Sound Devices v. Zaxcom

[Rob Castiglione]

Hey Trailangel,

I would also kindly disagree. I tell you what. Why dont you post some side by side comparisons. Identical recording chain but one recorded at 24/48, one recorded at 24/96 and one at 24/192. Let's see who can tell the difference.

We already have "HD" audio - 24/48 khz.

 

[Dave Blackham]

It's better to have than not. When shooting Red HD.. the audio needs to be HD as well. Next will be 3D audio.

If by 3D audio, do you mean surround. We already do this. We are using a variety of mic arrays with some of the best mics and preamps around, Decca tree for music work and various double MS techneques to name a couple of ways of going about it. 48/24 works perfectly well.

 

[Trailangel]

No, there is another audio component that is to be incorporated into the 3D glasses....tiny speaker or eatbuds that fold down receiving a wireless multiplexed transmission.

 

[Michael Panfeld]

So if 24/48k is the standard for dialog, why not take a good mixer (I have a Cooper CS104) and record the output on a cheap Zoom H4n (records at 24/48k using the XLR inputs)?

 

[JonathanF]

I think that's a different issue. There's a different feel with vinyl, and I have an open mind to the point where I'll concede that analog tape, digital, and vinyl all can sound different. But there's a difference between good digital and bad analog, and bad digital and good analog.

Marc,

I was being facetious.

J

Hey Trailangel,

I would also kindly disagree. I tell you what. Why dont you post some side by side comparisons. Identical recording chain but one recorded at 24/48, one recorded at 24/96 and one at 24/192. Let's see who can tell the difference.

We already have "HD" audio - 24/48 khz.

I spent a few years sharing office space with one of the top recording studios in the country and I can say that 192 will save your ass sometimes. This is the only reason I advocate it. Roger knows his stuff as well as anyone (and has the resumé to back it up), so his point is VERY valid. Like he says, 24/48 is THE standard, we always send 24/48 to post.

We do, however, let post know the original source is available at 96 or 192 depending on what was recorded, but give them 24/48. We have gone back to 192 a couple times to give the sound editor something he can really work on to get what he's after. Doing this is rare but the few times it's happened we've been glad we had it to fall back on. This is not so different than scanning film plates at 4k or double flashing on the Arriscan for some shots where you need the extra information in the scan (bad key, etc.) instead of at 2k single flash as we do for 80% of the films I've done. There are times when you genuinely need it to get out of a tough spot on some shots.

That said, some of us have systems that you CAN tell the difference with. I personally use an Audio Note M6 preamp and P2SE single ended amp, both using hand matched NOS Telefunken and RCA tubes. They drive a couple hand built Joseph Class A reference monitors. We have tested 48, 96 and 192 on this system and nearly every person with a critical ear CAN tell the difference and easily, repeatably picks the 192 from other recordings. To achieve this we used the .zax files to produce each of the individual playback tracks.

Most people don't spend this kind on scratch on preamps so this is an extreme situation to use as an example, but you get the point.

Again, the main issue is that IF anything needs to be done in post, speed ups, pitch change, any high level processing, etc. you can do a far better job if the source is as high of resolution as possible.

 

[Michael Panfeld]

And going in the opposite direction from my last post...

Someone earlier in this thread talked decimation. Here is a paper that discusses decimation as well as a lot of the issues in this thread: http://www.korg.com/services/products/mr/Future_Proof_Recording_Explained.pdf

So, is anybody using "1-bit" DSD recording for film work? I mean if 192k can "save your ass" as some have said, why not go to a unit like this which gives you all options?

http://www.bhphotovideo.com/bnh/controller/home?O=cart&A=details&Q=&sku=485879&is=REG

I also like the idea ousing this as a master archiving solution.

 

[Tom Visser]

If the audio files are compromised at 24/48, they will be compromised at 24/192 and there is no intrinsic aspect of the 192KHz sampling rate that will magically make something recoverable if it wasn't at 48Khz. If you are shooting slow motion and wish to play back sound without resampling, which can introduce some nasty artifacts, 24/96 will play back at 24/48 half speed with a simple modification to the sample rate flag of the file and 24/192 will allow one to play back quad speed material without resampling.

