what the "tweakhead option" means.

> I commend you in your civil way of disscussion. You keep your cool and talk with respect to your opposers. <

It's not even "opposers." We're having a professional discussion. Like many others here, I hang out to help where I can, and also to learn. The name calling I sometimes see is so "12-years-old" and I have no interest in that.

--Ethan

 

[Shailat]

First sjoko -

You have some points there....
I used the 001 as it was what a client had brought in to record and I took it from there.
BUT ! - I had heard some of the problems on a creamwere Pulsur/tripledat as well and a Paris system. Again - not as accuretly but it would probably bring the same results.

I use a PT TDM system for my self and I hope to try the same test with taking better care on the whole process...when I can find the time This next couple of weeks I'm in a session with PT running the show, maybe I'll find the time to do it again.

I had about a year ago tested the difference between the 001 converters and a 16 bit ADAT and a DA 30 converters and found them to be as good as the DA but better then the ADATS.
They are not up to par with the higher end converters in the market but they don't "suck" IMHO.....

Also the SPIDF cables I use are of high quality.

Every thing else you said I'll have to agree on. From the board changing the sound to the iternal clock (the worst) to the bouncing crap.
A huge note for all homerecorders would be to A/B their bouncing software compared to a mixdown to a different format like a DAT or Masterlink or what every they have.....they might ge a huge surprise..

As to the tweakhead....what do you use? the "best"