One For Ivan And Bogdan, trialing a new mastering compressor Options

[Saoirse O'Shea]

We've just been sent a new mastering compressor on evaluation for a few months by one of our suppliers. It's an Elysia Alpha and our supplier is pretty keen that we use it as well as/instead of our main compressor.

Anyway, Ivan very kindly provided me with an audio wave of a track he has recorded recently with Bogdan and their group Ljute Papricice called 'Nocni Gruv' to put through the Elysia. I should say at this point that what follows was aimed at evaluating the Elysia so Ivan and his group had very little chance for input in to what I did and so have been particularly kind to allow me to post the result here. Thanks guys .

Ivan has very kindly posted the original recording as an mp3 - see post no. 13 below. Thanks Ivan

Anyway here is my 'Nocni Gruv' Elysia test master for Ivan and Ljute Papricice.

 Nocni_Gruv_Master.mp3 ( 5.94MB ) Number of downloads: 301


-------------------------------------------------------------------------------------

And here is an edited version of the mastering log to give you all some idea of what went on and why:

-------------------------------------------------------------------------------------
Client: In house Elysia Alpha Mastering Compressor test.

Track: 'Nocni Gruv- Ljute Papricice'

Original: 96/24 Wave Required final format:44.1/16 Wave for Ljute Papricice plus MP3 [for upload to GMC]
Track length: 3:54 approx 110mb 96/24 wave.

[File was sent zipped via internet. Zipping here both reduces the file size and also provides a degree of data conformity checking.]


Internal DAW to 96/24 [to match the 96/24 wave; any final format changes required will be done as a final when I re-convert to 16/44.1/mp3.] Chain is :

Daw aux out -->Lavry DA1--->Elysia-->Ibis--->Lavry AD-->Daw aux in: daw main-->Lavry DA2 -->monitoring

Monitors and meters to K14 [I don't use peak meters for mastering. I use the K system and usually set out at k14 for rock/jazz. Monitors are pink noise calibrated on the K system. This means that I can accurately set, maintain and repeat spl, volume etc for mastering. This is important both wrt the Fletcher Munson effect and also so that I am much less likely to clip on transients.]

Normally a client would supply me with a reference track - a piece of music that they like the sound of that they want their recording to be like. I didn't have one and didn't discuss with Ivan what he wanted as this is about evaluating the Alpha.

Import into sequencer. No errors from sequencer. Save Session 1. (As most of the work here is OTB the EDL isn't as important. Instead I'm dealing with sessions - each one ending in a discrete audio print/take of the processed audio.)

1st audio play - impression - nice original studio recording number . Generally well recorded although, to me, has a slightly emphasised mid band. Levels fine.

2nd audio play [focus on what Elysia may be for]

Glue - bring the recording together and make a little more cohesive. Possible to bring up bass and add more punch to the drums as this should work for this type of funk. Final levelling to bring up rms for the 44.1/16 wave. Overall I'm looking to use the Elysia then three times - normally for most mastering sessions I tend to do one analogue pass. Most, possibly all of these three could be done in the Elysia in one pass but I've opted to do each one separately as it's easier for me to clearly hear what it's doing where and how.



Session 1 is the original, unprocessed 96/24 wave.

Session 2
I've put the Alpha in a very basic compression setting to bring the recording together. Threshold is down at -25, moderate ratio at 1:2, release at 280ms.

Save and AB with Session 1.

Playthrough 2- see previous comments. Note I'm AB'ing what I've done against the original - no point in making a change if it doesn't actually improve on the original. The Alpha actually allows you to monitor both the unprocessed and processed signal so you can AB in the Alpha. However, to keep things simple at this stage I'm going to monitor via the second DA on our Lavry and put down a temporary print between 'sessions'.

Session 3
Drums and bass - as this is quite a nice funk groove I want to bring both up in the mix and add a bit more punch, particularly to the drums. It would be possible to do this using a split band or multi-band compressor but MBCs very often introduce unwanted artifacts in to a recording. It's usually best therefore to use an MBC rarely and with caution in mastering. I want to add a bit of weight to the drums whilst maintaining a sense of groove to the track. To do this I've focused a lot on both the release time and the gain reduction limiter and I'm using the side chain of the Elysia on LP.

Save as Session 3.

Playthrough 3 - generally better to me - bass and drums have come up and there is more punch on the drums and they sound 'bigger' with a bit more weight and ambience. However, think some of the warmth of the original has gone and the track now sounds a little bright.

At this stage I would normally use a separate eq - either our Ibis or, if ITB a Flux or Sonoris plug-in. However, the Alpha has an Audio Filter so I'm going to use that as a sort of shelfing eq to bring the low end up by about 0.5 dB and the high end down by the same amount with a cross-over at about 1k.

