Who's tracking @88.2 and 96 KHz sample rates?

[malgfunk]

To date I've always tracked my projects @48KHz using Apogee Symphony using Rosetta 200 and DA-16X(s) and very happy. However, I'm getting some external projects farmed in that's @88.2 and 96Khz which seem to sound more smooth on the high-end with detail. I just want get a sample of who's using higher sample rates and why.

[raal]

i've been doing 96kHz for quite a long time now. i think it sounds better. nothing scientific to base that decision on, just my ears - haven't done blind A/B calibrated tests, so please take this as an internet opinion, nothing more.

[MichaelPatrick]

I almost always track at 88.2 or 96. I think many converters are optimized to work best at 48 and/or 96, or at least my ears tell me they are. I've recently changed my default tracking rate from 88.2 to 96 even though downsampling takes longer.

My reasons are like yours. HF detail and smoothness. If you want more control of high frequencies in the mix you need the extra information that you can't get at 44.1 and probably not even at 48. Most plugins just sound better at higher rates and when they get to work with the extra HF info.

I use 96 now for most everything. I'd never track at 44.1 except simple speech like a lecture. 48 for low end quick and dirty demo projects.

Since I'm on a PC, I use R8Brain Pro to downsample to 44.1. With the minphase and ultrasteep settings engaged the final prints retain virtually all of the HF detail and the precise tonal balance of the 96kHz masters.

[thethrillfactor]

When i can get away with it i do all live recording at 88.2khz. The end result even on an MP3 always comes out better sounding even though at a certain point i have to sample rate convert to 44.1khz for a CD.

My biggest issue though is during the mixdown most of my digital effects boxes that are connected digitally through AES won't work because the highest sampling freq they work is 48khz. So my processing choices are halved or i have to run them analog back into the converters which takes away I/O's i could be using for other gear. Basically i go from (7) pieces of processing gear connected by AES down to (2).

Oh well...i guess it will be my excuse next year to buy a Lex PCM-96 and a Bricasti M7.

[kittonian]

I track and mix everything at 96kHz. The plug-ins just seem to respond better at that rate and from what I've understood from a few software developers, 96Khz is the optimal way to run any plug-ins.

I'll be honest. Even when I get a project that's been recorded at 44.1/48 I always upsample it to 96kHz to mix it. Then I use Barbabatch to convert it to CD and MP3 quality. Sounds great to me.

[kinghand]

I've been working with 88.2k sample rate for a long time, it does make sence, I'm sure!

[u b k]

88.2k, the top can withstand a digital process or two and still sound analog.


gregoire
del
ubk
.

[henryrobinett]

Primarily 88.2.

[s12512]

still use 44/24!! most of the hits now are still being made on 44 or 48. for movies most of everything is being made on 88 or 96. kinda just depends on what ur doing. even if ur not doing movies its still nice to have the ability to do up to 96. even though most people that are puttin out hits everyday dont use it. i prefer good converters over 88/96.

[TornadoTed]

I've done my last 3-4 project at 96kHz and definately prefer it for the reasons stated by others.

Smoother top end and plugins sound better.

[andsonic]

Since we upgraded the computer, we've been running 96k.

I used to track at 48 & mix at 88.

Now I do everything at 96. Upsampling has worked for some projects that were originally at 48.

I'm tracking a Blue Grass album right now at 96. It sounds stunning.

[PheelTheMusic]

No disrespect intended, but can one of you 96k people tell me why you work at 96k rather than 88.2 when the "digital gods" agree that 88.2 to 44.1 is an easier conversion than 96 to 44.1?

[Duardo]

No disrespect intended, but can one of you 96k people tell me why you work at 96k rather than 88.2 when the "digital gods" agree that 88.2 to 44.1 is an easier conversion than 96 to 44.1?

"They" haven't agreed to that for a long time...with most current sample rate converters it's not an issue...

[MichaelPatrick]

Originally Posted by PheelTheMusic

No disrespect intended, but can one of you 96k people tell me why you work at 96k rather than 88.2 when the "digital gods" agree that 88.2 to 44.1 is an easier conversion than 96 to 44.1?

I used to record virtually everything at 88.2 and I agree with Duardo. Conversion is easy from any source rate if you have an excellent SRC.

Sound is far more important to me than ease of conversion.

Let's face it. 44.1 is the redbook specification for CD media and 88.2 is with us only because it's convenient multiple of 44.1.

48 and 96 are standard professional rates for film and broadcast audio.

We know 48 sounds better than 44.1 because there's more information conveying HF fundamentals and harmonics.

IMO plugins and converters generally sound better at 96 than they do at 88.2. Perhaps it's because the designers focus more on performance at the professional sample rates. I don't really why, it's just what my ears tell me after many hours of critical listening.

[Titans55]

I find myself always recording at 88.2. It is easier to convert down to 44.1 (cd). Personally I do not feel their is not a lot of difference in sound between 88.2 and 96 until you get to mixing. I agree with the above posts that 96 sounds better with pluggins.

[Mike Derrick]

Originally Posted by MichaelPatrick

Let's face it. 44.1 is the redbook specification for CD media and 88.2 is with us only because it's convenient multiple of 44.1.

Actually it's not true that it's more convenient.

SRC conversion formula involve oversampling to a common multiple regardless of source sample rate and destination sample rate. Therefore no sample rate conversion is more mathematically more convenient than any other. (see page 295 of Digital Audio Explained by Nika Aldrich.)

From null tests I've done, and of the plugins I checked, some plugins do process the exact same audio differently at 44.1kHz vs the same audio with the same plugin settings at 96kHz.

