Skip to main content

Your input on my setup (mostly the bit reduction part)

Member for

16 years 7 months
I wanted to see if there may be a better way to configure my gear to promote better quality. Here is what I have:

External gear:

- Ensoniq ASR-10, MR rack, ESQ1 and DP/4+
- E-MU E5000 Ultra
- Korg MonoPoly
- Soundcraft Spirit E 12
- Apogee MiniME
- M Audio Delta 66
- M Audio CO2


- Reason 3
- Digital Performer 5.11


- iMac 17" 2.0GHz Core2 Duo (running DP and Reason)
- G4 1.2GHz 7455 (running DP and Reason)

The iMac is my primary workstation. The G4 is used for a dedicated Reason machine and down-sampling bounces.

I have the optical in and out of the iMac running to the CO2 to convert to coax SPDIF. The digi out of the iMac goes to the Delta card (SPDIF in) which is in the G4. The card is set to hardware route the SPDIF in to a pair of balanced 1/4" outs. This is my D/A for the iMac.

The input of the iMac is both the digital out of the MiniME and the USB out of the MiniME. Using the digi out lets me get 96/24 whereas the USB only sports 48/24 (though I mostly record at 44.1/24). This is my A/D for the iMac.

The A/D D/A of the G4 is the Delta card. This is where I want to get things double checked. I use the MiniME and G4 to down-sample by setting the MiniME to 44.1/16. This applies the UV22HR process to the incoming signal which is leaving the iMac at 44.1/24 via the digital out. I'm recording in at 44.1/16 for the final stereo mix. I do this to down-sample bit and sample rates as I suspect the math that is applied by Performer to sample and bit rate convert is not desirable if it can be avoided.

Is bouncing like this to down-sample bit and sample rates better that applying DP's math to accomplish the same thing? There is another converter hit to go D/A then A/D but I suspect the UV22HR process may be worth it. I used DP's math to down-sample my last album recorded at 48/24 and it was OK but it lost some clarity, separation.


Member for

16 years 2 months

RemyRAD Fri, 01/26/2007 - 20:49
OK, so I'm not a Macintosh user but what you're putting yourself through seems to be highly unnecessary?

All of this resampling, sample rate conversion, debt reduction, etc. just can't be doing anything good for your recordings.

What is the purpose of producing in this manner? So you think you are doing things with a higher resolution that sound better?? Sure it does, at the source but in the end, what do you get? Scrambled math!

Personally I like my sound Sunnyside up. And so, I really don't think twice about recording at 44.1kHz, 16-bit. I record, produce, deliver and it stays that way, even when it ends up as MP3/WMV/iPod/broadband cellphone. That's what we used today and yesterday. What are we going to use tomorrow? What do I look like? A psychic? OH! You want to be prepared? Don't worry about it. When the new standard comes out, we will all know about it and then, you can make more intelligent choice.

Until then, set all your stuff to the same sample rate and bit Depth. Suddenly, you will notice, everything you do, WILL SOUND BETTER! It's like magic! Just don't over or under record. You have 96 DB. Tell me that's not enough? What good is 24-bit 140 DB dynamic range when the analog portion of your equipment can barely deliver 100 DB of dynamic range/signal-to-noise ratio? Right. It really doesn't make sense does it?

Magnificent desolation (Buzz Aldrin)
Ms. Remy Ann David

Member for

16 years 7 months

zerosin Sat, 01/27/2007 - 11:23
So what you are saying is that I will not notice any improvement mixing 30+ tracks in 24 bit vs 16 bit? And that Apogee's UV22HR processing will not provide any improvement in the final 16 bit signal because I won't notice any gain from using 24 bit in the first place?

The process is really very simple and takes less time then waiting for my DAW to down-sample. The whole thing takes place in real time. On the last album I used DP to down-sample the final 48/24 stereo track to 44.1/16. It wasn't the fastest machine so it took around 20-30 minutes to down-sample the bit rate and the sample rate for a 7 minute song. This process takes 7 minutes for a 7 minute song and there is no math applied by the DAW. Just a D/A A/D hit and the UV22HR processing applied to the final 44.1/16 master track.

Member for

16 years 2 months

RemyRAD Sat, 01/27/2007 - 12:12
The point is not to leave the digital realm. So you can do things faster by just feeding things through and by decoding, reencoding, decoding, reencoding. Instead of waiting the few minutes for your computer to do the math properly. So in a sense, you really are digital. You are analog in your production thinking. You're defeating the purpose of the equipment you have.

My microphones and preamps are analog being converted to digital. All other work is performed digitally in the box and only after all other processing and mastering is complete, does it come out of the box digitally as a CD or other digital compressed or uncompressed format. It's not analog again until it comes out of the CD player. So I only have 1 encode and 1 decode. And that's all you should have also because what you are doing is just plain stupid engineering for somebody that requires drive-through audio at McDonald's.