Playthrough 4 - better, track sounds warmer but Ivan's guitar sounds a little brittle. As I want to focus on his guitar I'm going to bring in an external EQ.

Session 4
Ibis added to signal chain. EQ cut -o.8 dB, medium q, 610hZ, bell. Note - normally I'd put an eq before the compressor but in this particular case it works better after.

AB - better, Ivan's guitar still, to me, sounds suitably funky but without the previous brittleness.

Save as Session 4

Session 5
Last bit for the Alpha - I'm going use it to level the track. I want to try to bring up the rms by a few dB and so give the track increased perceived loudness but without using a brickwall limiter or hardclipping. The Alpha has a soft limiting function and I'm going to set the overall gain stage so that fast transients clip out at -.5.I'm aiming at -.5 rather than 0.0 or -.2 (which I normally use) as I'll have to process the final track as an mp3 and LAME tends to add between 0.2 to 0.5dB to peak levels - if I aim at -0.3 LAME could easily force hard clipping and produce some nasty digital distortion on the mp3. [I did the gain stage here manually just by listening for overshoots rather than watching the meters].

Downsample to 44.1k dither to 16bit and print as CDA Wave for Ivan.

Final listen/quality check on 44.1/24 wav
Plays through ok but there is, to me a little distortion on a couple of fast transients at, for instance, 31 seconds. To me this seems to actually be intersample peak distortion and may have been partly as a result of the multiple, sequential sessions saves. Even though the set up, etc have all remained at 24/96 and no dithering has taken place until the final 44.1/16 print it's a possibility that the multiple saves has accumulated distortion (the final track is a forth generation.) A single pass through the Elysia Alpha and one session take would probably eliminate it.

44.1/16 bit wave sent to Ivan as a zipped file (approx 38 Mb)



Format change to mp3 VBR for GMC. Added linear fade out at Ivan's request.

Total time - approx 1 hour 17 mins. - slightly longer then normal as quite a bit of time was experimenting with the Alpha and some additional time connecting the Ibis in to the signal chain.

You are at GuitarMasterClass.net

Don't miss today's free lick. Plus all our lessons are packed with free content!

Don't miss today's free blues, jazz & country licks. Plus all our lessons are packed with free content!

This post has been edited by tonymiro: Feb 16 2010, 07:08 PM

[Saoirse O'Shea]

... most plug-ins etc. in the DAW domain doesn't operate at this sample rate (96Khz), I guess that f.ex a guitar with a delay added are already downsampled in the DAW ??? The same must go for every effect used in my belief, or am I wrong here?? Also, wouldn't the greater sample frequency put a heavier load on the computer, since a larger amount of data must be transferred between the discs and the RAM... ..... (just trying to motivate myself why I should record at 96Khz )

//Staffay

Most, maybe all (I can't think of one that can't) Mastering grade plug ins can be run in an over-sampling/high quality mode that upsamples it to 96 or higher. But you are right that most standard vsts don't/can't upsample.

With the Daw - the ones we use here for mastering we can set the sample rate and bit depth and it will then auto convert any stream to that on the fly. In this way you can have three audio tracks recorded at different sample rates - say 48 and 96 - processed by vsts at 44.1 and 96 and the daw will auto change the stream to what ever you tell it to. However, generally you don't want a lot of SRC going on so we tend to set everything to a single sample rate - usually 96 and so minimise the amount of src taking place in the daw - we don't let our daw downsample unless we need it to.

As an aside I can think of at least one major recording daw that routinely applies SRC and bit depth changes without user intervention - most users probably don't even know that it is tbh. (Not sure if it truncates or dithers for the bit depth though.)

Yes higher SR does increase the load on the pc. For mastering I'm not likely to have lots of lanes open and lots of plug-ins running and latency isn't much of a consideration for me. I have a quadcore i7, a lot of ram, and sata drives. The only software on that pc is the OS and mastering stuff - nothing else, no virus guard, no internet browser - nada. Background processes are kept to an absolute minimum. If I pull up a stereo main place it at 24.96 and put up say an LP EQ, a MP EQ, a compressor and a limiter and a dither plug the cpu trips up to between 15-32%. It's the consequence of running mastering grade quality stuff - it's hardware/pc intensive.

Other bit is file space - Ivan's number is just under 4 minutes and at 24/96 is about 110Mb. Five session takes and I've got 500Mb of audio stored on a hard-disc. We archive sessions for quite a while in case a client comes back to us in the future and wants a re-master. Fortunately hard drive disc space nowadays is pretty cheap .

One thing I find kind of interesting is that many home users know what the bit depth etc of their chosen daw is but know little about why and how it affects what they do.

You are at GuitarMasterClass.net

Don't miss today's free lick. Plus all our lessons are packed with free content!

Don't miss today's free blues, jazz & country licks. Plus all our lessons are packed with free content!