Sample rate is not like pixel density or resolution. This analogy is commonly made with audio, but it simply does not apply. The science behind PCM / sample rate and image / resolution, are not comparable.

DSD... is an interesting technology. It will never succeed as a mainstream format as long as Sony owns the licensing rights. It might find a niche market for audiophile labels and archival, but will remain pretty small. In my estimation, DSD is more or less equivalent in sound quality to 24/96. Most of the comparisons and marketing materials advertise DSD as superior to PCM, but are mostly referring to PCM at Redbook CD standard or 16 bit 44.1KHz. So even though the marketing is correct, it does sound better, 24/96 also sounds better than 16/44.1. First thing problem with Redbook CD, the 16bit wordlength does not have the headroom of 24 bit. Very high end audio reproduction systems may be in fact capable of resolving somewhere in the neighborhood of 20 bits or so. Ridiculously expensive systems may run as high as 22 bits or so. In these cases, the 16 bit wordlength of Redbook is a limiting factor. The second failing of Redbook are the brickwall and/or sampling filters used, and when speaking of 44.1KHz, those filters exist just on the edge of human hearing, around 20KHz. When using 96KHz sampling rates, the filters are moved up the audio spectrum well above human hearing thresholds to around 40KHz-ish range. So... the biggest advantage of utilizing 24/96 is not the fact that the sampling rate is increased, per se (or incorrectly advertised as "more resolution"), but that the filters required for linear decoding of PCM are not in the human hearing range. DSD also requires filtering, but is a completely different process than PCM. Its non-linear characteristic, I forget the exact frequency, but is something in the MHz range, I believe, completely undetectable (as is 24/96). DSD / SACD, is pretty cool, though, but simply not a feasible mainstream format, not due to technical reasons, but mostly because of Sony.

I do not know of any film projects that touch DSD. The Korg MR-1000 is a pretty cool unit, not too expensive. Its microphone preamps are substandard compared to other top=flight recorders, but are decent enough. It really excels more as recorder when using line level inputs, which undervalues its portable battery powered operation. The Korg MR-2000S would be the better studio archival deck.

 

[Michael Panfeld]

Excellent post Tom. I must say that I spent the better part of the day researching this as I am in the market for a recorder. The result of my research lead me to some farout forums on high end audio. The bottom line from the audio experts on those boards is that, when DSD is recorded in DSD and then played back as DSD, it blows away anything recorded under a PCM pipeline. But since no theater and almost no consumers have this capability, that difference is not relevant to filmmaking.

However, what might be of relevance is whether a today's PCM recorders with their hardware-based AD converters can yield a better-sounding 24/96 (or other sampling rate) file than a DSD recorder that records to 5.6 mhz DSD file and then, using a software based AD converter, converts it to a 24/96 PCM file. The constraint on the hardware-based DA recorders is that it must be done in realtime, whereas a software based conversion can be down later and at a less than realtime pace. They did do some comparisons to Lavry Gold DA converters (whih cost $8500) and most people could not tell the difference. Given the price differential, that makes the Korg a bargain.

I also read that, just as Sony has a stake in seeing this technology prosper, likewise the PCM folks want to remain on top. Its hard to know who to believe when everybody has a vested interest.

That said, I agree, its an interesting tech and one that may be excellent for master archiving, given that it can be delivered to PCM pipelines now, and presumably delivered using better tech into PCM or DSD pipelines in the future.

Its also interesting in that the limitation of the human ear seems to be the ceiling right now (Nyquist and all). But the tech may have some usage outside of humans (thinking computer analysis or animals) whereby the extra sampling might be very relevant. Of course, that would be outside the scope of this, a filmmaking forum.

Its informative that the Korg's pre-amps are inferior and that the line inputs are better. I am researching a good pathway. I just bought a Cooper CS104 mixer, which I have read has a "warmer" tonal quality than say a SD 442. So I am wondering if the Cooper/Korg pathway might be complimentary given the compliments that the recordings seem more natural (i.e, without any coloration). Can't I go from the Cooper to the Korg using a line out?