I have recorded at 96 and 88.2, but also record at 44.1, depends on the scope and needs of the project.

[tsvisser]

48KHz for the most part (mostly film/video) or 44.1KHz for music. If I know the performers are good and my setup is good, like I have enough time and resources to put a lot of effort into it, and not simply close mic'ing everything - a true stereo or surround recording I will do 96KHz, but ONLY if my client asks for it. If its for me, of course I do it, but I would never just deliver 96KHz without a specific request to do so.

[MichaelPatrick]

Originally Posted by MichaelPatrick

Let's face it. 44.1 is the redbook specification for CD media and 88.2 is with us only because it's convenient multiple of 44.1.

Originally Posted by Mike Derrick

Actually it's not true that it's more convenient.

But convenient for who? Users? SRC designers? Both?

I think both. Most users like me find it faster to downsample from 88.2->44.1 than 96->44.1. And it's simple logic that multiples of 2 are easier, i.e. more "convenient," for SRC designers.

My point is that we probably wouldn't even have 88.2 in our converters if not for the Redbook. Instead we'd probably have 64 or 80, which would make a designer like Dan Lavry and a user like me happier. Dan thinks it's sonically optimal and I'd like smaller file sizes and reduced CPU loads.

I could be wrong. If 88.2 isn't a convenient accomodation to the Redbook's 44.1 rate, why do you think PCM converters all have 88.2?

[MichaelPatrick]

Originally Posted by tsvisser

... I would never just deliver 96KHz without a specific request to do so.

For me 96k is just easier to work with through the production process making it easier to give customers a better result. Like you, I keep source tracks unless they ask.

[Mike Derrick]

Originally Posted by MichaelPatrick

I could be wrong. If 88.2 isn't a convenient accomodation to the Redbook's 44.1 rate, why do you think PCM converters all have 88.2?

for the clock. (not src)

[MichaelPatrick]

Originally Posted by Mike Derrick

for the clock. (not src)

Do you mean that designers use just one crystal or PLL (clock) for both 44.1 and 88.2? Is that why virtually all PCM converters support a rate of 88.2?

[Mike Derrick]

Originally Posted by MichaelPatrick

Do you mean that designers use just one crystal or PLL (clock) for both 44.1 and 88.2? Is that why virtually all PCM converters support a rate of 88.2?

As far as specifics, your question is probably better answered by Dan Lavry or other clock designers, but yes, I believe you are on the right track.
It is my understanding that most gear will have two different crystal clocks to support the base frequencies of 44.1 and 48. And that those crystals oscillate at 256x the base rate in order to cover the multiple of sample rates above the base rates.

[dkatz42]

Originally Posted by MichaelPatrick

BI think both. Most users like me find it faster to downsample from 88.2->44.1 than 96->44.1. And it's simple logic that multiples of 2 are easier, i.e. more "convenient," for SRC designers.

The algorithms are nearly identical. The basic algorithm is to upsample to the least common multiple frequency, which is done by simply inserting zero samples. Then you do a digital low-pass filter to eliminate any material higher than the Nyquist of the final sample rate. Then you decimate by tossing samples.

So to go from 96K to 44.1K the code inserts 440 samples of zero between each original sample, runs a low-pass filter on the resulting stream, and then tosses 959 resulting samples for each one kept.

To go from 88.2K to 44.1K you simply skip the insertion of zeros step, but you still do the low pass and the decimation (but only one sample is tossed for each kept.)

What is true is that the 88.2K version is going to be significantly faster, since the low-pass filter is run on 1/441 of the amount of data relative to the 96K version.

But neither is "easier" to the SRC coder since it's the same code for both, skipping the zero insertion step.

[MichaelPatrick]

Originally Posted by Mike Derrick

As far as specifics, your question is probably better answered by Dan Lavry or other clock designers, but yes, I believe you are on the right track.

Thanks. I thought that's what you meant and it's what I thought too. And like you, I'll defer to Dan on this stuff any day. Reading his papers and posts have made me long for a 60kHz sample rate.

[MichaelPatrick]

Originally Posted by dkatz42

The algorithms are nearly identical. The basic algorithm is to upsample to the least common multiple frequency, which is done by simply inserting zero samples. Then you do a digital low-pass filter to eliminate any material higher than the Nyquist of the final sample rate. Then you decimate by tossing samples.

So to go from 96K to 44.1K the code inserts 440 samples of zero between each original sample, runs a low-pass filter on the resulting stream, and then tosses 959 resulting samples for each one kept.

To go from 88.2K to 44.1K you simply skip the insertion of zeros step, but you still do the low pass and the decimation (but only one sample is tossed for each kept.)

What is true is that the 88.2K version is going to be significantly faster, since the low-pass filter is run on 1/441 of the amount of data relative to the 96K version.

But neither is "easier" to the SRC coder since it's the same code for both, skipping the zero insertion step.

That's more detail than I can understand, but what I get of it clarifies why it's not hard to convert uneven sample rates. Thanks!

[ISedlacek]

I do everything in 96 kHz. Sounds a way better than 44kHz and still better than 88kHz to me ...

[drundall]

Just got a session today to do overdubs on that was 192Khz. Sounded good. Pain in the ass though.

[Chapiro]

Ever since PT hd came out i started tracking and mixing at 96 k 24 bit for about 2 to 3 projects and then switched to 88.2 it sounds better to my ears and its ready to take advantage of any high resolution format.

[MichaelPatrick]

Originally Posted by drundall

Just got a session today to do overdubs on that was 192Khz. Sounded good. Pain in the ass though.

It's a good thing disk drives are cheap these days. You may need yet one more

[Flying_Dutchman]

+1 96kHz