What gear should I buy?

SchoepsCMC641Four

Probably, the most posted topic in any technical forum.

It’s really an important question that, unfortunately, is hard to answer objectively. Most of the time, you have to take the decision blindly because is usually hard, if not impossible, to try the gear before buying. 

I am not going to talk about the objective (technical) reasons that should guide a purchase. I leave that to other discussion. I fact you can read more about this angle in the rest of the blog. 

So, what can we trust? We should do like Indy in his third movie, walk over the bridge although we don’t see it. That is, try to add several subjective opinions hoping that, at the end of the process, will emerge some kind of objectivity. The problem can be worst when the buyer can’t tell the difference between products because his/her untrained ear or inexperience.

So we can find two, not very objective, reasons to decide a buy: product popularity and product placement in professional contexts and professionals. They are, respectivelly the quantity and quality reasons.

In the first case, is usual to apply the following logic: if lots of people have it, it should be good. And this, more or less, can be true. But in a capitalist market full of oligopolies this idea can be dangerous. Brands can be very strong and what genuine quality doesn’t achieve, can be achieve with marketing.

This is now even more relevant because the companies are merging and buying each other. We must assume that Neumann is not anymore a company, but a Brand (Sennheiser bought it in 1991). Or KRK which is now far from the genius of the engineer who give the brand its name (Keith Klawitter) since Stanton bought it. 

krkvxt01_l

Surely, this doesn’t means that this products cant’ be great. But in a certain way we are taking the buying decision based on the past quality, which may not be the same now. An this can be even worst. Everyone wants an U87 or an U67 and, for sure, they are great mics, but old schoolers will argue that their vintage models sound way better. So, are we basing our purchase in a prestige and quality that can be, partially, in the past?

Of course, companies are well aware of all this and they will always use the classical brands names. So, the key, resides in how much the company respects the brand legacy. Some times this respect is missing and we can usually find an economical shift form the manufacturing process to the marketing department. 

Another example of this, is Pro Tools dominance. It’s a great software, no doubt about ut, but also Logic, Cubase and many other are very good options. So, is objectively better? In my opinion every DAW does basically the same and the best would be the one you are used to use. Maybe because your own workflow has grown around it.

But Avid (previously Digidesign) knows very well that they are the industry standard and the use this position to keep the prices high, specially in the HD product line. 

The other tool we have to decide a purchase, which is even more relevant, is the quality of the people that uses it. We reads an interview to a respected engineer and we are hoping that they will ask him about what gear he uses. If he mention some of ours, we can’t avoid a little smile. We are fine. This guy with 2 oscars use the same pre-amp that I do. 

Companies take advantage of this and they spend a lot of money in endorsements with the best professionals possible.  

This last quality factor can be a good reference but we should not forget that is very important to know how to use the equipment properly. In my opinion, this can be even more important that the gear’s intrinsic quality. If you know what sound you want to achieve and your tools, I think this is the differentiating factor. Poor man’s consolation? Maybe.

Before going, I want to mention a different case. Line Audio is a small (garage small) company form sweden. Its existence has passed from person to person, almost like a secret. Without advertising or a strong distribution, their sales have gone up to the point that the owner and manufacturer had to stop taking orders last christmas.  

IMG_0650

Without talking much about the sound quality (wich I often hear is great) the surprising thing is the prices. The CM3, for example, is a small-diaphragm microphone popular in classical music recording and sound design. Without taxes and shipments it is worth 120€. A equivalent microphone from Neumann or Schoeps would be around 1000€. In this thousand € how much is paying the design, components and fabrication if a guy in a garage makes something similar for the tenth of the price? Is the price so artificially inflated like whan we buy a pair of Nikes? Maybe is a mistake to put the CM3 and a Schoeps in the same ground, specially when I haven’t try the microphone. But, you know, I have two subjective tools. A quantity one and quality one (I and II)

 

 

 

Make audio sound far

In the daily work in sound post and design we often face this situation: You want to introduce a library or foley sound in a scene and, although, the sound itself works it doen’t fit the scene context, it feels unnatural. 

The problem lies in the acoustic positioning. We often try to fit in the scene a sound which is recorded in a different acoustic plane. In many occasions (and, in fact, this is the ideal situation) the audio we are dealing with is dry and sounds “in the face”. This kind of sound is maybe not the best to give credibility in a determined context. Especially when we want to position it behind the foreground.

If, otherwise, the sound you want to use includes certain reverb or some “perspective” and distance, getting rid of this vibe is quite difficult, maybe impossible. In this case, one option would be to use some reverb removal plugin. This may never work nice but if the case is not very extreme it might save lots of work. I find one of this plugins is especially effective in removing not only reverb but also some “out-of-focus” quality. I’m talking about “Univeil” by the Garman company Zynaptiq. In certain applications the plugin honors the company motto: “Science, not fiction.”

