Pianoteq Pro offers the "internal" sample rate of 192kHz (which stage and standard do not include).
Have you tried the "internal" sample rate of 192kHz and if so what do you think of it? Can you hear the improvement with 192kHz vs. 48kHz internal sampling rate in a blind test?
Osho
Pianoteq Pro offers the "internal" sample rate of 192kHz (which stage and standard do not include).
Have you tried the "internal" sample rate of 192kHz and if so what do you think of it? Can you hear the improvement with 192kHz vs. 48kHz internal sampling rate in a blind test?
Osho
I have never tried it but a couple of years ago a user provided some files testing this thing and I could hardly hear any meaningful differency. Higher sample rates does not mean better quality and this is also true of Pianoteq. I can barely hear up 16-17 kHz anyway...
Thanks for the input.
Anyone else? I am interested in just playing piano - not in mixing it with effects etc..
Osho
Well, if you just say hypothetically there is a difference, only a body having unusually keen hearing might hear it and through headphones or monitors capable of 192kHz output.
You've likely heard another question, "When a tree falls while no one is around, does it make a sound?"
Some argue a 192kHz sample rate is affected mostly by filters.
I tried to find differences between the different frequencies of internal calculations. The difference was found in the long decay of complex polyphonic chords. I found more naturalness in sound space. But this difference is subtle. (After 20-30 seconds of sustain, differences can be discerned). It's not about the presence of high frequencies, but a 4-times more accurate calculation of the available frequencies.
in fact, I met more differences when I tested the PME Babyface sound card. Then I heard more details compared to my card.
As a result, it’s more important for me to go to such sound settings that will make the delay as small as possible. This is 96,000 for the card and 64,000 for the internal calculation in Pianoteq. As a result, I have 1 ms delay. And I catch the crunch very rarely.
Too often I fail a blind test to be certain that I like it more.
To me, the great advantage of 192kHz is that it allows overhead for less lossy destructive editing and mixing but the most important part is when mixing down from a final master to a consumer rate like 41kHz.
On outputting a lower res file like that, depending on your DAW or mastering software you can choose from a variety of dithering options. That's 'kind of like' choosing how compressed you want the final file to sound (jagged edges, anti-aliasing, blocks and altered color blocks, simplified to patterns, less colors in the palette etc.).
In the case of certain final file types like MP3, you might not want to use certain dither or related options for a cleaner copy. MP3 for example has its own compression mostly out of our control in that process, so less dither is likely better. Outputting to WAV or flac or others without lossy compression will possibly sound better if particular output choices are made to that format. Different music may inform those choices.
For synth projects I sometimes like working at 48kHz (destructively editing for effect) then choosing a dithering and other output options which add the most interesting digital texture to the sound, hinting at the glitch scene I suppose, or the way 8bit art plays with digital artifacts from the past. Thinking of it as a digital form of valve distortion.
When loaded in the DAW at 192kHz I love it and like scherbakov.al mentions, the tails of the sounds do feel different - but I do think I've felt it's different on different systems (capable of 192).
Systems with soundcards only capable of lower rates are probably giving very processed results during the 'fake' digital to audio downsampling. The audio gear needs to be capable of 192 to hear 192.
But once it's been output at 41kHz with rational settings, I fail to tell the diff too often to know for sure if I can discern the diff more than a small % of the time.
Science tells us our ears are pretty universally at a peak with 41kHz and I'm actually fine with that now. In the past I would really care more to demand 192kHz from players - but now it's pretty moot.
Can you hear the improvement with 192kHz vs. 48kHz internal sampling rate in a blind test?
No.
What makes no difference are broad claims against hard evidence. I need to see some proof. Speaker and headphone models will do! Interfaces will as well! Incidentally, according to a new version of your user manual, both Standard and Stage have been limited to a 48kHz sample rate as of PIANOTEQ v.6.7.
Now I’m saying show me the facts, your entire signal chain. When you claim you hear a difference, please let’s start with the microphones you used to conduct your experiment. Those you omitted from your entry (report), as you reported your findings.
If you selected one of the modeled microphones, you’ve neglected to include any info from its original manufacturer’s spec sheet, whether or not it’s rated for higher resolutions upwards of 41.1kHz to 192kHz —and maybe beyond. That might support your claim. If you say you used a perfect microphone, you assume it’s perfect for 192kHz as an input. Although its specs in this regard are nowhere mentioned inside your user manual.
