Recording Engineers : You molest our recordings !

[PeterSt]

Quite a statement, right ?

 

With a wink to this thread : 24/96 & Above now I've got this for you. For me it has been a most astonishing experience, actually not believing my own ears. But since we were with four of us, we all must trust it is true. Completely true ...

 

I have never been into recording other than for measurements, but since I got the request from my audio pal in loudspeakers to record a base drum in 16/44.1 and in uncompressed fashion, I thought to take two 50 meter XLR runs, and connect them to my FireFace800 on one end, and to two measuring microphones on the other end, positioned at the face of the drum kit upstairs.

For the recording GoldWave was used, and actually that was all there was to it. And a few drum sessions of course.

 

 

 

In between the lines I will admit that I am using playback means which is not available to any of you, and that this *will* matter to the results perceived. But, this is not all that much important, as you will hopefully get from the below.

 

I can be brief about at least one thing : at playing back the first recording we all didn't know what was happening to us. It was by miles and miles and miles the utmost best recording out of the thousands of CDs, DVDAs and hires material otherwise I have. So, undoubtedly this can be sold as the most "representative" recording ever made.

 

But there is more;

There was no one single thing that each of us could complain about the reality of it all. "It is as if the drum kit is in the listening room !".

 

There was no difference in dynamics, color, metalness, and the difference between the 5 toms used were as real as possible, never heard from any other recording. Each cymbal sounded exactly as it really sounds, and yes, the base drum too ... 100%.

 

So here we go. One first stupid 16/44.1 untouched recording - though made by professionals in audio *playback* -, and there's just no single differentiation from reality. None. Zilch. Not that four people could hear.

 

Today a(n audiophile) person came by, not knowing about this recording excercise. I let him hear the drums, next the recording from yesterday, and he said "maybe you know where the differences are, but I sure don't hear them".

 

Now, don't blame this on the playback means. Oh yes, it is of utmost importance, but of course it is about what every other recording engineering department etc., is apparantly doing WRONG.

It also tells that hires is not needed at all; reality = reality. Nothing much more to say. And nothing more needed.

 

It all must be about *some* appliance to at least the production version of a CD, and I don't know what it is. Notice though that this is not about "just shouting" and "are you deaf ?". No. On that matter I guess it just *is* important that one can hear back properly what came from the production. So, without really wanting to say it myself at this time, and from Mani in the thread I winked to :

 

And judging by the sound he's getting from his system (I heard it around 6 months ago), there's a LOT that the rest of us can learn from him (there was nothing at the Munich show this year that even came close IMO).

 

Thus, for now assuming this just is so, no sound engineer is to blame for doing things more or less wrong (what to do if you can't hear it back properly). But still it is wrong. Now what to do ?

 

Anyway, I refuse to believe this is about me using measurement microphones, and otherwise ... start using them guys.

I also refuse it is us and our good recording capabilities (although it should be in the blood somewhere, my late father owning a recording studio). Maybe it's because I used his XLR cabling ? nah ...

 

I mean it, this is not by a small marging ... the difference is completely ridiculous. I mean "real" ... did you get it ??

 

 

 

 

 

May it be of interest, at the bottom I put some pictures of the headroom used. Indeed, not all of the available dynamic range is used. I can tell you, this matters.

... as I always found with the best recordings. They stay well under 0dBFS, and they sound the best.

 

 

 

As you can see here, the normal average volume at a certain SPL for a "commercial" album would be -18dBFS; for our recording here, the output has to be -4.5dBFS in order to create the same average SPL. So, a difference of 13.5dB in average SPL, really telling about the difference in compression.

That there's also a headroom of 1.5dBFS in relation to the loudest sample is another matter (you will believe me without picture).

But it still can't be about this only, because I own numerous CDs with these figures ...

 

Now, who of the sound engineers around here can come up with at least a piece of this puzzle ? Just try something ... dither, close mike only, we record louder, anything ... up to we use software of which we don't exactly know what it does.

It must be something.

 

In the mean time I will do the same for a baby wing I have here, and a violinist if I can find one. Maybe the drums are too easy (although I'd think the contrary). But this latter still won't explain the CDs (including hires) to be so relatively "lousy".