My field recorder choices so far (from cheapest to highest) are

Zoom H4n (24/48 capability)
Tascam (24/196 + timecode)
Korg MR-1000 *on sale at B&H for $1019!*
SD 702T
SD 722
SD 744


(* I am certain this will stir up something, but I brought the Cooper into Vark Audio for servicing and they said that their bench testing has the Cooper easily besting the SD 442).

Cheers

 

[Dave Blackham]

Excellent post Tom. I must say that I spent the better part of the day researching this as I am in the market for a recorder. The result of my research lead me to some farout forums on high end audio. The bottom line from the audio experts on those boards is that, when DSD is recorded in DSD and then played back as DSD, it blows away anything recorded under a PCM pipeline. But since no theater and almost no consumers have this capability, that difference is not relevant to filmmaking.

However, what might be of relevance is whether a today's PCM recorders with their hardware-based AD converters can yield a better-sounding 24/96 (or other sampling rate) file than a DSD recorder that records to 5.6 mhz DSD file and then, using a software based AD converter, converts it to a 24/96 PCM file. The constraint on the hardware-based DA recorders is that it must be done in realtime, whereas a software based conversion can be down later and at a less than realtime pace. They did do some comparisons to Lavry Gold DA converters (whih cost $8500) and most people could not tell the difference. Given the price differential, that makes the Korg a bargain.

I also read that, just as Sony has a stake in seeing this technology prosper, likewise the PCM folks want to remain on top. Its hard to know who to believe when everybody has a vested interest.

That said, I agree, its an interesting tech and one that may be excellent for master archiving, given that it can be delivered to PCM pipelines now, and presumably delivered using better tech into PCM or DSD pipelines in the future.

Its also interesting in that the limitation of the human ear seems to be the ceiling right now (Nyquist and all). But the tech may have some usage outside of humans (thinking computer analysis or animals) whereby the extra sampling might be very relevant. Of course, that would be outside the scope of this, a filmmaking forum.

Its informative that the Korg's pre-amps are inferior and that the line inputs are better. I am researching a good pathway. I just bought a Cooper CS104 mixer, which I have read has a "warmer" tonal quality than say a SD 442. So I am wondering if the Cooper/Korg pathway might be complimentary given the compliments that the recordings seem more natural (i.e, without any coloration). Can't I go from the Cooper to the Korg using a line out?

My field recorder choices so far (from cheapest to highest) are

Zoom H4n (24/48 capability)
Tascam (24/196 + timecode)
Korg MR-1000 *on sale at B&H for $1019!*
SD 702T
SD 722
SD 744


(* I am certain this will stir up something, but I brought the Cooper into Vark Audio for servicing and they said that their bench testing has the Cooper easily besting the SD 442).

Cheers

The Cooper is an excellent Mixer. it should match well to the Korg if you go that way. The spec of it looks very good, certainly they seem to be using best of breed components. I dont see any DSD plug in's for Post systems though which would be my only concern, Any one know if these are available ?

 

[Tom Visser]

I would take it a step further, Cooper made some of the best mixers in the world. It is a shame that they decided to cease operations. The Cooper Korg combination would be a really cool combo for music remotes. The Sound Devices is no slouch, but it is no match for the sound of a Cooper.

If you want to work with DSD in post, you have a couple of options...

1) Sonoma audio workstation. It's very expensive and highly proprietary. Plugin support is highly restrictive. I believe when you do things like crossfades, the software converts DSD to PCM for the actual fade itself.

2) Sadie had a solution, they went bankrupt, and were then purchased by someone and are now back in some form.

3) Pyramix is the most popular and cost effective solution, but still not cheap by any means. Pyramix basically works in DXD which is more or less PCM at (I think) 384KHz. People should not be afraid of the PCM based workflow, I don't know the full science behind it, but those that are smarter than me have given very thorough explanations why DXD is DSD friendly and doesn't compromise at all, and is backed by the sound. Again, plugin support is certainly not like ProTools running 48 or 96 projects, but is better than its competitors.

