• If you haven't done so already, please add a location to your profile. This helps when people are trying to assist you, suggest resources, etc. Thanks (Click the "X" to the top right of this message to disable it)

Recording an Accordion

When you say lifeless, that might be the reason I'm asking in the first place, just something felt like it might be missing. My ukulele only recordings feel the opposite to me but I'm guessing it's because the accordion may need different treatment due to being a very different instrument. Any ideas on how to change that?
I do have a few suggestions.

So the room is a small rectangle (9x10' maybe?) and has stuff in it but isn't treated and doesn't have lots of soft or dense items. I think what I can do with the room will be relatively limited. I sit at a desk facing one wall.
Does the room have a closet with some clothes hanging on hangers in it? Is there a bed or anything else in there? Any carpets? Any large windows with or without cutrains?

I do push the gain on the 2i2 to try and raise the volume until I think I'm just avoiding hiss, 47db is what I've been using recently. To record I just use Audacity and bump the gain on the recording by another 10/15db but nothing else. I feel like I need to figure out what other post processing may improve things.
OK, so here are a few of the basics, you may be doing some of these things already and they are something that you could experiment with...

1 - When recording you do NOT want to be pushing the recording levels anywhere close to 0db... indeed I make all my recordings (for both audio and videos) somewhere between -12 and -18db as a maximum recording levels. One doesn't need to worry about final levels until the last step of the process as it depends a LOT on what the destination use will be.

2 - Sound controlling the room is likely the biggest thing that most people don't pay attention to and then wonder why things sound off. Live reverberant rooms are sometimes a but hard to tame, but there are a few ideas. Now, to be clear, we are talking sound CONTROL, not sound deadening. They are 2 different things. Sound deadening is where are cut down exterior noises from coming in to our room... sound deadening is where we control the sound better from bouncing around inside the room where you are recording. Both are important, but deadening is way harder and more expensive.

- If you are making an audio recording only and are in a bedroom, place your recording position where you have an open closet with clothes hanging and play directing the sound into the closet (imagine mic(s) just outside the closet, you sitting down facing the mic looking at the clothes. The clothes serve as a device that reduces room ambience.

- If you have some moving blankets (thick blankets that are used to cover furniture when moving) are great at cutting down unwanted echo. In some cases, one can purchase them very cheaply. I purchased 20 blankets from Harbor Freight for $3 each on sale. They are a bit smaller but I can place them on a hand made wooden frame and cut the ambient size of a room in half. For me, I used them more as protective covers to place over my Corvette in the winter, but I did experiment with them in my basement where they did not make a big difference. Where I tried them once and heard a nice improvement was in the area between our living room and dining room... cheap and easy. Sometimes you can find places where they have the big heavy moving blankets that are damaged for near nothing, and with a little easy repair work, can make some very easy dividers.

3 - I mentioned it before... get the mics closer. I was doing some experimenting today upstairs in my dining room, and the sound difference between the sound that the cellphone camera caught versus the two small pencil condenser mics (using the cardioid heads) that were less than 2 feet in front of me was very pleasantly surprising. camera mic caught all the echo and sound was weak and thin. The condensers caught the sound without echo and was a little richer and clearer.

4 - Post processing. Training yourself to have a good ear and practicing so that you can improve the sound captured is worth the time and effort. If you go to YouTube and listen to the best sounding videos, you can bet that they have passed through the hands of someone that tweaked the sound. I have a little motto here... "a little does a lot", so don't go nuts with any one EQ or effect, it is easy to go over the line.

5 - Final recording levels. KNOW WHERE YOUR FILE IS GOING TO BE USED. Different locations have different demands on the final levels. For example, if I am making a recording in MP3 format to be placed on a CX, or a WAV file for a CD, I normalize the files to max out at -1db for best output (loudest), but still never touch that 0db mark because anything above that is clipping and in the digital audio world, we distort in a bad way and that is something we never want to do.

In the "old days" using analog recording methods, it was desireable to go past 0db on the VU meters as they had additional headroom and this made the recordings more "crunchy" and added a natural compression. You CANNOT do this in the digital world, you just clip and the song sounds like crap.

There are different places/ways that a file will be used. In the broadcast world, most channels recommend a max volume of -12db so everything you watch on the TV is set to these levels (that gives them 12db to raise the commercials so those are heard better and are more invasive). For this reason, audio at -12 dB sounds normal. On the web, virtually all producers target 0 dB, and web viewers are used to this higher volume... but it should NOT always be so because some places modify your file in potentially very negative ways and that is something we want to avoid if possible.