 

Thanks,

Peter

 

 

 

 

 

 

 

 

 

[manisandher]

Hey Peter,

 

Really interesting stuff. I think I've got to come over and take a listen myself at some point.

 

Just a couple of questions:

 

1) Were you upsampling the 16/44.1 to 24/705.6 on playback in XXHE?

2) At what upsampling rate does it stop sounding real?

 

Cheers,

Mani.

 

[PeterSt]

Hi there Mani,

 

1) Were you upsampling the 16/44.1 to 24/705.6 on playback in XXHE?

 

to 32/352.8.

 

2) At what upsampling rate does it stop sounding real?

 

I can imagine your interest here, but I never tried it.

Also, it is of less interest for myself hence the (by accident) subject itself. There is too much different for others to make answers to this useable. Or at this moment anyway.

 

Only few people can come over more than one time. But of course people how buy the XYZ two times always can.

But even without ... we'll have a lot of fun and talks anyway.

 

Regards,

Peter

 

[manisandher]

Hang on, I've got three licences for XXHE (and I intend to buy a NOS1 - probably two). So, you owe me at least 3 more visits.

 

But if your cooking is anything like the first time, I might just pass ;-)

 

BTW, I'm assuming Bert was one of the 'gang'? Please say hi to him next time you see him.

 

Back on topic, would it be possible to try upsampling to just 24/176.4? I'd appreciate this...

 

Mani.

 

[barrows]

Thanks for posting this. Two track, direct to digital, so very little in the signal path, and your environment may be somewhat cleaner than a studio environment from an RFI perspective. This is a little depressing though...

I suggest you post this at one of the Pro Audio Forums, where many of the professional engineers discuss things-it may stir the pot, but I suspect you will also get answers from some of the better engineers as to exactly what is going on.

Anyone here know which pro forum is best, I cannot remember?

 

[cfmsp]

 

 

Thanks for posting this.

 

What "measurement" mics were you using?

 

Barrows,

I can't recall that I ever found the "better" pro recording forum.

 

Clay

 

 

 

[Christopher]

Hi Peter,

 

Thank you for sharing this. Cool kit.

 

Allow me to mention that recording drums is not an easy task. In this area, I have a few hours of experience ; )

 

My initial observation is that your mic set-up is similar to a spaced-pair approach. This approach is kinda common and, indeed, very convenient when you have a limited number of mics.

 

In my experience, some recording studios prefer more control and a lot more mics for drums. Your approach can also be less frequently implemented during live sessions because of space constraints, possible feedback from other equipment and ambient pick-up of audience noise.

 

Question: are you hearing a lot of cymbals...maybe too much...or more than expected? Also, do the drums sound mono-ish...is there a lack of stereo image? If so, mic placement may be the reason.

 

If interested, you may wish to consider adding more mics with some isolation.

 

If more mics are not possible and/or feasible, you can try different mic locations. Attached is a rendering of a two-mic X/Y approach. Please try it...you may be surprised.

 

Just trying to help. Thank you, as always, for sharing.

 

Best regards,

Chris

 

P.S. personal opinion only: If you want serious snare drum snap & snarl with bass drum kick and tom-tom punch, try recording vintage Gretsch drums with real skins. I’m also a fan of modern dw kits. As far as cymbals, vintage Zildjian from the late 60’s and early 70’s are super cool. As I’m sure you are aware, no two kits sound exactly the same. Just my passion and accompanying 2 cents.

 

 

 

[PeterSt]

Hi Chris,

 

Thank you for all your thoughts. But I tried to read them in the context "so how could others fail on it ?". This, because there's nothing much failing here. Nothing actually, although I want to work (just for fun now, because it looks like a nice next hobby) on the best image of it all.

 

So, this is the recording of the drum kit only, which is not much realistic in practice, unless it concerns a drummer only. So, IMO you'd always be stuck with imaging which is in between one of the speakers and the middle, or the drumkit physically (on stage) being in the middle and the image only allowed to be relatively small (never reaching each of the speakers).

The additional fun for me/us was, it totally worked to see exactly where everything was, apart from it being too wide if not careful.