Do you assume an internal sample rate naturally means the same thing externally? Possibly, it is available as an option merely to meet DAW requirements and plugins that might otherwise use any upsampling.
Oh, another word about sample rates, somebody confuses them with bit rates that can necessitate a dither pattern in high frequencies. Like, FLAC, AIFF, WAV, OGG, AAC, MP3, and a lossy format dither is of no consequence whatsoever while you’re playing especially.
There is a reason why 48000.
A little video on the argument which perhaps will help unaware users understand better.
https://www.youtube.com/watch?v=-jCwIsT0X8M&t=1362s
Interesting topic.
Apart form the internal sample rate (where this topic is about), there's also the sample rate in the 'Devices' section. This sample rate is connected to the audio buffer size and the latency.
Now, I'm completely unfamiliar with sound engineering. My question is: what is this sample rate? What is the difference between this one and the so called internal sample rate? Also, what is the relation between sample rate and audio buffer size?
Sorry if I somehow hijack this topic, but I cannot find an answer on this question. Neither in the manual, nor somewhere in this forum.
Thank you Chopin87, for such an excellent video.. really shows what I'm talking about (around the 10min mark).
Oh, another word about sample rates, somebody confuses them with bit rates that can necessitate a dither pattern in high frequencies. Like, FLAC, AIFF, WAV, OGG, AAC, MP3, and a lossy format dither is of no consequence whatsoever while you’re playing especially.
Incorrect Ra, thanks.
If you go the that 10min mark, you'll see the type of signal manipulation I was referring to, (it shows it more clearly than anything I've come across describing it) grit at lower rates, cleaner at 192 in the DAW, for example for some high freq. esp. pertinent to me with some synth things but can be applied to thinking about 192 (or up-sampled FX) for piano.
From there, the music and type of manipulation you made at that stage will inform your choice of dither type on the end, including from those which may involve some variety of noise shifting (moving noise up to non-human audible range) - or choose no dither in the case of keeping another kind of clean output - really - that's absolutely beside my point Ra - and a choice at the end I alluded to.
Hoping it shows that even if you have something you love the sound of in the DAW at 192, you can output it to something sounding just as good at 41.. but it may depend on your entire chain of processes to produce that.. in other words, some poor choices along the way can make for better or worse 44.1kHz vs. 192kHz comparisons.
The main reason it's a mystery topic is because it's full of technical intricacies IMO.. but I love that I can still fool myself with all kinds of sample rates and outputs.. it's really a constant joy really to wrangle all these things for effect - and right or wrong, choices can be artistic or technical - I like a mix erring to taking some artistic liberty with all of it, which is more fun for me than being very clinical along the way (comes from analog days - also I wouldn't be surprised if I occasionally quoted incorrect things - I'm not googling it like some others).
@Viridis - buffer size relates to how audio is handled by CPU, the size of 'info' segments to be consumed (kind of how much is jammed through in what sort of frame size).
A larger buffer means more info per whatever cycle (so fewer cycles), and tradeoff is higher latency but more stable sound.. but a low buffer means less latency but you could end up with crackling etc. if too low.. more smaller cycles for the CPU to compute.
But finding a nice balance can make the latency round trip feel realistic (like a physical grand) - I like some latency so often have a high buffer size, even though I could make it 1.3ms if I wanted (I like around 10 which to me feels like a big grand action I like and I do change it for some music when I remember to do it). For synth things, less latency for sure can feel better if the playing is more about MIDI tick tock rather than classical musicality.
Sorry if I somehow hijack this topic, but I cannot find an answer on this question. Neither in the manual, nor somewhere in this forum
I hope you never worry about that IMO - ideas lead to other ideas and without that, I'd never have seen that video posted by Chopin87 which I really enjoyed all too much. I love those guys, have seen some other vids they made - and although haven't used their plugins, I'm taking a look later. They seem switched on.
I know I've read many threads here in the past about much of this - but maybe search for this copy paste the line as is, then scroll a while for some obv. titles to jump out..
sample rate buffer size
Ra wrote:Oh, another word about sample rates, somebody confuses them with bit rates that can necessitate a dither pattern in high frequencies. Like, FLAC, AIFF, WAV, OGG, AAC, MP3, and a lossy format dither is of no consequence whatsoever while you’re playing especially.