A couple examples:
- YouTube and now Spotify use the -14LUFS standard, meaning that if your song is set to average out at around 0db, they will known it down -14db often adding some compression or using a lower quality level.

Know your destination and adjust your levels to match the requirements of where/how that file is going to be used.

CDs use a 44.1 Hz level, DVDs and most movies today use 48Hz level.

Did you know that YouTube now supports 24-bit 96Hz audio? That is what I upload to YouTube... using the -14LUFs standard, of course. Now, in music, most people won't hear a big difference between 44.1, 48 or 96Hz, but for me, in back to back tests, I found that I could hear smoother highs and generally more dynamic range when using 24-bit 96Hz, so that is the lowest level I record at. My audio interface can capture at 192hz and when using that, I capture at the highest level and output it to 24/96 for upload purposes.

People listening to cell phones or low quality speakers will never hear the advantages, but I can and enjoy knowing its there for those that can appreciate it.
 
Last edited:
With dynamic mics the preamps make all the difference. WIth condensor mics the preamps that matter most are the ones inside the microphone, to give them a boosted signal. For my humble applications I wasn't going to spend twice the money for C414 mics over the C214.

You may be using less gain, but your device preamps override the higher output of the phantom powered mic preamps in the chain of events and it still will have a lower dynamic range and more noise than using a quality audio interface because all is transferred from mic preamps to device preamps, modified and then converted from analog to digital.

Also, if your mic usage is more or less basic, using the C214 is perfect for your needs, but since the C214 only has the single cardiod pattern, you cannot take advantage of other kinds of micing styles like Blumlein, mid-side and other micing techniques that can take advantage of all the polar patterns like the Omnidirectional, Figure-8 and HyperCardioid patterns.

Playing around with different configurations is both fun and adds another dimension to the sound. Example of the mid-side setup that uses the figure-8 pattern for the bottom mic and cardioid or hyper-cardioid for the top mic.

1734386609556.png
 
With dynamic mics the preamps make all the difference. WIth condensor mics the preamps that matter most are the ones inside the microphone, to give them a boosted signal. For my humble applications I wasn't going to spend twice the money for C414 mics over the C214.

You may be using less gain, but your device preamps override the higher output of the phantom powered mic preamps in the chain of events and it still will have a lower dynamic range and more noise than using a quality audio interface because all is transferred from mic preamps to device preamps, modified and then converted from analog to digital.
That reads mostly as word salad.

If you want to have low noise, it pays off to boost the signal strength early on so that it dwarfs the inherent noise of successive stages. For that reason, "hot mics" can get by with less impressive preamps afterwards. Electret microphone elements have super-low output, so a single JFET preamplifier is placed within few millimeters of the element, forming an integrated part of an electret microphone capsule. Dealing with the signal afterwards is a lot less critical. "Proper" condenser mics may or may not boost the phantom power voltage level for higher gain, but they will either way add mic-internal preamplification afterwards, and those producing a "hot" signal will work admirably with comparatively noisy audio interfaces.

One device where the difference is very obvious are ribbon mics. Traditionally they come without a preamp and require very good preamp stages for your soundcard or other device; often needing gains along +60dB. Modern ribbon mics can include a preamp in which case they require phantom power (phantom power on an old-style ribbon mic can kill it due to switching artifacts, so you better get it right). And then the subsequent preamp is a lot less critical, and you can get by with higher noise levels.

Of course the noise floor and the gain of a preamp are only one characteristic; other qualities may impact the signal at any gain.
 
...
CDs use a 44.1 Hz level, DVDs and most movies today use 48Hz level.

Did you know that YouTube now supports 24-bit 96Hz audio? That is what I upload to YouTube... using the -14LUFs standard, of course. Now, in music, most people won't hear a big difference between 44.1, 48 or 96Hz, ...
I always record in 16 bit 44.1 kHz so that it can go straight onto a CD. At my age (65) I doubt that I could still hear the difference between this and 24-bit 96kHz (but I have not tested that). The only time I use 48kHz is when I do tuning because my phone cannot measure the C#8 reed when sampling is set to 44.1kHz but can when set to 48kHz. Al long as I can still hear the C#8 reed well (it is the highest note for which reeds exist) I am satisfied with my hearing. I guess I will be able to keep playing longer than I will be able to do tuning, because my playing is mostly on the bass accordion ;) .
 
