How should I use my patchbay? Why can I only play one sound with my Octapad? How does DAT error correction work? Can I avoid the animation on your web site?
For more hints, tips and problem‑solving visit the
SOS Discussion Forum https://web.archive.org/web/2015..." target="_blank
I have recently bought a Signex patchbay (balanced). I am uncertain as to when I should be using the connections isolated or normalised. I need to use it for mixer channels and inserts, effects and soundcard connections. Could you also explain how halfnormalising is achieved?
Editor Paul White replies:The topic of patchbays and the best way to use them has been covered on various occasions in SOS. You should find all the information you need in 'Bay Watch: All About Patchbays' in the March 1998 issue, and 'Patching Up', an FAQ article about patchbays which we ran in December 1999. Both are available to read free on the SOS web site (https://web.archive.org/web/2015..." target="_blank). However, I'll also cover some of your points briefly here
As a rule, non‑normalised patchbay connections are used when you want to bring the inputs and outputs of a mixer, effects boxes, recorders and so on to a patchbay. This allows them to be linked using short patch cables. Normalised connections are used where you want the signal flow to continue uninterrupted when nothing is plugged into the patchbay (inputs bottom row, outputs top row). Normalising (or semi‑normalising) is used for all insert points, but may also be used where you want to establish a default connection, For example, where a mixer output feeds a recorder input unless a plug is inserted to break the circuit.
Most normalised patchbays are actually semi‑normalised so that the signal flow is only broken when you plug into the input but not the output. This allows you to take a feed from the output by inserting a patch lead without breaking the signal flow.
I have recently purchased a second‑hand Roland Pad 80 Octapad II so that I can play overdubs using samples of my drum kit. The samples are in the 8Mb SIMM RAM of my Soundblaster AWE32 MIDI synth on my Pentium PC.
Having configured the Pads to trigger the eight sounds I wanted, I found that each pad triggered the same sample, the one set to 'Record' on my sequencer. This soundcard only allows one MIDI channel to be set to 'Record' at a time, so although I could, say, play a snare‑drum part properly, I couldn't play a run‑down on the toms, or a snare/hihat break. I sought advice from several retailers, one of whom told me that I needed to buy a new soundcard which could record on eight channels simultaneously, thus solving the problem.
While researching which soundcard to buy, a second retailer told me that I should be able to set all eight samples to the same MIDI channel (1), set all the pads to the same MIDI channel, and use a function called 'key range' on the pad controls to limit the playback to the sound I wanted.
Having read the Roland handbook from cover to cover, I found no mention of this. I did find a control called Program Change, so I set up all eight samples on MIDI channel 1, and entered the appropriate Program Change number to each pad, which I presumed would change the sample within the same MIDI channel.
No such luck, so please enlighten me. What do I need to do (or buy) to enable me to perform the apparently simple task of playing eight drum samples back simultaneously from an Octapad?
Martin Walker replies: There are obviously some less well‑informed retailers out there, as you have been given some conflicting advice. First, each input port on any MIDI interface can accept data on up to 16 channels simultaneously, as well as sending out data on 16 channels. It's your music application that determines whether or not you can record several at once, and (for instance) Cubase VST provides the 'MultiRecord' function to do just this. However, this is normally used to record the simultaneous performances of several players, where you want to keep the MIDI data from each instrument on a separate track for possible editing.
In your case, the second retailer is quite correct. Multiple drum sounds are nearly always 'mapped' on to different MIDI notes, but all on the same MIDI channel, so that they can either be played from a keyboard, or from other MIDI controllers, such as drum pads. This is how drum and percussion sounds are arranged inside the vast majority of synths, and in the case of General MIDI there are even standards on how the different sounds are mapped — bass drum on note 36 (C1) and snare drum on note 38 (D1) for instance — so that tracks can be easily tried out with different drum 'kits'. Convention also normally places the drum sounds on MIDI channel 10, although you don't have to stick to this in your own music — only if you want to play back standard MIDI files.