 

The theoretical problem with the measurement mikes is that they are as omni as can be. So, I didn't expect much of it at all. This includes hollownes of the room, even with glass in the roof. Strangely enough (ok, to our surprise) at a further distance all was as dry as could be. Too dry. At the distance you see on the picture it's somewhat too hollow.

Notice that I just compare with being in that room itself, and want the same hollowness. This sure can be done, but it's a fight with the width again. So, the current setup assumes the knowledge of the distance between you and the speakers, and this is how the mikes ended up where they are, for distance to the kit and for the distance in between them.

Image is ok, but sound is too hollow (read : the space the kit is in is perceived larger than reality).

 

But we really shouldn't make a problem of this. It is only fun to get it 100% right, where this time 100% includes the projection of the space. We can't have it all I guess, but somewhat more directional mikes will help. I think.

 

In my experience, some recording studios prefer more control and a lot more mics for drums.

 

Of course. And this may be one of the key reasons why at least the drum kit sounds the best I ever heard on a CD (etc.);

When each device contains its mic the first thing to do is mix them correctly. And miximg means -no matter what and no matter how little- destroying. Still I can't imagine that I don't own CDs which used this same spaced air approach.

 

are you hearing a lot of cymbals...maybe too much...or more than expected?

 

No. It's all just in the normal balance you'd perceive at being in the room itself. The most freaking is the metalness which -so far- always have been a difficulty for me with normal CDs. The metalness exhibits by means of harmonics, and I think I have that quite allright. But the REAL metalness comes from the attack. And I think it is this where the big difference is. So if this is true, it must be about the transients which are preserved better compared to random recordings.

I myself (not being the drummer) hit everything and all by all variations I could imagine not to hear back, but no matter what, it was all there and I saw myself back doing it to the detail (which you sure have forgotton if you're doing this 8 minutes or so without a real "melody").

 

Thanks again,

Peter

 

 

 

[PeterSt]

Clay, simple Behringer EMC8000 (very often used for measuring).

 

But I'm just thinking of something ...

 

I have high sensitive speakers (115dB) for an explicit reason : they are the most easy to drive - or better : move. So, this is (good) for resolution and high transients.

This implies that less sensitive speakers will have less of this ...

 

These measurement microphones are the most sensitive; you actually won't be able to imagine if you didn't use them for measurement, but tiptoeing on a carpeted (30cm concrete) floor, the mic 12 meters away, and it won't go unnoticed.

So, this must imply "hi transient recording".

 

So maybe it's just about using the most sensitive microphones instead of the other way around (which latter I think will be happening most).

 

Hmm ...

 

[I. G.]

You should make binaural recordings with a Kunstkopf! Jaaaaaaaaaaaa!!!

 

 

 

[Christopher]

I think you raise an excellent point; namely...

 

“When each device contains its mic the first thing to do is mix them correctly. And miximg means -no matter what and no matter how little- destroying.”

 

I tend to agree with you. It seems there are pros and cons...compromises and trade-offs...with most approaches.

 

I’m not a recording engineer and I don’t know how to mix. My comments and observations are only based on playing a few kits that were mic’d pretty well.

 

Always interesting and educational to read your posts. Thank you, as always, for sharing. Cheers to your new passion!

 

Best regards,

Chris

 

 

 

 

 

 

[ziggyzack]

I spoke to an engineer who said that the quality of the drummer is a factor. He said if the whole kit is played by a virtuoso, it becomes a single instrument, and a couple of mics may well be all he needs. In other words, he didn't feel he needed to remix Steve Gadd!

 

I'm sure there's no need to remix your cacophony either, Peter

 

[cfmsp]

Some random thoughts:

 

Microphone Data doesn't say much about the Behringer, but it does show output sensitivity, which is average. So, I don't think the microphones are contributing factor here.

 

I think you're on to something with the comment "implies that less sensitive speakers will have less of this..."

 

Perhaps what you're experiencing is an extreme example of the sonic attribute (of Horns and SET amps, etc.) that some refer to as "immediacy", esp as it relates to your completely uncompressed (WRT dynamic range) recording.

 

Compared to the commonly highly compressed music of today, your own recording played via horns is likely to offer sound quality with an order of magnitude improvement irrespective of sampling frequency, IMO.