I always record in 16 bit 44.1 kHz so that it can go straight onto a CD. At my age (65) I doubt that I could still hear the difference between this and 24-bit 96kHz (but I have not tested that).
Converting between 44.1kHz and 48kHz sampling frequency causes more artifacts than going down from 96kHz. Also tone control creates warped results (higher attenuation/gain per octave) near the Nyquist frequency, so you get more natural results when working with 96kHz samples. 24bit gives you a lot more dynamic range for recording/mixing/processing.

44.1kHz/16 bit is fine as a format for the end results in my book. As a working format, it is too limited. It is suitable for playback, not for recording and mastering. I've settled on using 96kHz/24 for that. For practice recordings that are only going to see leveling/cutting but no tone control, reverb, etc, i record at 48kHz/24. Those are never going to see a CD.

By the way: high quality converters all do oversampling internally and then bring the frequency down in the digital domain. That requires a compromise between quality of results and latency. If you bring the frequency down in post-processing, latency is not an issue. So leaving the critical part of the downsampling for the mastering stage may get marginally better results than letting the converters/soundcards do the job online. It is also likely to give you less monitoring latency if you are monitoring.
 
Last edited:
I always record in 16 bit 44.1 kHz so that it can go straight onto a CD. At my age (65) I doubt that I could still hear the difference between this and 24-bit 96kHz (but I have not tested that). The only time I use 48kHz is when I do tuning because my phone cannot measure the C#8 reed when sampling is set to 44.1kHz but can when set to 48kHz. Al long as I can still hear the C#8 reed well (it is the highest note for which reeds exist) I am satisfied with my hearing.
Well, thats when your experience shows... using the tools to do the heavy lifting for precise tuning projects and using your skills to accomplish that tuning. CD standards are 44.1 but they did raise the standards for broadcast devices to 48k. I see using 96 and 192 in the same manner like using RAW video files straight from the sensor, you capture EVERYTHING and filter it down to what is needed. It's easy to remove... impossible to add one that data is stripped away. For me I can hear the differences between a 48hz recording and a 95hz recording when it's the same thing played back to back. I agree that no one on earth would be able to hear a piece of music and say "that was recorded at XXhz".

I guess I will be able to keep playing longer than I will be able to do tuning, because my playing is mostly on the bass accordion ;) .
Aren't you glad that you didn't choose to play a piccolo ? :D :D
I've tried a Hohner bass accordion, lovely sound. I think that in a ensemble or orchestral accordion environment it adds a lot of the secret sauce that makes it sound complete and richer. For me, my hearing is good so far, but like everyone else it will eventually degrade, if I am lucky enough to live that long. :)
 
For me I can hear the differences between a 48hz recording and a 95hz recording when it's the same thing played back to back.
Assuming you are actually talking about sampling frequencies of 48kHz and 96kHz (rather than 48Hz and 95Hz) and no difference other than good-quality anti-aliasing filtering and downsampling, I am pretty sure that you could not guess statistically distinguishable from random in a double blind test (where neither tester nor yourself knows what they are dealing with). And it saves you from annoyances your cat and dog cannot get away from. In the times of CRTs, the world was divided in people annoyed by the 15.625kHz whine of normal TV signals and those impervious to it, and age tends to move the former to the latter eventually. 20kHz, in contrast, is a pretty safe bet. However that does not mean that a device working with 48kHz sampling frequency will actually reproduce all the way up to 20kHz, and it does not mean that it will do so without artifacts.

You need a rather high-quality A/D conversion to get close to those theoretical limits. And when your device does not have that quality, feeding it with the equivalent 96kHz signals may lead to a better conditioned signal. That is not a problem of the sampling frequency but of your replay device, however.
 
If anyone has an interest, when I get a couple hours time, I could do a few experiments with some different setups in my dining room (a not so nice place to record) to show that one still can get pretty good results (... or not?) in a very imperfect location/situation.

I am personally curious as to what the results will be. I could grab my pair of mid-range condenser mics and then do a couple of tests with the SM57 and SM58 clone (dynamic mics) and compare that all to a cellphone mic.

I was doing some thingking... to help make it a it easier, I could use the recording function from my lowest quality level recording device, my QSC TouchMix-8 because it maxes out at 48hz (it's very good, but not amazing), but creates a 32-bit floating WAV file, which means even if the settings on the mixer are bad the recorded files will NEVER be clipped or distorted.

Also, because the mixer has up to 8 inputs, I can make one recording for the 4 mics in a couple different setups (how do the condensers work at 18 inches and 4 feet away), and show how things work.