Roland UK told me that this is exactly what the Pad 80 presets do, for different styles of kit, but that you can also create your own user kit if required. So the answer is to make sure that the SoundFont you use has its samples mapped in the same way as one of your existing presets, and create your own user kit, or instead use your XG synth, which will already have its drum sounds correctly mapped on MIDI channel 10. By the way, if anyone out there has a Pad 80 without a manual, Roland UK can still supply them, as well as a patch listing for the instrument. Telephone them on +44 (0)1792 515020.
Please could you answer this question for me: why does DAT error correction not always work?
Technical Editor Hugh Robjohns replies: It's a short question, but needs a long answer I'm afraid — although if you really want the potted version it is simply that some kinds of data damage are simply unrecoverable.
All recording media suffer from degradations of one sort or another. Any error‑correction system must be designed to cope specifically with the kinds of errors expected on a given medium — the likely error rate, duration, number of adjacent damaged blocks, type of damage (missing data or corrupted data, for example), and so on.
The system used by DAT is the doubleencoded Reed‑Solomon system, which uses the same basic concepts as that used by CD, but is far from identical, as it has been optimised to cope with error bursts of a very different nature.
Error correction relies on two stages: the first is to detect the fact that the data has been corrupted in some way, and the second is to correct the corruptions. To ease the workload on the error‑correction system, pretty much all digital audio systems (and other data‑handling systems) also use a process called 'interleaving'. This is a method of re‑ordering individual data blocks within the data‑stream before recording, to offer greater protection against a large burst of errors trashing adjacent data. On replay, when the datastream is de‑interleaved and the data is reorganised into its original sequential form, a large block of adjacent errors is translated into a number of isolated single errors, much easier for the error‑detection and correction systems to deal with.
However, certain kinds of damage can elude the detection system completely — although this is relatively rare — and suggest either severe physical damage to the tape, or excessive head wear (or dirty heads) reducing the capability of the system to recover data from the tape in good condition. In this situation, as the corrupted data is not recognised, it will be passed on to the output to become audible as a splat or glitch of some sort.
If the error is successfully detected, the appropriate correction strategy is employed. The restored data is output, and you will be completely unaware of the process taking place. Some professional machines provide an indication of how hard the error correction is working, since this is the only indication a user can have of the condition of the medium and machine. A higher‑than‑usual error rate would suggest either a dirty replay mechanism or a worn‑out tape. In the case of the former, routine maintenance is required, and for the latter, a cloned copy should be made of the original tape. If the clone is only made when errors become audible, some of the original data has been lost forever!
A third error‑handling stage is also used in most digital audio systems. If an error is detected but the damage exceeds the capability of the correction system to repair it, a further process is employed. This is called 'concealment' or 'interpolation' and relies on the fact that the bulk of audio material changes relatively slowly (at least in terms of the rate of digital audio samples) and is largely predictable.
If the error correction is exceeded and concealment has to be used, the defective samples are discarded and a replacement calculated by performing an intelligent average over preceding and following samples. In this way, the replaced sample (or samples) is (or are) likely to be very close in amplitude to the damaged samples being replaced.
Concealment inaccuracies can be heard very occasionally, but usually only on very loud high‑frequency material. This is because a rapidly changing signal shows up the deficiency of the relatively simple averaging process and, being loud, the inaccurate averaging tends to be heard as a small click. However, 99 percent of music has little energy at high frequencies, so concealment is generally inaudible.
Of course, concealment results in inaccurate data and so is completely unacceptable in pure data (ie. computer) applications. In these cases, the usual solution is to employ an even more sophisticated design of error correction which uses typically three‑layered strategies instead of the two (plus concealment) used in most audio systems.
The added error protection employed in pure data systems has to be 'paid for' with less efficient data packing, as more of the recorded data is tied up with error protection instead of carrying the wanted information. The error‑protection information is often called redundant (or redunancy), as it takes up space but is not carrying any wanted data — it is simply there to look after things.