Incorrect Ra, thanks.
If you go the that 10min mark, you'll see the type...
If smoke ever clears, maybe then —and only then— you will begin to discover the errors of your own ways.
Then I understand that this sample rate (in the Devices menu) is the amount of samples that is send to the CPU per second. This sounds logical, since 96000 Hz (per s) x 2.0E-3 s = 192 (samples).
Also, what is the relation between sample rate and audio buffer size?
192kHz has sometimes less latency than 96kHz. At 192kHz a buffer of 32 samples happens twice as fast as it does at 96kHz.
If you’re proposing 192,000 samples per second makes an audible difference compared to 96,000 samples in the same time frame, you present no evidence supporting such a claim.
You may look at internal sample rates as computational speed occurring inside the software itself and device sample rates as any audio interface device or soundcard settings. I know I do.
When I change the Sample rate under the Devices tab, it immediately changes on my audio interface. From the software it affects the audio interface. Conversely, whenever I change the Internal sample rate under the Perf tab, my connected interface is unchanged. Internal sample rate as it implies is internal; it is inside the software. It has no affect on the hardware audio interface. Which is an external hardware device.
The difference was found in the long decay of complex polyphonic chords. I found more naturalness in sound space. But this difference is subtle. (After 20-30 seconds of sustain, differences can be discerned). It's not about the presence of high frequencies, but a 4-times more accurate calculation of the available frequencies.
Have you verified ‘a more accurate calculation’ via a spectrogram? One which shows comparative differences in a long decay, and in some histrionic.
scherbakov.al wrote:The difference was found in the long decay of complex polyphonic chords. I found more naturalness in sound space. But this difference is subtle. (After 20-30 seconds of sustain, differences can be discerned). It's not about the presence of high frequencies, but a 4-times more accurate calculation of the available frequencies.
Have you verified ‘a more accurate calculation’ via a spectrogram? One which shows comparative differences in a long decay, and in some histrionic.
You can listen to an example:
Converting to MP3 retained some of the difference. Only space has become flatter.
An example from 2017. Since then newer versions of Pianoteq have been released, maybe now there will be no difference .. I do not know. Did not check again.
With all due respect scherbakov.al, anybody even hears obvious differences between 44.1kHz sample rates and 48kHz sample rates. You do obviously at lower rates, like 22050Hz and 11025Hz. I was expecting to see something showing 192kHz versus 48kHz. Now from you I got instead maybe a 192kHz audio example and a 44.1kHz one, and besides no real answer to my question.
Can you hear the improvement with 192kHz vs. 48kHz internal sampling rate in a blind test?
But this difference is subtle. (After 20-30 seconds of sustain, differences can be discerned). It's not about the presence of high frequencies, but a 4-times more accurate calculation of the available frequencies.
And yes if I do say so now myself, differences are subtle.
Anybody has an idea whether or not inharmonicity is affected when you go from 44.1kHz sample rates just to 48kHz rates? That as you listen might account for differences perceived as prevalent in your auditory system. That inharmonicity perhaps affects Spectrum profile in turn. Which in itself more than likely depends on Internal sample rate Options you selected before or after any Output.
Pianoteq Pro offers the "internal" sample rate of 192kHz (which stage and standard do not include).
Have you tried the "internal" sample rate of 192kHz and if so what do you think of it? Can you hear the improvement with 192kHz vs. 48kHz internal sampling rate in a blind test?
Osho
Honestly, I was unaware you could just use 192kHz as an Internal sample rate even though your audio interface might max out at 96kHz like mine!
Thank you, Osho.
Oh, another word about sample rates, somebody confuses them
That is what I referred to as "false" - I believes ya knows that - don't ya buddy?
In the past I posted about dithering. For one example, here's what Ra said in one thread:
Have just done some close examinations from a few of Qexl's points. They to me are excellent!
.
His own words show, I brought these ideas to him.
Link to that quoted thread here.
Here's a pertinent part where I informed him of these notions around dithering.
Then when exporting (down) to 24 bit rate and 48kHz sample rate which would likely invoke some choice about dithering, no dithering and/or normalization which might need a few changes for best results.