 

clay

 

PS, presumably you have a "fast" amplifier to drive the horns?

 

[PeterSt]

Hi Clay,

 

Microphone Data doesn't say much about the Behringer, but it does show output sensitivity, which is average. So, I don't think the microphones are contributing factor here.

 

I am not sure, but wouldn't it be so that when the output sensitivity is "small" I just crank up the input sensitivity at the analogue part of the recording device ?

If so, I wouldn't know anything about that sensitivity (as long as I avoid comparing with others). Thus, all we did is adjust the input so there would be a nice digital headroom left.

 

... esp as it relates to your completely uncompressed (WRT dynamic range) recording.

 

Well, may be ... can be ... but it still doesn't tell me why others wouldn't record like this, when it is so obvious that it matters the whole world.

Indirectly though you might come to the same what my suggestion is : once the engineer won't perceive it because of his playback capabilities (of the equipment used of course), then he may think it all doesn't matter anyway.

But it does, and possibly -at this moment- I am the only one who can "see" it.

 

Compared to the commonly highly compressed music of today ...

 

True again. But I don't listen to today's recordings only. There's just no comparison with anything. Now if the difference was marginal ... but it isn't.

 

Yes, my whole system is explicitly built around "speed" because it will imply my 1:1 thinking best. This already starts in the software, of which by now it is commonly known that the more speed (as a derival of latency) the more detail there will be.

Btw, detail seems to compromise emotion, but that's something for a thread two years after today I think.

 

Thanks Clay.

Peter

 

 

[Elprior]

Btw, detail seems to compromise emotion, but that's something for a thread two years after today I think.

 

Funny Peter, I've been having this discussion with Christiaan Punter (from hifi-advice, hope this is allowed to mention it Chris, if not, could you just delete that part ? thanks).

We share your views on this.

 

Elp

 

[Audio_ELF]

Peter ... you (as part of your whole post) post the question "it still doesn't tell me why others wouldn't record like this, when it is so obvious that it matters the whole world" but surely very few engineers will be recording just drums solo, it's part of a whole group of instruments and perhaps its at this point (with everything interacting) that the drums (and other instruments) loose their (for want of a better term) natural characteristics on the recording?

 

There are some recordings where they have been made with a couple of mikes in an acoustic type setting similar to what I imagine you were doing. Have you ever heard any of the Isomike recordings from (IIRC) Ray Kimber? Not sure but my understanding is that these are recorded similarly (though most are 4-channel).

 

I don't know of this is the answer ... just something to think on.

 

Eloise

 

[PeterSt]

Hang on, I've got three licences for XXHE (and I intend to buy a NOS1 - probably two). So, you owe me at least 3 more visits.

 

At first I wrote XXX but thought you would think of XXHE in that case, so I changed it to XYZ. Not that it helped.

But I meant NOS1.

 

But if your cooking is anything like the first time, I might just pass ;-)

 

I had a last line in there "but it is your turn for dinner this time", next asked the family here whether they recall you have something with cooking, and the No made me scratch the line.

 

BTW, I'm assuming Bert was one of the 'gang'?

 

Yes.

 

Back on topic, would it be possible to try upsampling to just 24/176.4? I'd appreciate this...

 

Ok, just spent an hour or so on it, which is because the differences were more and different from what I expected, and they are not easy to analyse. I mean, it appears not easy to put in words a difference compared to reality (which is new for me now).

 

In very general (and don't laugh) ... the colour changes. And no, I don't say this because you observed similar (altough not consistent with what's happening here), but it does anyway. This is the most apparent.

Listening more closely, you can see in your mind that the bended shape (looking from the side) of a cymbal has turned more flat, or maybe it is better to say that you can't see the shape anymore.

Very generally this was with all, knowing that I started off with 176.4 and couldn't recognize so much of the rythm I had in my mind during playing the "piece" myself (there's a take of me too). So, you could say it didn't "work" anymore. But I only noticed that clearly at running it again at 352.8 (when all fell in place again).

 

A most apparent thing - but I had to hop over to 44100 to confirm that - was a lack of dynamics. The attacks were less, and the metall was less. This is what I not quite understand yet, but it may take a beer or two to get there. Btw, 44100 was quite messy, but still better than (more "clear" or so) normal CDs).