Obviously, the stronger the errorprotection system, the more redundancy necessary, and the less space available for useful data. This manifests as a lower recording duration or storage capacity of a given medium.
The pragmatic answer is that DAT errors become audible when the tape is damaged (or just worn out), or the replay (or record) machine is defective. This could be because it is simply dirty and needs cleaning with an approved cleaning tape, or the heads are worn out, or the machine is out of alignment. Few people ever have their DAT machines or CD players serviced — they seem to think that 'digital' means 'perfect forever'. The truth is that the mechanical tolerances in a DAT or CD player are much finer than for video recorders or any analogue system, but the amazing strength of the errorcorrection systems can successfullly cope with huge misalignments. A tape which produces audible errors and splats on one replay machine may work perfectly well on another if its alignment or cleanliness is better. Most mastering rooms have at least a pair of machines (from different manufacturers) for this very reason!
Another couple of handy tips: always record at least two minutes of tone or silence at the start of a DAT, so that the important audio is kept well away from the beginning of the tape (which is almost always damaged by its joint to the spool hub). Also, never eject a tape without rewinding it to one end or the other. I usually wind it forward to the end before ejecting. If the tape gets eaten by the machine as it is ejected, the damage is well away from the crucial material and it is relatively easy (and safe) to repair. Also, the tape has to be rewound before playing, which evens out the tape tension with many machines.
I recently subscribed to your magazine. Absolutely fantastic journalism. But I find it very hard to read any article (on your web site) with the animation 'July 2001 Current Issue' flashing constantly! Could you change this as soon as possible?
Producer, Blink Music Inc
Publisher/Webmaster Ian Gilby replies: Tim, the SOS web site only contains one animation — in the navigation panel across the top of each page — and in this day and age it's an oasis in a desert of Flash‑animated sites. However, here is a way of avoiding it: Click the front cover icon on the left, then click‑and‑hold the mouse pointer on the monthly contents list which subsequently appears in the righthand frame below the navigation panel. Up should pop a menu allowing you to 'Open frame in new window' (or similar wording). Select this option and the full contents list with article links appears in its own window, minus the navigation panel. Bookmark the URL of this contents page (Add to Favorites) and just use that. The July 2001 contents page URL is: https://web.archive.org/web/2015..." target="_blank
Tip: you can access past contents lists by changing the month/year reference, so the June 2001 issue would be: https://web.archive.org/web/2015..." target="_blank
And September 1999 contents would be:
Hope this helps increase your enjoyment of the SOS site and articles.
I have been getting conflicting advice regarding this issue from my local music stores and am turning to you in desperation.
I have a beige G3/233 Mac with 192Mb RAM, which I would like to set up to record to HD. I've chosen the sound card (MOTU 2408), but need to install an AV‑compatible drive that is fast enough to play back video and record at least eight tracks at the same time. I've been told that it is no longer possible to get a SCSI card and drive for this model of Mac and that I have one of two options: purchase an IDE card and mount internal IDE drives in the Mac; or purchase a Firewire card and mount internal IDE drives inside external Firewire enclosures.
What is my best option? Are there any others? I would prefer an external solution.
Assistant Editor Sam Inglis replies: If your beige G3 is anything like mine, you shouldn't need to add an extra IDE card in order to fit an additional IDE hard drive. Like most PCs, the beige G3 Macs have two IDE busses, each of which supports two devices (Master and Slave). You'll probably find that the built‑in hard drive, built‑in CD‑ROM and built‑in Zip drive (if you have one) are connected to the IDE busses, which leaves one device slot spare for an additional hard drive. I installed an additional 20Gb hard drive in the extra IDE slot on my Mac, and it works fine, although it was a bit of a pain to install. The two hard drives should be on different IDE busses, and should both be the Master devices on their buss.