The truth is that making things sound far is tricky and there isn’t a magical formula. Every case is particular and would require specific technics. In indoors scenes it might be a little easier because adding reverb would give us a great head start. But this is usually not enough, if we want the effect (or voice) to really feel coming from the background we must apply different processes carefully. 

Let’s see some of this technics:

The romans combined the architecture with acoustics to make the actor’s voice feel closer. (Mérida Roman Theater, Spain).

Reverberation:
Is kind intuitive. If something is farther it has more reverb. Or being more precise the direct sound/reflected sound ratio is lower. That being said, we must be very careful with the amount of reverb. Too much may sound artificial and over processed, especially in outdoor scenes. 

An interesting strategy would be equalizing the reverb return trying to attenuate the frequencies that can reveal the effect. The mids zone (800 – 2000 Hz) is a good candidate for this. 

On the other hand, convolution reverbs like Altiverb can use outdoor IR’s which can sound quite real and natural. If the plugin allow us to differentiate between direct signal, early reflections and tail we can play around with these values to find what we are looking for. The early reflections can be especially helpful. 

One last tip is inserting an expander before the reverb. We then adjust it so that only the louder events could make it to the reverb. This way we still have the “wet” flavour but in a much cleaner way. 

Stereo Image:
An audio with a stereo nature (like music) would sound almost mono played from a far distance. It doesn’t make sense to perfectly hear and difference the L and R channels if the sound is diegetic and coming from the background. So, closing the stereo image or going straight to mono is going to help a lot if we are introducing in the soundtrack a stereo sound. 

Equalization:
The air (or any other material) absorbs the sound energy proportionally to the frequency. For that reason a LPF would be helpful to create the feel of farness. A daily example of this is when our kind neighbour is listening to music and we hear it from the other side of the wall. That being said we must be careful with this, the air is far less efficient absorbing sound energy as a brick wall is. Use with moderation or it will sound muddy.

If we study the Fletcher and Munson curves we can see that the 1-5 KHz area is the more sensitive for the human ear. Thus, when we are equalizing, enhancing this range could work very good, especially attenuating at the same time the highs. 

On the other hand, the low-mid region is very important to give an out-of-focus or “roomy” feel. This would be the kind of sound that every good boom operator would try to avoid and maybe in this case we are trying to achieve.

Also, we must try to avoid that the lows equalization give the viewer a Proximity Effect feeling which will not help at all.

Compression:
Changing the signal dynamics is another good technic we can use. Far away sounds are less dynamic and they feel flatter. Moreover, we can use a short attack to quit punch to the signal making the effect even better.

Delay:
Sometimes the use of delay works very well but in some other occasions it feels unnatural. We will use the Haas Effect, taking advantage of the way that the brain localize sounds.

I usually use a short delay (20-70 ms) with some feedback (0-40%). This way, we create kind of a controlled reverb that can create the illusion of sound going farther. This technique give me great results introducing in a scene TV or radio sounds.

Edition:
When someone talks to you from a far distance part of the intelligibility is lost. This is because we stop hearing little details like mouth clicks, esses or letters t, k, c, etc. If we edit the speech removing this details we can help the sound a lot in feeling farther. This concept is also applicable in the little noises of, for example, a foley movement. 

Well, I’m contradicting myself. If we use the compressor like I explained before we are going to make these little sounds more audible. So a way to go would be editing first and then applying compression.

Old Style: 
When nothing works we must come back to the old tricks. We can reproduce the sound we are working with in the studio monitors and record it from any place in the room. Playing around with the microphone setup can lead to different timbres and colors. This technique is sometimes very convincing. It has been used a lot in post studios that usually have an echo chamber for this purpose. 

Echo chamber in Abbey Road. Notice the tile wall for more reflection and the columns for diffusion.

These are the technics that I use in the daily work. I hope they are useful for you. If you happen to know others, please leave it in the comments.

Using shotguns microphones indoors

Chinese (中国的) version: http://107cine.com/stream/9197/

I have been researching an idea that I have been hearing for a while:

It’s not a good idea to use a shotgun microphone indoor.

I want to check this question out and, eventually, make some tests. Here we go!

 

Shotgun microphones

The main goal to these devices is to enhance the axis captation and to attenuate the sound coming form the sides. In other words, make the mic the more directional as possible in order to avoid unwanted noise and ambience.