Dynamic mics are easy... they suck at anything over 18 inches away and take a ton more gain, so the closer the better and I'll likely just max out the gain on my QSC TouchMix 8 (60db of analog gain and I have an additional 15db of digital gain if needed) and keep them VERY close.

Let me know if anyone finds this interesting or not. :)
 
  • Like
Reactions: Tom
I was doing some thingking... to help make it a it easier, I could use the recording function from my lowest quality level recording device, my QSC TouchMix-8 because it maxes out at 48hz
You mean at 48kHz. 48Hz is the frequency of a note somewhat below G1 (the E1 that is the lowest note of a bass guitar and a number of accordions is about 41Hz).
(it's very good, but not amazing), but creates a 32-bit floating WAV file, which means even if the settings on the mixer are bad the recorded files will NEVER be clipped or distorted.
That it will not get clipped or distorted after conversion to digital from the point where samples are converted to 32-bit floating point (which on numerous devices is not even the processing format but merely a storage format option). There is plenty of opportunity to clip and distort in the analog domain and/or during analog-to-digital conversion and at any point before the conversion and processing to floating point happens.

I'd need to read the advertising materials for the Touchmix to make an educated guess (and it will not be more than an educated guess since engineering and marketing tend to be only loosely on the same page) as to what significance the floating point option actually has for processing.
 
I've finally had time to have another play. I changed the mic postiion and angle, tweaked gain, playing volume etc. and I think I'm happy with the recording now, examples here: https://filebin.net/na75y2d1xr21wyy0
Any recording feedback is welcome.

Of course, I still need to improve my playing, get my instrument looked at, avoid breathing into the mic and learn about post processing but atleast on the recording side I feel like I'm going in the right direction with the SM57 in terms of volume. I did also realise I should check the playback volume is at what would be reasonable e.g. atleast half volume if not closer to full.

If would be good to see any more videos that might help with recording though.
 
I've finally had time to have another play. I changed the mic postiion and angle, tweaked gain, playing volume etc. and I think I'm happy with the recording now, examples here: https://filebin.net/na75y2d1xr21wyy0
Any recording feedback is welcome.

Of course, I still need to improve my playing, get my instrument looked at, avoid breathing into the mic and learn about post processing but atleast on the recording side I feel like I'm going in the right direction with the SM57 in terms of volume. I did also realise I should check the playback volume is at what would be reasonable e.g. atleast half volume if not closer to full.

If would be good to see any more videos that might help with recording though.
I started a new thread with a LOT of added info and posted a video on Youtube comparing the SB57, clone SM58 and Neewer NW-410 condenser mics... BTW, all have a cost of about $100US... the Neewers are $100US for TWO condenser mics and 3 heads each (cardioid, super-cardioid and omni)!

To save you the time looking... here is the video:



In short, I found about 25db difference between condenser mics and dynamic mics, and a slightly nicer/more natural sound sound.

After this, I am not all that impressed with the Shure mics all that much. Too much work to get them to give you a clean signal. I had my mixer cranked all the way up to maximum AND had to add 15db in post. My cheap condensers are about 25-35db louder and smoother sounding.
 
Last edited:
I started a new thread with a LOT of added info and posted a video on Youtube comparing the SB57, clone SM58 and Neewer NW-410 condenser mics... BTW, all have a cost of about $100US... the Neewers are $100US for TWO condenser mics and 3 heads each (cardioid, super-cardioid and omni)!

To save you the time looking... here is the video:



In short, I found about 25db difference between condenser mics and dynamic mics, and a slightly nicer/more natural sound sound.

After this, I am not all that impressed with the Shure mics all that much. Too much work to get them to give you a clean signal. I had my mixer cranked all the way up to maximum AND had to add 15db in post. My cheap condensers are about 25-35db louder and smoother sounding.

Interesting video. I think as you said, the Shures are probably less forgiving, I know I definitley have mine closer than you did and I've tried different angles too. I think as you said before, the stereo is probably the biggest improvement worthwhile, probably because of the nature of the accordion I.e. it throes out sound in a field in front of it.
 
To save you the time looking... here is the video:



Thanks for sharing, especially at a time when I'm taking a new look at my recording facilities.
Until now, I've had the choice of Sennheiser dynamic E835s, Samson C03 condensers, or the built-in condenser mics on my Tascam DR40X.
The Samson mics were first used at a recording studio, and I was impressed with how they recorded my vocals. Not crazy expensive (about £70 each) they have three patterns on a switch - cardiod, 8 and O shapes.