 

There is one thing which REALLY puzzles me, and with the knowledge from today most probably a difference with reality in the base;

There are parts hitting a cowbell, and at listening back earlier I already imagined the cowbell to carry something which I didn't hear at it playing myself. If you imagine the snare of the snaredrum to vibrate on the sound of the cowbell you will be pretty close. But, I didn't accept this as a difference, because at playing the cowbell voices towards you (or at least the mouth points at you), while at listening it is just not.

Now, the strange thing is that the in fact unrecognizeable noise is the most profound in 176.4, while it is far less there in 352.8. And the beauty of it is, that it changes with the angle I hit the cowbell, each hit changing the angle more towards horizontal. The effect of it is that with 352.8 some 10 hits without the noise are there, while with 176.4 there are only 2. So, this is clearly an anomaly somewhere.

 

Do we care ? no. Are we going to improve on it ? yes maybe, but the only real solution at hand would be playback at 705.6, as the logical next step. But I'm not sure I want to go that route, because it can't be readily utilized (NOS1), no matter how I wanted that.

 

But of course at this stage it would be more interesting to see what happens with 24 bits to that same "anomaly". Or with 176.4 recording for that matter.

 

Well, I'm glad I found something afterall, because it will tell me later what matters how.

 

So, to be continued.

Peter

 

 

[Tipper]

 

 

A great many years ago me and a group of friends used to put bands on at a local youth club.

(Our greatest moment was putting on AC DC long before they got famous. I think it was their first gig in England)

Anyway, we did record a number of bands (unfortunately not AC DC) using just two microphones that belonged to one of the guys PA system to a Revox reel to reel (I think it was an B77 but tbh I cant remember) on 1/2 inch tape.

The acoustics at the club were far from ideal and none of us had much knowledge about mike placement or recording techniques.

We did record a band called The 64 Spoons that had entered the Melody Maker local band competition that newspaper held back then.

I had a Thorens TD160 back then feeding a good quality valve amp and a set of Rogers LS2’s if my memory serves me.

What I do remember very clearly was how impressed we all were with the sound quality compared to even decent Vinyl we got from our recording efforts played through my stereo system.

As recording moved from 4 track into multi track, and then to digital, the recording engineers have had more and more opportunity to influence the sound of the final mix.

Maybe there is a lot to be said for Peter’s proposition that modern recording techniques involve excessive processing and it kills the original emotion and accuracy of music.

 

 

[barrows]

This might be getting somewhere. See my thread on "The Soul of Music" here:

 

http://www.computeraudiophile.com/content/Soul-Music

 

Perhaps digital editing processes are not nearly as benign as we have been led to believe, and, just like in analog recording/editing, the amount of processing done degrades the music. We accept in playback that different software (even when bit perfect) engines affect the sound, it logically follows that every processing step changes the sound, even if we do not understand why.

Maybe: "direct to digital" should be the digital recording standard that everyone tries to come closest to.

 

[Soundproof]

How true.

 

Tipper and Peter write: Maybe there is a lot to be said for Peter’s proposition that modern recording techniques involve excessive processing and it kills the original emotion and accuracy of music.

 

I spend 80% of my time listening to music that has been recorded live, or that has been recorded as if it was performed live. I am extremely wary of music that's been through the mixing process and even dislike the use of close-miking, unless it's by one of the true recording artists out there.

 

In contrast, look at 2L's process. They spend a lot of time finding the perfect recording venue for the music to be recorded - and now that they've been doing this for quite a while, they have a few to choose from. They do not use dampened studios, because they want the acoustics that a real venue deliver.

 

Then they set up for a high-resolution capture, according to a system they have developed, where the goal is to do as little as possible with the captured signal before release. The processing they carry out is that required to adapt the material for the various formats it is to be distributed through.

You can't call their releases "studio masters" - they are "recording masters" untouched by today's many music post-production tempations.

 

Contrast that with people who want the high-res headroom in order to have leeway for manipulating what they've recorded, before its final release on Redbook - and who then go on to manipulate away like mad, removing any semblance of REAL from the reels.