To achieve this, the system cancels unwished side signals delaying them. So, the operating principle is based in phase cancellation. At first the system had a series of tubes with different sizes that allows the axis signals to arrive in the same time but forces the off-axis signals to arrive delayed. This idea, created by the prolific Harry Olson eventually evolved in the modern shotgun. 

In the Olson’s original design, in order to keep improving directivity you had to add more a more tubes making the mic too big and heavy to be practical. To solve this, the design changed in to a single tube with lots of slots that behave in an equivalent manner to the old tubes. These slots make the off-axis sound hit the diaphragm delayed, so when it is combined with direct sound a noise cancellation occurs. 

This system has its limitations. The tube needs to be long if we want to cancel low frequencies. A typical 30 cm (12″) mic would start behaving like a cardioid (with a rear lobe) under 1,413 Hz. If we want to go lower the mic would have to become too big and heavy. 

Electro Voice 643

Electro Voice 643, a 2 meters beast

On the other hand making the mic longer makes the captation angle narrower, so the more directive the mic is, the more important is a correct axis alignement. The phase cancelation principle also brings consequences like comb filtering and undesirable coloration when we go off axis. This can work against us when is hard to keep the mic in place. 

In this 416 simplified polar pattern we can appreciate the directional highs (in red) curling in the sides. The mids (in blue) show a behavior somewhere between the highs and a typical cardioid pattern (in green) with a rear lobe.

Shotgun Polar Pattern

In this other pattern we can see the general shotgun response. The side irregularities and the rear lobe are the consequence of the interference system.

Indoor usage

The multiple reflections in a reverberant space, specially early reflections, color the signal. Ideally, the mic, depending of the incidence angle will determine if the sound is relevant (signal) or just noise. When both the signal and the noise get reflexed by nearby surfaces they enter the mic in “unnatural” angles (If we consider natural the direct sound trajectory). The noise then is not properly cancelled since it does not get correctly identified as actual noise. Moreover, part of the useful signal will be cancel, because it is identified as noise.

For that reason a good acoustic treatment is a great way to improve the sound of a scene, specially with these mics.

 

Another aspect to have in mind is the rear lobe that these mics have. Like we saw earlier this lobe captures specially lows so, again, a bad sounding room is something we want to avoid with a shotgun mic. When we have a low ceiling in a set we are sometimes forced to keep the mic very close to it so the rear lobe and the proximity effect combines and can make the mic sound nasty. This is not a problem in a professional movie set where you have high ceilings and good acoustics. In fact, shotgun mics are a popular choice in this places. 

Last, the shotgun size can be problematic to handle in small places, specially when we want precision in the axis. 

The alternative

So, for indoors the best options would be a pencil hipercardioid microphone. In one hand it eliminates the size problem (they are quite smaller) and they are more forgiving in the axis placement. Moreover, they don’t hace interference tube, so we don’t get unwanted colorations from the room reflections. 

         Pencil and Shotgun mics polar pattern comparision

Is worth to say that these mics have a rear lobe too but not so pronounced. Also we need to keep them closer to the source than a shotgun.

Despite that, hipercardioid pencil mics are a great choice for indoor captation.

In a next post I will make an A/B comparative between the 416 and the Oktava MK-012 to try to illustrate this.

Working with video in your DAW

Lately I have been working in Post-Production for some short films. Sometimes the video behavior wasn’t the best (sinchrony errors, bad screen refreshing…) and in another moments the videos wasn’t very CPU friendly. So I started a little research about codecs, formats and other stuff to optimize the audio work with video reference.

This is what I have learned. 

Intel:

First, we need to get all the info out of the video we have. We can do this in Quicktime. Window>Show movie inspector. It will open a window where we can see al the data related to the video.

Now we know what we are dealing with. So, which parameters are the best to work with?

Format:

I would recommend .mov. In some cases .mp4 may also work but in may experience can cause problems. 

If you are streaming the video via firewire to a DV Box, DV PAL/NTCS would be, obviously the ideal format. You would save CPU power and improve the screen refreshment rate. Here are two DV Box model if you are interested: “Canopus ADVC-100″ and “Sony DVMC DA2″. That said, using DV format without an external device doesn’t seem very desirable (I haven’t tried this).

 

Codec:

If we are working with “native” video I would recommend Photo-JPEG. It looks fine, behaves good and is light to the CPU.
I wouldn’t recommend H.264 for being CPU consuming and cause problems with Pro Tools.

Size:

Using a small size is a great way to save same space and resources, but you have to be careful with the quality. If you make the video too small you may lose some crucial details for your audio work. Something like 640×480 should be fine.

 Audio:

The audio track is going to be just a reference so you could compress it if you want. Personally, I leave it uncompressed. I just change the sampling rate and bit depth to match my session.