At that time, I was happy to have shared this knowledge which might help him in his ongoing learning.
If I ignore the insult and all, here's more info..
I recommend to watch the whole video posted by Chopin87 at top.. at around, or maybe before minute 10, or from the start really, it's worth it.. he's absolutely nailing a presentation of the way that working with frequencies in a DAW is different at the different sample rates.
All I tried to add to the understanding here is that, depending on output which might include dithering options, a downsampled 44.1kHz version of something which was once heard at 192kHz could be very different because of all the technical choices during downsampling.. so comparisons are quite often not 1:1 - just an observation.
Anybody has an idea whether or not inharmonicity is affected when you go from 44.1kHz sample rates just to 48kHz rates? That as you listen might account for differences perceived as prevalent in your auditory system
Absolutley related things (inharmonicity being high frequency related) - like I suggest Ra, go to that video, and see how the demo with sine waves sounds (he shows these things beautifully and it's why I mentioned 'synths' lot of high freq's. - the simple wav forms coupled with these artifices esp. in high frequencies, can be powerful components in your sound design, more than just layering a track over another.. these things can be complex quilts woven quite intricately - and playing with the technical areas of it can lead to some artistic satisfaction - all I'm trying to gift people with here).. any high frequencies exactly relate to my earlier postings. Inharmonicity may show up in high frequencies and be subject to some intermodulative processing.
Like scherbakov.al mentioned, you may hear a sweeter 'tail' to chords etc. at 192 - one reason being exactly what I was hoping to express.. these 'artifacts' of intermodulation can make 'pure' waves noisy ast lower sample rates.. again it's both good and bad though.. and I pointed out, I like 48.. and Chopin87's video made my day - I didn't know before this, that 48 has such a nice technical situation which applies to my desires.
So, working on piano at 192 is good.. I hinted at that earlier..
pertinent to me with some synth things but can be applied to thinking about 192 (or up-sampled FX) for piano.
With different sample rates in the DAW, all kinds of frequency effects run rampant esp. with effects on top. It may be different for given instruments, effects, the genre - but your choices can help or hurt the outcome.
You choose the rates you want to work in, partly because of what you want to either keep, or destroy in the working digital signal. You can create 'tape' vibe with some simple destructive digital planning and I love going 'glitch' to degrees.
Next, you mix - whatever other production you want to apply (again, you can choose certain compression types, which variously help to keep or destroy values in the signal)..
Next, if you want to output to a consumer sample rate 48 for example, you might apply at that last stage, any given type of dithering, or not.
In the case of keeping a cleaner output to a non-lossy format (like flac or WAV) - you may avoid dithering - but a lot of the time, the process of downsampling really does leave some harsh areas, masking problems - so. This is one reason dithering is done.. to help the human ear - the translation downward is pertinent to what we hear.
If outputting to and MP3, the MP3 format's own compression routines may cause some negative consequences for the downsampling too.. so you may not want to dither. Again, variables incl. different music types, what rates it was 'worked on' and more, just an example.
I hope you understand I'm trying pretty hard (over a long time line) to give information you can take onboard.. not trying to just "say stuff" - and please quit with the insinuations OK?
In the end.. it does depend on hearing the same thing on the same equipment producing the same level of authenticity to the signal, otherwise we can be comparing apples to oranges - just hoping to shine that light on it.
Sincerely - just hoping this helps people - not trying to 'win' stupid prizes mmmmkay
Man, you open with a paraphrase, misquote, or omission, or, each.
Anyway, in any event I'm hardly misled by your assertions!
Now what interests me is some of the following:
There is a reason why 48000.
A little video on the argument which perhaps will help unaware users understand better.
No change log indicates an adaption of a 48kHz ceiling under Options. Please correct me if I'm wrong; 48kHz is a new Sample rate limitation of both PIANOTEQ Stage and Standard. I take it if I am right end users essentially accepted a downgrade and without so much as a single protest!
Nobody appears at all concerned about an omission in the change log.
EDIT: I left off a period: picked up a bad habit. Have to steer clear of bad influences.
Please correct me if I'm wrong; 48kHz is a new Sample rate limitation of both PIANOTEQ Stage and Standard. I take it if I am right end users essentially accepted a downgrade and without so much as a single protest!
Nobody appears at all concerned about an omission in the change log
Internal sample rate of 48 kHz has always been the maximum for stage and standard.