 

That's a problem, a big problem.

But it doesn't mean that every recording studio is doing this, and it just means that you have to be picky with what you listen to, and what you should expect from it.

 

Every day, as many as a few hundred new live recordings are posted here:

http://www.archive.org/details/etree

 

Sure, you do find duds, but there are thousands of FLAC full-resolution recordings there, of whole sessions, and many of them are stellar. And even those that aren't offer more immediacy than most store bought items.

 

The people who post them have recorded sessions, usually with the agreement of the bands playing, or even because they are tasked with doing so by the venues. And because they know it will never be officially released, they place the music there.

Go hunting, enjoy the show!

 

(Also from archive.org - ISGM releases all its concerts, which have been recorded at their venue. Unfortunately not in 24/96 FLAC, but there are some truly engaging performances there: http://www.archive.org/details/ISGM_Podcast)

 

[CG]

Gosh - two posts in one year for me. What am I doing?

 

Peter has made a great observation.

 

When my daughter was much younger, she played in the elementary school band. I decided to record their concerts, just for us. So, I bought what was one of the better portable DAT machines of the time (a Sony PCM-M1) and some simple and practical stereo microphones (Sonic Studios DSM). These are hardly state-of-the-art, but they were really the right thing for the circumstances.

 

The results were staggering! Even using the built-in playback facility of the DAT machine right into the stereo playback, it was clear that these recordings were better than *any* other recordings we owned. This was not because the venue was ideal, that the microphones were placed optimally, that the equipment was so great, or even because of the musicians (!). It had to be something else.

 

The stage was very clear. The image was perfect. Depth? Tons of it. Little tinkly triangles were as they sounded. Bass drums had oomph but no bloat. I can't think of anything that wasn't very, very realistic.

 

Given the dunce who was running this modest equipment and all, the only differences had to be the lack of processing and the use of simple microphones with only two tracks. Nothing else makes sense.

 

[Soundproof]

If you want to hear fantastic recordings, CG, pick up birdsong releases by some of the well known recording ornithologists.

 

They go into the field with equipment similar to what you describe, and with good microphones, which they will place on stands - record to two channels and do minimal processing.

Fantastic soundscapes that really show off what your loudspeakers can do in the room.

 

However, before we wax lyrical as to how brilliant amateurs are compared to pro's, do remember that a lot of the so-called singers out there have lousy voices that require a lot of post, and that the whole release is often colored by this fact ...

Not many of today's singers who can weather a truly acoustic release.

 

[Miska]

and connect them to my FireFace800 on one end

 

One dictating feature in this case, is the downsampling and SDM->PCM conversion done by the Fireface 800 converter chips in this case, especially if the interface was running at 44.1/16. I don't know what chips RME uses, but I have my guess...

 

Naturally biggest impact coming from the recording space, mic placement, mics and mic preamps. In that order.

 

Have you experimented close miking with mics 5-10 cm under each cymbal or 30 cm above?

 

 

[PeterSt]

No. But I'm not quite sure why I should ?

 

But if I try I may find out.

:-)

 

[Mister Wednesday]

Greetings all,

 

I've been following this thread over the past few days and while I applaud the efforts made in recording a single instrument (the drum set) in a single room environment it seems that the question of why multiple instrument recordings will generally sound different (worse) than this single instrument recording hasn't been addressed.

 

For those of you who may not already know there is a very interesting and well written book on this subject by mastering engineer Bob Katz titled "Mastering Audio, Second Edition: The art and the science" (ISBN-13: 978-0240808376). I am in no way associated with or have any financial interest in this book or Bob Katz's business.

 

One of the more interesting things included with this book(at least the first edition I have) is a reproduction of a chart showing the frequency range of specific commonly used musical instruments so that you can see quite quickly when there will be conflicts in specific frequency spectrums when recording 2 or more instruments. And it's frequency spectrum conflicts like this, amongst other factors during recording, mixing and mastering which Bob Katz describes, which lead recording engineers to make a compromise decision during these processes which might account for the original poster's position on this question.

 

Perhaps recording engineer Bob Diamante from the LIO-8 thread might want to comment on this also?

 

Regards,

 

Mister Wednesday