Frame-Rate:

When you convert the video, leave the FPS the same. Then, in the DAW select the same FPS so that the time code keeps in sync.

Software:

Well, where I convert the files? A good option is Quicktime itself. The brand new Quicktime X doesn’t have anymore cool exporting capabilities so we have to switch to Quicktime 7. You can find it in the Apple website. When you try yo install it, it will tell you that you already have a more recent Quicktime version and close the installation. Playing around with files contained in the installer I have managed to make it work without loosing Quicktime X. I don’t remember exactly how I did it but if you are interested, let me know and I will explain it.

If Quicktime fails (sometimes it doesn’t behave well with certain codecs) I would recommend MPEG Streamclip as backup.

Finally, to improve the performance of both Quicktime and MPEG I would suggest to intall Perian, which add more codec support to our system.

With this advices you should have Pro Tools or Logic working fine and smoothly with video. I will appreciate your own advices, suggestions, critics or ideas. Thanks!

 

Test: Pre-amp gain levels

Spanish Version
Versione Italiana. (Thanks to Mirko Perri)

With this test I want to try to solve a doubt that I sometimes have when I record.

It’s pretty obvious that when you turn up the gain you are also turning up the noise. But in the other hand you also pick more useful signal. So, the question is:

 Does the Signal/Noise ratio keeps constant as you rise the gain?

When you record you always tries to avoid signal peaks which could saturate the pre-amp. Then, should you force the pre-amp in order to get some dB’s more? Or it’s better to get those dB in post-production? 

Instead of making theoretical analysis I would try to dispel my doubts with some tests.

Crank up the volumen!

Test 1: Room Tone

I’ve used the Tascam HD-P2 with the Sennheiser MKH 416 recording with differents gains and the matching levels with Pro-Tools.

To my surprise the biggest difference is in the lowest gain take (the first). The sound has a clear 12 Khz tone that can be also heard (barely) in the second take and it’s indistinguishable in the other takes. Also, the highs are overemphasized and the birds, wich we can clearly hear in the other takes, sound distant, specially in the 3-5 KHz range.

In the other takes, as you rise the gain you get some clarity in the bird sound and I don’t appreciate that you lose any sound quality, although maybe is hard to tell since it’s just ambient noise.

In the Izotope RX’s analysis we can clearly see the 12 Khz tone plus two other high bands above. We can also see that the 7-15 Khz range is overemphasized compared to the other takes. Also there is a clear high emphasis in the take’s borders (fades) Why all these appears? Is Pro-Tools producing it when we rise the signal? I have turned up the takes using the Pro-Tools’ native equalizer. I have also made the same operation with other plugins with the same result.  Just to make sure I had made it with the Pro-Tools’ offline AudioSuite too. Same results. Thinking that maybe is a Pro-Tools thing I’ve tried the same in Logic. Exactly the same.

It seems like rising the signal digitally introduces errors in it or maybe this errors are due to record with low gains. However it looks like it’s cleaner to earn those dB’s with the pre-amp gain.

You can also see in the second take a clear DC Offset (or so i think). It appears anyway I turn up the signal so it seems that it is in the original take. An A/D converter fail? Studying the complete take I realized that in that exact moment I was still rising the gain which can explain that converter behavior. To the ear it’s barely noticeable. In the RX’s analysis you can see a low end increment while in the soundcloud’s waveform you can see a clear level amplification.

Test 2: Signal

For this test I’ve record a Youtube dialogue (quite noisy) played in the monitors and captured with the same recorder and mic. 

We hear again something similar to the previous test. In the 3 Gain take the highs are again emphasized and we can hear the 12 Khz tone. 

We can also hear a 7 Khz periodic tone with its armonics. You can hear it as well in the first test. It’s very strange. I don’t think is an ambient noise (like a bird) as its periodicity is very exact (2,4 s) and we can’t hear it in the other takes which were made immediately after. I don’t find any explanation to this sound. We can see it in the RX’s analysis. 

For the rest, aside from the little ambient noises that change in every take, the 3 other takes are quite similar. I don’t hear any sound deterioration as you rise the gain.

So it’s seems to me that the best option is to record at high gains (being careful not to saturate). The tests tell us that the signal/noise ratio keeps constant as you rise the gain.

I have recorded these test at 16 bits and 48 KHz trying to keep the same sound quality that I usually use in the set. Given that I have recorded at 16 bits and low gain and I haven’t used many of the disponible bits this could traduce in converter errors. It’s the best theory I can find now. I will try to confirm this with futures tests at high resolution.

In the other hand I don’t find any explanation to that periodic noise at 7 KHz neither to the 12 Khz tone. Maybe inherent recorder noise? I leave this questions here in case someone could answer them.