I've a PIANOTEQ v.3.6.8 copy right now on my desktop. Correct, 48kHz is its Internal sample rate maximum. I was looking under Devices.
While under your Perf tab you can downsample, any upsampling is yet unsupported. However, on my copy of PIANOTEQ PRO, it permits a 192kHz sample rate in file exports (although my audio interface maximum is 96kHz).
Thank you Viridis, for your prompt reply.
If anybody’s sharp enough to catch it, my mention (above) of file exports is about piano files (recordings) I play. I’m trying desperately to meet a criteria here:
Thanks for the input.
Anyone else? I am interested in just playing piano - not in mixing it with effects etc..
Osho
I was just playing some piano music!
Here’s some music; as I play it I can of course just faintly hear a piano part.
Thanks all for the input. I think this thread settles it for me. It's not worth to upgrade to Pro for $300 just to get 192kHz for home usage where one just wants to play the piano live. I would rather spend the money on buying some instruments or for the upgrade to Pianoteq 7 (which should be this year as per the historical trend...)
Thanks again all.
Osho
Thanks all for the input. I think this thread settles it for me. It's not worth to upgrade to Pro for $300 just to get 192kHz for home usage where one just wants to play the piano live. I would rather spend the money on buying some instruments or for the upgrade to Pianoteq 7 (which should be this year as per the historical trend...)
Thanks again all.
Osho
You're wise. The main benefits of Pro are in the control of the Presets rather than the increased sampling.
The only use case where I've encountered needing a higher internal sampling rate is if PTQ is being integrated live into broadcast or professional production pipeline hardware that has a specific inbound rate requirement either for convenience (like in a production pipeline where data is being shared in multiple workflows through multiple production iterations and shouldn't be getting dithered as its sent back and forth between technicians and contributors) or for live broadcast hardware requirements (where again dithering/resampling either may not be possible between inputs on some hardware or where it could create artefacts or other problems because of downstream mismatches). While needing 192 is usually fairly rare as a requirement in any setting, I regularly have studios request 96 so they can downmix and dither at the end of their pipeline rather than try to mix in something that has to be upsampled before entering the pipeline. As more and more is done in-the-box even for live or production workflows using circuitry that encodes and decodes using calculus instead of statistics (like delta-sigma DACs), the practical reasons for different sample rates will likely decrease significantly over the next ten years as more legacy digital hardware is upgraded with newer, faster, more flexible hardware--especially if the trend towards streaming instead of broadcast continues.
There are particular format reasons for why these sample rates are set to what they are (NTSC American audio formats linked to multichannel color at 44056, Audio CD max file rates at 44100 because of Sony's PAL/NTSC average, early ADCs and DACs at 48000 for higher-fidelity than CD, etc.), but any additional sampling above 50kHz doesn't improve the listeners sound of something, it just increases the headroom for the amount of FX, dubbing, processing, and downsampling a track can handle without the foldback artefacts mentioned above. As delta-sigma converters and their pulse-density modulation have replaced most traditional pulse-code modulation in digital signal processing and CPUs and GPUs doing more and more heavy-lifting in digital audio production, this whole question could functionally disappear within a few years. Already in most of today's hardware both consumer and professional, data compression ratios and codecs are the most important factors for sound quality, while concerns around sample rates and bit depth have evaporated. Even network bandwidth and datastream size are functionally irreverent as 4K video formats/codecs/encodings are so many Mbps that it doesn't matter how large the audio packets are, because the video will always be so many magnitudes larger in the datastream nobody cares how much space the audio uses anymore--a far cry from the SD DVD days. The problems with codec containers and bit rate get even worse when trying to compare any shared examples because nearly any website that accepts uploads will re-encode, compress, and resample them anyway.
My experience is that if you need this feature someone else will tell you as part of any agreement to contribute to a workflow pipeline, but if you're working on your own or just want to play the piano, the difference is practically non-existent.
Further reading on sample rates and delta-sigma DAC algorithms for anyone interested:
https://en.wikipedia.org/wiki/Sampling_...o_sampling
https://en.wikipedia.org/wiki/Delta-sigma_modulation
That's such superb and thorough info and advice tm, thank you.
if you're working on your own or just want to play the piano, the difference is practically non-existent.