4) and the way that i personally would handle DSD if I had any interest in going this route... Just treat it as a tape machine and work with hardware. The Korg MR2000S has the ability to lock multiple units together to record and play in synch. So if you purchased 4 units, you'd in effect have an 8 track recorder. Run this rig exactly like you would a Studer tape machine in the old days for effects and mix down to another DSD recorder, another 2000 or a 1000 or one of the Tascam machines. Edits and sequencing could be handled in software.

 

[Roger Morris]

Most Cooper mixers were designed for something very different than a Sound Devices 442.

Although I own (and love) my SD442, it's first and foremost a "bag mixer". Sound Devices doesn't market it as a "cart mixer" for Features the way Cooper was marketed and used (on Features).

The headroom on a Cooper is mythical, and there's really very little to compete with it. Certainly the SD442 doesn't have headroom in practical use that equals the Cooper (although the SD442 has headroom for 99% of the audio situations a pro will run into).

Sadly, Cooper is no more. There are Sonosax mixers out now that are quite nice, ........but alas Cooper owners will know that there's nothing like a Cooper!

The Cooper went for over 10 Grand, and that was for console only. These days that 10 Grand will buy you a SD788 and it's CL-9 "sit down mix tab"...with enough change left over to buy tons more "sound stuff".

It's worth pointing out that Cooper went out of business not due to any lack of sales, or because the console was being surpassed by a superior technology, but simply because the owner wanted to retire in peace.

I, along with many audio folks, keep my eyes open for a mint Cooper for sale.......but most who own them know they'll never be marketed again, and keep them very close, and definitely not for sale!

Having said all of that, I'd not ever trade my SD-442 and SD-702T away.......when used for what they were designed for, there's nothing to equal the combination of the two.

 

[Marc Wielage]

I, along with many audio folks, keep my eyes open for a mint Cooper for sale.......but most who own them know they'll never be marketed again, and keep them very close, and definitely not for sale!

Much of what you say is true, Roger. But trust me, used Coopers are out there if you look around. I think I saw a couple of Cooper CS208's available within the last three months for about $10K (well below $16K list).

It's a great analog mixer, but I think the PSC Solice comes very close to matching it in terms of quality and features. Also, more and more mixers are going all-digital with units like the Deva Mix-12, but they depend on the inputs of the recorder itself. This is one big advantage of Zaxcom over the other guys, but Sound Devices recently came up with the CL9, which is considerably cheaper (and uses a very similar design philosophy). I'd call it any SD vs. Zaxcom comparison a "Cadillac vs. Mercedes" situation, with the Cantar being a BMW.

I agree the 442 mixer and 702T will work fine in straightforward 2-track situations. The problem is that production is demanding more and more tracks and very complex situations to allow more flexibility in post. It's ironic to remember that some of the greatest movies ever made were shot in mono, and that 2-track only became popular for feature dialog in the 1990s. In the past 5 years, though, at least 4 tracks are expected as standard, and 8+ tracks are becoming commonplace.

BTW, I'm very much a Zaxcom user and love the company, but I also own and use SD products (and just bought a 552 for smaller projects). Sometimes, it's nice to own a Mercedes and a Cadillac if you can swing it.

 

[Dave Blackham]

It's ironic to remember that some of the greatest movies ever made were shot in mono, and that 2-track only became popular for feature dialog in the 1990s.

Theres a point here. When I was taught to mix Mono audio was always thought to be harder to mix than stereo. I think this is still completely true. On the face of it it may be considered easy as theres less routing and so on going on but from a creative point of view theres much less to distract the listener so the mic placement and mix has to be absolutely perfect or any slight imperfections will be heard. Your ears are a lot harder to fool than your vision.

 

[Tom Visser]

Theres a point here. When I was taught to mix Mono audio was always thought to be harder to mix than stereo. I think this is still completely true. On the face of it it may be considered easy as theres less routing and so on going on but from a creative point of view theres much less to distract the listener so the mic placement and mix has to be absolutely perfect or any slight imperfections will be heard. Your ears are a lot harder to fool than your vision.