The Tascam is great for recording a single track - the built in mics are good, but while it is in theory 4 track, it's only has two independent tracks. Assuming I've been good and not naughty, I'll have a Zoom H8 tomorrow. 6 independent tracks, plus the option to buy a 4-track expander module. I'll set aside some time to do the same as you have in the video.
 
Interesting video. I think as you said, the Shures are probably less forgiving,
It's not so much of being less forgiving as much as I need to push my audio preamps to 100% and still not having it loud enough to get proper recording levels. If I didn't record with a $1500 digital mixer that had very low noise, I would be spending 30 minutes alone on EVERY recording removing hiss.

My advice is that with the $100US you want to spend on a 2nd SM57, go straight for the matched pair Neewer NW-410. Double the number of mics and you get 3 times more polar patterns on each mic. If you take the time and save about $150, I recommend the Shure AT2020. That's just my suggestion, of course.

Until now, I've had the choice of Sennheiser dynamic E835s, Samson C03 condensers, or the built-in condenser mics on my Tascam DR40X.
Assuming I've been good and not naughty, I'll have a Zoom H8 tomorrow.
Of the 3 mics that you mention, the Samson C03 will definitely get you the best results. My concern is their price... a single C03 is double the cost of 2 NW-410s and about $20 more than a Shure AT2020. For ultimate budget and still get a good sound, the NW-410 just cannot be beat, or of you are saving a little longer, like mentioned above, a pair of AT2020 is what I would recommend. The AT2020 are the cheapest LARGE condenser mics with a really good sound (as opposed to the NW-410 or Samson C03, which are small condenser mics).

I see many musicians prefer the small pencil condenser mics due to their ability to record musical instruments with great quality and as you may know Piotr uses the Shure AT2020 (large diaphram condenser mics) to perfection, so you can see how those sound!

6 independent tracks, plus the option to buy a 4-track expander module. I'll set aside some time to do the same as you have in the video.
When I was shopping around for a straight recorder, the H8 was one of those that I "interviewed". Back then it was nice, but not the quality level that I wanted. A little bit plasticy, I had concerns about if it was dropped and the noisy preamps.

They've since improved the preamps on the H8 since then. They are now a little quieter, but the gain is still not as high as I would wish it to be (at around 56db). This is NOT an issue with condenser mics but would be with dynamic mics. I'd have to call the H8 the ultimate PORTABLE recording device that gives you a lot of options and quality for a reasonable price.

In general I like the company (ZOOM), they made my F4 field recorder, and I really enjoy it and use it every chance that I get.

For comparison purposes, my Mackie 1640i console has 65db of gain, the QSC Touchmix8 has 60db (with an additional 15db digital gain available) and the Zoom F4 has 76db of gain. More importantly, all have preamps with very low self noise of over 70db This means that none of my devices will introduce audible hiss even while at full gain settings and that is what I was looking for (as a starting point).
 
Of the 3 mics that you mention, the Samson C03 will definitely get you the best results. My concern is their price... a single C03 is double the cost of 2 NW-410s and about $20 more than a Shure AT2020.

We were in the slightly unusual position that the studio we used were using the C03 - I liked how they recorded with my voice, so didn't bother looking at any other model, just bought the one I knew.

They've since improved the preamps on the H8 since then. They are now a little quieter, but the gain is still not as high as I would wish it to be (at around 56db). This is NOT an issue with condenser mics but would be with dynamic mics. I'd have to call the H8 the ultimate PORTABLE recording device that gives you a lot of options and quality for a reasonable price.

It was the portability that swung it for me. I almost bought a Tascam Model 12, which would have done everything I needed. But then I did some recording with a friend with a Zoom L-12 LiveTrak and that made me think about Zoom products. The portability of the H8 won me over as I'm planning to do some outdoor recording next year, as well as it's just so easy to throw it in a bag and run it from batteries.
 
Is that the Audio teknica At2020?
Sorry yes, thats what happens when you get under 1 hour of sleep in a night... lol.

Piotr uses the Audio Technical AT2020:
Screenshot 2024-12-24 at 4.54.39 PM.png

For me personally if I wanted something from Audio Technica I would go for the AT4050 because I like the multiple polar patterns. I listened to many mics, and honestly the one that won me over was the matched pair of the sE Electronics sE4400a:

Screenshot 2024-12-24 at 5.04.50 PM.png
 
Last edited:
Back
Top