Downsampling

Hiya!
After sending some recent work to my ME, I asked for some feedback. He told me that the mixes were pretty impressive and wondered why I recorded at 44.1 and not 88.2 or 96k. I told him simply because I use UAD plugs a lot and using a higher sampling rate would be too much of a headache.

He told me to record at a higher sampling rate and downsample when it comes time to mix.

How would I go about doing this. I use Sonar 7, and I'm not sure how to do this....

ORIGINAL: fooman

Hiya!
After sending some recent work to my ME, I asked for some feedback. He told me that the mixes were pretty impressive and wondered why I recorded at 44.1 and not 88.2 or 96k. I told him simply because I use UAD plugs a lot and using a higher sampling rate would be too much of a headache.

He told me to record at a higher sampling rate and downsample when it comes time to mix.

How would I go about doing this. I use Sonar 7, and I'm not sure how to do this....

The only reason to record at a higher sample rate is if your tracks sound bad now, and would sound better if you recorded at the higher rate. You would need to do a test with the gear you have and verify that you can hear a difference, after that decide if that difference is required or desired for the type work you do. Using a higher rate just to be doing it doesn't make sence. After all there is some change in quality when you have to downsample so you also have to consider if the gain in quality you got on the tracks is going to make it through that. There is a chance it will sound worse if the downsample does more damage then the tracks did good.

Also be aware that resampleing will thorw off lip sync if you do sound for film, so don't record at a rate other then what the final product will be.

Back to quality. Some folks do record at higher rates because the sound card they have is fairly cheap and 44.1 doesn't sound as good as the higher rates. However, if you have a card with really good converters (Lynx, RME) 44.1 may sound so good there is no reason to go higher. In that case just upgrading the sound card might do more good then going to a higher sample rate.

There is one exception here. If you mix with an analog console and A/D again on the other side then there will be no resample or downsample. In that case it might be good to record at the rate your converters or sound card sound best at (that's not always the highest rate) but most folks mix in the box now so if you record at a sample rate other then what the final product will be you run the rist of damage from the downsample.

Yes, higher rates are available now and almost all software, plugins, and hardware can do them. But that alone is not a reason to use them. You need to verify that it really is better all the way through to the final product, not the first stage.

It's being done all the time. Most of us, I think, record at 48 Khz and 24 bits. Exporting the project you just choose the sample rate and bit depth you want (normally 44,1/16).

I would not recommend the sample rates you mentioned, though. You don't actually gain anything by going over 48 Khz IMO. But many do use 96 Khz, which takes quite a slice of HDD.

This is very basic. You'll find info in the manual, I suppose.

Firstly, in my humble opinion, the higher sampling rates are really not needed. It all depends on what style of music you are doing. My personal preference is 24/48. Anything higher is really not needed for the music I play, which is rock/metal. The higher bit/sample rates are incredible for cleaner music. Some jazz, recording an orchestra, acoustic instruments that have more dynamics than volume....that's when I'd use the higher ones.

But downsampling is pretty simple in Sonar. It's called "dithering" and you will see it in your options when you go to mix down your project if the project is at a higher bit/sample rate than 16/44.

But lets start from the beginning. First you need to set up your soundcard for the higher bit/sample rate. Go to your card options and set the bit rate for 24, the sample rate to 48. Since you use UAD (I do too) go with 24/48. Then open Sonar and go to advanced audio options. Set bit rate to 24, sample rate to 48000. Then, click the wave profiler button so Sonar finishes setting everything up for you. You may need to adjust latency settings if you are using WDM, or your sample rate setting in your ASIO control panel once you make this new change. Next, after wave profiler is done and you click ok and it exits, click ok on the audio options panel to get out of it. You will most likely have to restart Sonar.

Once you restart, go to options, then global. Once there, go to the audio data tab. Select 24 as your record bit depth, and you can either set render bit depth to 24 or leave it at 32. Click ok and exit, and you should be in good shape as far as setting up goes. Now there are a few ways to dither and downsample, but I'll give you the easiest one as this can be an artform all in its own.

Once your project is ready to be exported out of Sonar, and you click on the export option, you will see options in that box. One is dithering. There are a few ways to dither and there are good tools you can use to do this, but for quick purposes you can just use Sonar to do your dithering. In a nutshell without getting technical with you, dithering assists with proper downsampling and allows the file to keep its original sound as it is downsampled to a lower bit and sample rate. Set your dithering to like powr3, and then change your bit and sampling to 16/44 then click export. When the file is done being exported, it will be a stereo 16/44 that has preserved your 24/48 sound quality. Like I said, there are other ways to do this, but this will get you started. Hope this helps, good luck!

Note: One other situation (other then an analog mix) you might want to use a higher rate. Live to two track (stereo). If you are recording a live performance where there will be no more tracks added (no mix) then you might want to consider using a higher sample rate (and at 24bit !). You should get the best quality your converters will do and the recording ready if at some point in the future there is a popular consumer media format that is higher quality then what we have now. Then do all your (processing) mastering at that rate and resample down to the target product rate as the next to last step. Then bit reduce with dither and you and you have the final master.

With modern converters and sound cards the big gain in quality is going from 16bit to 24bit, not the higher sample rates. So you always want to go 24bit when you can no matter what sample rate you use.

What I was told to do was downsample prior to mixing so I could use the UAD stuff when mixing without much issue.

I understand how to set everything up, and how to do mixdowns. This is all simple.
But to record at one sample rate and mix in another is something I've not done.

I will experiment. I have Lynx Aurora 16's convertors... and I've never actually recorded above 44.1k. It might be worth my while, might not be.

Foo, downsampling prior to mixing is not a good idea. You can use your UAD's at 24/48. The higher bit/sample rate gives you better quality. You WANT to mixdown during this time and dither/downsample to the lower bit rate to still keep the original quality. I'd say either your friend doesn't really know what he's talking about, or you've misunderstood him...unless I've been doing it all wrong all these years, I can't understand why you'd want to record at one bit/sample rate, and then mix at a lower one. It makes absolutely no sense at all man.

great thread guys.....well informed stuff in here. This thread should help lots of new people coming to the forum.

Figured... didn't make much sense to me.
I think UAD needs to come out with a new card.... but thats a whole 'nother thread ;)

ORIGINAL: Kalle Rantaaho


Most of us, I think, record at 48 Khz and 24 bits. Exporting the project you just choose the sample rate and bit depth you want (normally 44,1/16).

naaayy -- i bet most record at 44 - at least i do