I think that the mono mix is as relevant as ever. It is critical for dailies, guide track for later replacement, or in some cases, for final delivery (yes that still happens even in this day and age). Now that we have multi-track recorders, it hasn't superseded the need for mono mixes. The ISO tracks tend to be more of a post asset that is necessary, especially now that less attention is being paid to really getting clean location sound, but the mono mix is the sound mixer's perspective on dialogue and FX, from the only person on set who is actively monitoring sound... a very valuable perspective for post sound to have. Now that most of our mixers are at least "stereo" mixers, and camera hops are "stereo", I have been treating this more as "dual mono", and giving me the ability to provide what is more or less an isolated boom feed (my teams best foot forward of what we could capture on set, minus negative factors such as compromised boom position due to production variables, ambient noise pollution, or operator error) and a lav mic mix (or 2nd boom, ambient etc...). Maybe it is simply my ability, but I have never really been satisfied mixing boom mics and lav mics together, so welcome stereo camera feeds to allow me to provide 2 mono tracks in this fashion. The raw playback of such tracks can be a little distracting, but its rarely done on anything but rudimentary monitors and other professionals can quickly get past the non-stereo tracks played as stereo at least for dailies viewing.

If I've done my job competently, over 90% of what post sound needs can be drawn from these two mono tracks, versus having to dig through lots of ISO tracks that need to be mentally decoded. On another side note, are any of you sound mixer recording coms feeds? I wonder if this would be helpful to post since it gives them an on-set perspective of what was happening at that point in time. "Hold sound - helicopter"... although it's obvious once you start hearing it, that verbal cue could save people time and scrub forward to the "set" call rather than having to decode the information. Actors handling lav mics, mixer giving cues to the operator, etc... all these things that have not traditionally been part of the deliverables but now perhaps could be a convenient set of running notes? Anyone think this is a good idea?

 

[Roger Morris]

For me, when the need for more tracks exceeds the two on my 702T, I use my FireFace 800 and Metacorder in a custom rack set up.

Metacorder gives me more tracks than I could ever imagine using, and is the preferred software based solution in many (if not most) post houses.

Worth pointing out though is that my choice of hardware for two track and software for multi-track is just a personal preference based on years of experience with other packages.

Generally, I use the SD package in a bag, and the Metacorder/FireFace package on a cart.

It all depends on what the production is calling for......or more often, what they're willing to pay for!

 

[Marc Wielage]

On another side note, are any of you sound mixer recording coms feeds? I wonder if this would be helpful to post since it gives them an on-set perspective of what was happening at that point in time. "Hold sound - helicopter"...

I just say that verbally as part of the tail audio slate. Since the AD and the director are hearing my feed, they know exactly what's going on.

We're crowded as it is for maximum channels without also having to record the comm feeds. I also think there might be some invasion of privacy issues, because some stuff said on the comms (especially walkies) shouldn't be repeated or recorded.

 

[Tom Visser]

I just say that verbally as part of the tail audio slate. Since the AD and the director are hearing my feed, they know exactly what's going on.

We're crowded as it is for maximum channels without also having to record the comm feeds. I also think there might be some invasion of privacy issues, because some stuff said on the comms (especially walkies) shouldn't be repeated or recorded.

That is so true... about the privacy thing.

 

[Dave Blackham]

For me, when the need for more tracks exceeds the two on my 702T, I use my FireFace 800 and Metacorder in a custom rack set up.

Metacorder gives me more tracks than I could ever imagine using, and is the preferred software based solution in many (if not most) post houses.

Worth pointing out though is that my choice of hardware for two track and software for multi-track is just a personal preference based on years of experience with other packages.

Generally, I use the SD package in a bag, and the Metacorder/FireFace package on a cart.

It all depends on what the production is calling for......or more often, what they're willing to pay for!

This is quite true, Metacorder is great software and offers a very good alternative to the bespoke recorder approach. It also offers the choice of alternative A to D's and Preamps. For my use it tends to get used for music work rather than features, in part as its not so portable.

With Tom's point about Mono deliveries, its still the principle reason M&S mic set ups are used for good mono compatibility. Nothing else comes close. With complete respect to post sync work which often has incorrect perspective or acoustic which is distracting, I still believe if there were more consistently good audio captured on location then more of it would get used in the final mix to give a better end result. Im not so sure about recording comms to indicate whats going on on set. If the audio cant be used then don't record it or if you do make sure its noted on the sound sheets.