You can be stupid but I don't have to be because I'm not
Ms. Remy Ann David

Member for

21 years 2 months

Pro Audio Guest Sat, 01/27/2007 - 16:36
zerosin wrote:
The process is really very simple and takes less time then waiting for my DAW to down-sample. The whole thing takes place in real time. On the last album I used DP to down-sample the final 48/24 stereo track to 44.1/16. It wasn't the fastest machine so it took around 20-30 minutes to down-sample the bit rate and the sample rate for a 7 minute song. This process takes 7 minutes for a 7 minute song and there is no math applied by the DAW. Just a D/A A/D hit and the UV22HR processing applied to the final 44.1/16 master track.

Sample rate and bit depth reduction by going from Digital->Analog->Digital is one of the worst methods available and should be avoided at all costs. Not one of the original samples in your recording will be preserved in the reduced version.

Imagine you're at a party with 100 of your favorite celebrities and suddenly they all turn into celebrity look-alikes. That's what you're doing to your recordings. I can't imagine why you would be losing clarity or separation... :roll:

Bit depth reduction is a simple, quick, straight-forward calculation using software and gives great results when applying dither appropriately.

Good sample rate reduction is cpu intensive but gives the best results IF it's implemented correctly (ie, massive oversampling of the reconstructed wave to provide integer ratio downsampling which preserves as many of the original samples as possible). Most sample reduction software today uses this method.

If you think that sample rate reduction takes too long, then find software that can do it in batch mode and let your computer do it while you sleep or do something else...

Member for

16 years 7 months

zerosin Sat, 01/27/2007 - 18:02
Bit rate reduction is fast, and seems to have little audible change to the audio. Sample rate conversion sucks, and you can only hope that the math is tip-top. This is why I am recording at 44.1 now instead of 48. I sample field recordings as 88.2 because the math is simple to down-sample to 44.1 and my source files remain in higher quality to maintain a high end sample library for future use as well.

My main goal was to apply the UV22HR process to the 24 bit files as they are converted to 16 bit. In part due to using Reaon 3 at 24 bit and tracking it in at 24 bit. I figured I'd keep some of its quality. To do this with the MiniMe I have to get the signal to the analog inputs of the MiniMe. Sounds like this may not be worth it though. DP's signal processing and down-sampling truncation have always been pretty darn good so I guess I won't sweat it.

Though this process is essentially what would be done if one were to use an analog summing bus for the final mix. Except in this case, I would not need the summing bus but the amount of converter hits would be the same. Initial A->D then D->A->D. Never mind the fact that most of my instruments are based on sampled instruments, so add an A->D->A before the initial tracking in for each instrument in my rig. I use outboard FX so I am going D->A->D for that as well. The same thing happens even if you are completely analog and use outboard solid state FX. A->D->A.

Thanks for the input.

Member for

17 years 8 months

Cucco Tue, 01/30/2007 - 06:12
Okay....a LOT of things going on here that need attention.

1 - it doesn't look to me as though you are going into the analog realm at any point. You're exiting in one digital format, converting that digital format then using another hardware device as sort of a router then entering yet another digital device to perform your dithering. There's absolutely nothing wrong with that except for the fact that it's wholely unnecessary.

2 - There is absolutely NOTHING wrong with using analog as a form of dither. Many if not most mastering houses leave the digital domain and re-enter after analog processing. Many of those mastering houses re-enter the digital domain at 16 bit, not 24 (personally, I prefer the sound of my dither algorithms to straight 16 bit, but this is a choice, not a scientific necessity).

3 - Converting from 88.2 to 44.1 is no more simple math than converting from 48 to 44.1. This is a misconception - that you only have to remove every other sample...nope. The math that is involved is not very complex at all and any standard FP CPU can handle it with ease. (You're basically doing multiplication of two large numbers then dividing it by another large number.) However, I'm a BIG fan of 44.1 unless otherwise needed. And unless you're doing video, please don't do 48 - it's not necessary. For high-end projects (things that will end up on DVD-A or SACD or even DVD-V, I'll bump it up as high as needed, otherwise, if I know it's bound for CD, I'll leave the recorder at 44.1).

4 - Chances are, your recording software has some dithering algorithm. I would use it. I know you'd like to use the algorithm in your Apogee, but put simply, the steps you're going through are quite excessive. I have yet to have a client ever state:
"Wow, I love the recording, but I'm not a big fan of the dithering....couldn't you have used POW-R 3 instead of UV22HR?"
(The day that happens is the day I retire!)

Anyway - I like your creativity in finding a solution. That kind of innovation is what the recording industry needs. Keep it up!