I had two main difficulties in installing the new drive. The first was the physical problem of mounting the drive and cabling it up. Standard IDE cables have three connectors — one at each end, and one in the middle. One end connects to the motherboard, the other two to your two IDE devices. However, I couldn't find a standard IDE cable that had the middle connector in the right place, and eventually had to add another middle connector to it (you can buy these from Maplin et al — they sort of clip through the cable itself). You also need to get a plastic mounting bracket to fit the drive in the drive bay. Apple's recommended price for these is exorbitant, so try to scrounge one from a dead Mac.
The second problem I came across was that Apple's own drive‑formatting and management software wouldn't recognise the drive. I had to use a different program and driver (SilverLining Pro), which worked. This will probably change depending on the drive you choose. I was also told that if you're installing a large drive, you should partition it, so my 20Gb drive appears on the desktop as two 10Gb volumes.
Of course, the problem with this approach is that it's not an external solution. I doubt it's practical to use IDE drives externally, so if you're committed to having an external drive, you will need to go down another route. I'd steer clear of FireWire drives at present, unless you have definite confirmation that they will work for streaming audio. I'd be very surprised if you really can't get a compatible SCSI card, though — if nothing else there must be millions of second‑hand ones around!
Q I've been offered an old Allen & Heath 16‑channel, 8‑buss recording desk, around 1975 vintage. The seller wants me to make an offer on it and I'm not sure how much to offer. I don't want to rip off the seller and at the same time I don't want to rip off myself! Do you have any ideas on value? It's a huge piece of kit and I could probably get something new a tenth of the size, but at the same time it sounds really good and is ideal for my small studio setup. Greg Fullarton
Q I saw a Turnkey ad in SOS July 2001 for the Roland U8ST. I have searched for more info on this USB mixer/interface but I can't find anything. It doesn't seem to be on the Roland web site. I was wondering if you knew anwhere I can find a full spec for the U8ST, and if you think that it might be a (cost‑cutting) option for recording multitrack audio since I haven't yet purchased a multitrack card. Nick Barlow
Q I am relocating to the UK from South Africa. I have been a sound engineer in satellite radio for about four years, using various tools (Soundscape, Yamaha O3D and 02R, RCS radio software). I was wondering if you could tell me if there is a newspaper or magazine that publishes jobs for sound people. Andrew
Q I read in SOS ages ago about a keyboard convertor and a monitor adaptor for the Atari. These would be used to connect PC‑type keyboards/monitors to the Atari. Could you please let me know where I can purchase them? Ashley Bovan
Q I'm thinking of buying a digital desk (I'm considering an 01v) but have been told not to buy second‑hand digital equipment. Is there any reason why I shouldn't? Mick Cape
Q Which sound module gives the most "realistic" sound quality, in your experience? Some people say Korg, while the next person insists it's Roland and yet another person prefers Yamaha. Is there a brand and model that most people seem to agree produces the truest sound? I'd like to keep the price under $1000. Michael Kramer
Q Is it even remotely possible to find a valve recording console, say at least 16‑channel, 4‑buss, for less than £2500? We all dream of TL Audio's VTC console, but sadly £23k is a tad out of my league. Craig Walker
Q Our church pianist stopped playing for us, so now our church is in need of a machine that can play a CD or cassette and take out the voice, leaving just the background music for our choir to sing with. We purchased a karaoke voice canceller from Radio Shack, but all it did was turn down the music coming from the CD or cassette. We need something to take out the voice and leave the music. Freddie Lofton
A I'm afraid no such machine exists, and it's unlikely that one ever will. So‑called 'voice cancellers' basically work by putting the left and right stereo signals out of phase with each other, so that any signal in the middle of the stereo field is cancelled out. This usually at least reduces the level of the vocals, but it's rare that you will be able to eliminate them altogether by this method (it depends upon the track). Better find yourselves a new pianist... Sam Inglis