16/44 vs. 24/96 comparison

If you have a universal player you can hear the difference with the Hires Music LA Four disc. I may differ with Barry here one 96 v. 192 or 176. I find there to be a slight improvement with the higher sampling rates.

 

Someone said that winamp automatically downconverts to 16/44 on the fly--really? Even if you select "24-bit playback"?
What about Foobar?

My soundcard (an Audigy) supports 24/96 playback, so how do I avoid this?

 

That's the first I've heard of this. I use Winamp all the time and I've done comparisons with Goldwave playing hi-res files and never noticed any difference.

 

There are some "fan made" recordings using a microtrack from M audio. You can record '"things" in 24/96.

Some/most of these recordings are downsamples to 16/44.1 to burn on a standard CDr.

I make DVD-As from the 24/96 files and have compared them to the 16/44.1 on a CD.

The 16/44.1 sounds like a transistor radio comparing it to the 24/96.

Just my 2 cents.....

 

Hi Mal,

I currently record at 24/96. (My converters will do 192 but I am still testing this.) In my experience, audio encoded at 24/96 will always sound better than audio that is originally encoded at 16/44 too.

Well actually, I'd need two identical converters, would have to test them to acertain they are identical (or close enough) and I'd have to split my mic feed. This would serve the interests of scientific discover, no doubt. But it would not serve the fidelity of that particular recording.

My mics feed the converter directly, with no splitter, console or anything else between the two. Perhaps one day I'll find the time to conduct such a test when the recording itself is not critical. With my current schedule and time being as short as it is, I need to ensure that when I make a recording, I'm getting the best signal I possibly can. For me, that means nothing between the microphones and the mic-pres/A-D converters.

I understand what you're saying and agree in principle.
Do keep in mind however that the 16/44 version was done with an src algorithm I have tested and found transparent - I don't find most to be even close. The primary "loss", if we can call it that, would come from the dithering. In this case, with Pow-R, one of the best (i.e. most transparent) I've heard so far and in any case, the dithering resulted in a 16 bit file that sounded a lot more like the original than not dithering. Based on my experience, I would go so far as to say the 16 bit file derived from the 24 bit original sounds a great deal more like the high res version than one that would be encoded directly to 16/44.

Also, both files came from the same mastering, which is something I've not encountered before in any comparo involving high-res material.

You raise valid points of course. Hopefully, these files will prove as useful to others as they have been to me. Note that before creating the final 16/44 version, I created several, using different combinations of src and dithering algorithms, which I am always testing (and beta testing for a number of designers). I chose the ones I found to be the very best after extensive evaluations.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

This all depends on the soundcard and the driver software. This is especially true if your soundcard is a Soundblaster. Up until the Audigy 2 ZS and Xfi cards, a lot of their 24/96 playback ability was compromised. The chips were there, but the signal path actually ran through 16/48 chips too, so the end result wasn't truly, unmutilated 24/96 data.

If you have Foobar, I suggest using the KS or kernel streaming output driver. There's a mention in the software about this being beta and unpredictable, but on my system, it sounded noticably better when I had an Audigy 2 ZS. I even notice a slight improvement compared to the WDM or Directsound output drivers with my current soundcard, an E-MU 1212M. It's most apparent in the high frequencies, which sound much finer, sharper, more focused and smooth.

I'd avoid the Directsound or WDM drivers for playback since resampling often takes place inside these drivers. The most reliable driver short of the kernel streaming one would be the ASIO drivers. I know Foobar v9.x supports ASIO so you could try that. I'm not that up on Winamp any more. Supposedly hires drivers exist, but I can't speak for their quality. I do know (and love) Foobar. I can tell you that it does 24/96 very well. I even use it to upsample 16/44.1 material to 24/96 and just leave my E-MU driver set all the time to 24/96.

Regards,
Steve.

 

Hi gloomrider,

It would be a mistake to say I'm "not quite sold" on 192k as an improvement over 96k. Much depends on the converters and the filtering. What I've said is that I'm currently testing this out and have not reached any conclusions yet. (I tend to take my time with these sorts of things.)

The ULN-2 is fabulous, nothing less. I use one and love it. I'm also testing another device that can do 192. When I've reached a conclusion one way or the other, I'll report. Right now, if I was going to record this afternoon, I'd be doing it at 96. Not because I find this better but because I'm positive about the results I can get with it and have not completed testing on something that may in fact turn out to be even better.

Hope this helps.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

Barry,

I find more "air" around the instruments at 24/176 or 24/192. Do you find that as well?

 

Barry, this mirrors what I've observed using modest home equipment (although still "pro" quality converters, an E-MU 1212M) and I believe Metoo has also noticed that the same with his E-MU 1616M (which has the same converters). I recall experimenting with some SACDs I wanted to convert to MP3s for my pocket player. I recorded directly to 16/44.1 and then to 24/96 and downsampled using Cool Edit Pro and 0.7 bit triangular dither. The resulting downsample sounded noticeably better than the direct 16/44.1 files. The biggest difference was in the highs, which were significantly muffled on the direct file when compared. With ideal studio conditions and high-end pro hardware, the difference is probably much more apparent (I'm anxious to get home and hear your samples).

 

You should notice better soundstaging as well.

 

Well Steve you've gone a bit over my head.
I have a basic Audigy, the SE. It says it can do playback of 24/96. How do I know it's for real?
I've been using Foobar over Winamp, and have selected the 24 bit playback; how do I know which driver (KS or whatever)?

 

huh? Foobar over winamp? I think you just worded that ambiguously- anyway, if you are just using foobar, go get the asio plugin, and use that. It will force it to use 24/96 and bypass any of the internal creative filtering (and the eax, and all that other crap).

 

Sorry, I mean, I have both Winamp and Foobar, but use prefer to use Foobar. I'll look for ASIO plugin.

 

Try the KS plugin. You can find it in the Output section of the Preferences (Ctrl+P, find the Playback branch of the tree on the left of the Preferences dialog box and expand it to show the Output branch. Select that and look at the Output Device dropdown dialog box on the right side of the box at the top.

The KS driver bypasses all internal drivers and writes the data directly to the soundcard hardware. Second choice after it would be an ASIO driver. To do that, you need to select the ASIO Virtual Devices branch under Output then on the right, click the Add New button. If your card supports ASIO, you'll be able to set up an ASIO "device" in this dialog. If not, you'll get a message "No ASIO drivers found.

Even if you don't have an ASIO driver, you can still emulate ASIO using a freeware program called ASIO4ALL. This emulates ASIO on Windows machines and will basically provide an ASIO-type input and output to the limits of your soundcard's abilities. It's not as good as real ASIO, but it's better than Directsound, from what I've read.

This may all seem like a lot to digest, but it's really not that complicated once you practice a bit. The reward will be the best sound possible from your specific equipment.

 

Hi Lee,

Sometimes. I'm not positive (yet) this is real air.
There are occasions when spurious harmonics can lead folks to the conclusion they're hearing more information. They are but it isn't information that entered the microphones and hence is not, to my mind or ears, real. (This can happen with some balanced to unbalanced cable adapters too -if they are not scrupulously wired.)

Most converters are also not as good jitter-wise when asked to go faster (i.e. higher sampling rate). I'm currently listening to a most promising proto-type. Will report when I've reached a solid conclusion.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

Hi Steve,

Are the highs muffled on the direct file or is there a rising treble on the converted file due to the addition of noise shaping, which is part of many dither algorithms?
Or perhaps, a bit of each?

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

Thanks Barry. Look forward to your thoughts...

 

I think the noise shaping adds some information, almost like a sense of air (but constant and not to be confused with the sense of air surrounding instruments), but in comparing both to the original, the direct file sounded muffled, whereas the downsampled file sounded like a slightly fuzzy or "blurred" version of the original but with the same basic frequency response (there's another computer graphics reference. I really do use Photoshop too much in my work! . The downsampled versions also seemed to retain a little bit more of the solid transient thumps from the bass

 

I've tried EQ'ing before the dithering, and it doesn't work for me.

Not to skew the topic, but i've had success preserving ambiance while processing in 16-bit by adding a slight bit of reverb and EQ to the sound going into the soundcard.

 

Hi Grant,

I guess I got lucky. Outside of using SRC to get the sample rate down to 44.1k and PowR-3 to dither down to 16 bits, I didn't do anything at all for the CD version of "Lift" (very much on purpose) and I'm very pleased with how the original sound of the church is there to be heard on the pressing.

Hearing the CD on a system with outstanding soundstaging and imaging capabilities (using Nola Exotica Grands), everyone was exactly where I remember them, in full 3D (no flat images) and even the back wall and side walls of the church are in full evidence.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

Again, you did a great job on the CD, considering you did almost nothing to manipulate the results of it.

 

Well that's not preserving the ambiance, it's adding some artificial ambiance.

's alright I'd do the same sometimes. If it's for your own listening pleasure, do whatever makes it sound good. If I'm needledropping an especially noisy record that requires so much NR it zaps the life out of something, I'll add a bit of gentle exciter to restore some sense of the highs and then very slight reverb using an impulse response from a real space or from a Lexicon 960. The results might not rival an original recording, but they often sound much more pleasing to the ear than that "recorded live in the vacuum of space" sound!

 

Hi Grant,

Thank you for your kind words.

If I was to attribute anything to the sound of that recording, I would say it is exactly because I did "almost nothing to manipulate the results".

That is the idea behind Soundkeeper Recordings. It is a test of how well I can get the gear "out of the way" and then follow it, leaving just the performance. (This asks much of the musicians. They have to be able to perform a three minute song in three minutes. )

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

Same here, this is why I normally record at a higher resolution than I plan to end up with. If it is 96/24 I, thus, usually start with 192/24. To some people this might seem like overkill, but I have found that the same thing Barry mentions when going from 96/24 to 44.1/16 is present when doing that: you get a sound closer to the higher bit rate (although still quite not there), but always better than recording directly at the target rate and bit depth. Just my 2 cents.

 

And slightly smoother sound with a bit more detail.